framework,version,device,op_name,kernel_source,gemm_dtype,m,n,k,latency
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,65536,16384,24.82626749674479
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,65536,10240,15.277792358398438
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,65536,12288,17.394563802083333
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,65536,16384,26.034574381510417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,65536,12288,18.96243693033854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,65536,12288,18.65221150716146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,65536,16384,22.759214274088542
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,65536,8192,10.184027099609375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,65536,10240,13.959281412760415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,65536,10240,15.226302083333334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,65536,8192,12.538848876953125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,65536,8192,9.608363850911458
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,65536,7168,8.521646118164062
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,65536,7168,8.454668680826824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,65536,6144,6.867800394694011
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,65536,7168,12.6120361328125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,65536,6144,6.747097778320312
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,65536,6144,9.791269938151043
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,65536,5120,5.426720174153646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,65536,5120,8.599312337239583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,65536,5120,5.415989176432292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,65536,4096,4.2699030558268225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,65536,16384,12.803988647460937
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,65536,4096,6.416694641113281
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,65536,4096,3.778967539469401
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,65536,3584,3.5343221028645835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,65536,12288,9.826837158203125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,65536,8192,6.581892395019532
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,65536,3584,5.40992431640625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,65536,3584,3.336996205647787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,65536,3072,3.006579081217448
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,65536,3072,4.658751932779948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,65536,7168,5.600797017415365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,65536,3072,2.542397816975911
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,65536,10240,8.314791361490887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,65536,2560,2.5140970865885417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,65536,2560,2.20542729695638
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,65536,2560,4.101735432942709
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,65536,2048,2.000941848754883
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,65536,2048,3.300248463948568
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,65536,2048,1.7375935872395831
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,65536,1536,1.526253890991211
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,65536,6144,4.74535878499349
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,65536,1536,2.4494272867838545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,65536,1536,1.2821375528971353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,65536,1024,1.0836341857910157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,65536,4096,3.275279998779297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,65536,5120,4.20495351155599
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,65536,1024,0.949069849650065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,65536,1024,1.6714324951171875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,65536,768,0.8690709431966146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,65536,768,1.2726720174153647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,65536,768,0.7786176045735677
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,65536,512,0.6978506724039714
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,65536,512,0.5656384150187175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,65536,512,0.8850325266520181
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,65536,512,0.6356170654296875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,65536,256,0.5933418909708659
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,65536,3584,2.903875223795573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,65536,256,0.5769909540812175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,65536,256,0.5044949213663737
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,65536,128,0.48034133911132815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,65536,1536,1.2592821756998698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,65536,128,0.44226452509562175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,65536,128,0.4313045183817546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,65536,2560,2.11484375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,65536,64,0.48767999013264973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,65536,64,0.4105418523152669
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,65536,64,0.4501066525777181
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,65536,32,0.5118240038553874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,65536,32,0.4007552146911621
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,65536,32,0.45687360763549806
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,65536,768,0.7625792185465495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,65536,3072,2.3673728942871093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,65536,2048,1.7460372924804688
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,16384,16384,6.704669698079426
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,16384,65536,34.62806803385417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,65536,27.468585205078124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,16384,6.60186767578125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,16384,4.918103535970053
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,16384,12288,5.157460021972656
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,65536,26.590940348307292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,12288,4.994381713867187
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,12288,3.397129567464193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,16384,10240,4.224399058024089
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,10240,4.148047892252604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,10240,3.053496551513672
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,16384,8192,3.067841084798177
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,65536,1024,0.92651735941569
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,8192,3.35523935953776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,8192,2.0363157908121745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,16384,7168,2.6088404337565105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,65536,128,0.44539947509765626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,7168,2.555450693766276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,7168,1.768705113728841
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,16384,6144,2.281177520751953
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,65536,256,0.474348799387614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,6144,2.377129618326823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,6144,1.4335914611816407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,16384,5120,1.8625418345133464
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,16384,16384,3.3629313151041664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,16384,65536,13.141791788736978
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,5120,1.1453024546305337
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,5120,1.9583253224690753
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,16384,4096,1.530170694986979
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,4096,1.9718143463134765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,4096,1.0178719838460286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,16384,3584,1.3021845499674478
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,16384,12288,2.4092971801757814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,3584,1.323305638631185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,3584,0.8662570953369141
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,16384,3072,1.150162124633789
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,16384,10240,2.1241845448811847
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,3072,1.1589888254801433
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,3072,0.7054474512736003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,16384,2560,0.9497909545898438
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,16384,8192,1.6650250752766926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,2560,0.9390506744384766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,2560,0.5714848200480144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,16384,2048,0.7624373118082682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,16384,7168,1.3587168375651042
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,2048,0.7587146759033203
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,2048,0.4525290807088216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,16384,1536,0.5744277318318685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,16384,6144,1.1733375549316407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,1536,0.5738911946614583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,1536,0.3359018643697103
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,16384,1024,0.3973141352335612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,16384,5120,0.983670425415039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,1024,0.3958944002787272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,1024,0.2398837407430013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,16384,768,0.31371307373046875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,16384,3584,0.686192003885905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,768,0.3270847956339518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,768,0.2127317269643148
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,16384,512,0.22674239476521812
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,16384,512,0.16539306640625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,512,0.21921812693277992
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,512,0.15818559328715007
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,16384,256,0.16995733578999836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,16384,4096,0.8002826690673828
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,256,0.1347274621327718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,256,0.12858346303304036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,16384,128,0.14588267008463543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,16384,2560,0.508461856842041
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,128,0.11391359965006512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,128,0.11147200266520182
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,16384,64,0.14618132909138998
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,64,0.10567359924316407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,64,0.11851306756337483
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,16384,32,0.1474463939666748
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,32,0.10356693267822266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,32,0.12179306348164876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,16384,3072,0.6174346923828125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,12288,65536,28.022930908203126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,16384,1536,0.3336480140686035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,16384,2048,0.4196149190266927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,12288,16384,5.700886535644531
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,65536,19.31290079752604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,65536,18.828861490885416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,16384,4.806813049316406
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,16384,4.197176615397135
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,12288,12288,4.2099978129069004
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,12288,3.711080423990885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,12288,2.742552439371745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,12288,10240,3.4795519510904946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,16384,768,0.20710399945576988
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,10240,2.640207926432292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,10240,2.152429962158203
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,12288,8192,2.589146677652995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,16384,1024,0.24501652717590333
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,8192,2.1470667521158857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,8192,1.4631125132242837
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,12288,7168,2.2580586751302083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,16384,128,0.11433706283569336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,7168,1.8691786448160808
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,7168,1.3538570404052734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,12288,6144,1.940335973103841
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,16384,256,0.12151467005411784
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,6144,1.662176005045573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,6144,1.1453472137451173
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,12288,5120,1.6128309885660808
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,12288,65536,10.44710184733073
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,5120,1.4069588979085288
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,5120,0.8956053415934244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,12288,4096,1.279750442504883
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,12288,16384,2.4919156392415363
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,4096,1.110377629597982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,4096,0.7176810582478841
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,12288,3584,1.1252639770507813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,12288,12288,1.9218719482421875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,3584,1.0148864110310873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,3584,0.6192874908447266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,12288,3072,0.9717664082845052
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,12288,10240,1.458571751912435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,3072,0.8054655710856119
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,3072,0.5418912251790364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,12288,2560,0.8431242624918619
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,12288,8192,1.1768671671549478
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,2560,0.6807125091552735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,2560,0.42915627161661785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,12288,2048,0.6514634450276693
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,12288,7168,1.0516063690185546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,2048,0.5594677607218425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,2048,0.34198827743530275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,12288,1536,0.5021877288818359
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,12288,1536,0.25288000106811526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,1536,0.41102720896403
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,1536,0.25818026860555016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,12288,1024,0.3403530756632487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,12288,6144,0.909177589416504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,1024,0.30753278732299805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,1024,0.1962997277577718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,12288,768,0.2759968121846517
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,12288,5120,0.7399232228597005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,768,0.2228223959604899
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,768,0.15436800320943195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,12288,512,0.19710720380147298
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,12288,512,0.12665066719055176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,512,0.16190826098124186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,512,0.12531519730885823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,12288,256,0.14148586591084797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,12288,3584,0.5279242515563964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,256,0.10174506505330402
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,256,0.0993397315343221
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,12288,128,0.11909866333007812
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,12288,2560,0.388268788655599
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,128,0.08701653480529785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,128,0.0871392011642456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,12288,64,0.1187007983525594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,64,0.07940053145090739
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,64,0.09219733079274496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,12288,32,0.11894293626149495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,32,0.0785973310470581
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,32,0.09460373719533285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,10240,65536,24.804288736979167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,12288,4096,0.6173674901326497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,65536,15.953117879231772
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,65536,15.156959025065103
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,12288,3072,0.4584010759989421
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,12288,768,0.1589471975962321
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,10240,16384,4.982439676920572
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,16384,4.019896443684896
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,16384,3.455754597981771
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,10240,12288,3.712968444824219
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,12288,3.0829259236653646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,12288,2.0104981740315755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,10240,10240,3.099779256184896
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,12288,2048,0.32263358434041345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,10240,2.4365460713704428
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,10240,1.700708262125651
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,10240,8192,2.378061930338542
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,12288,1024,0.19017066955566406
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,8192,1.9913536071777345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,8192,1.2797344207763672
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,10240,7168,2.09006716410319
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,12288,128,0.08727786540985108
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,7168,1.6555573781331379
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,7168,1.0711594899495442
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,10240,6144,1.771178690592448
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,12288,256,0.09458346366882324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,6144,1.3176138559977213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,6144,0.9509834925333658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,10240,5120,1.467352549235026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,10240,65536,8.382637023925781
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,5120,1.1561247507731118
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,5120,0.7845759709676107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,10240,4096,1.1996170043945313
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,10240,16384,2.0960458119710283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,4096,0.928978157043457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,4096,0.6098954518636067
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,10240,3584,1.046239980061849
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,10240,12288,1.5461738586425782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,3584,0.8239018758138021
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,3584,0.5198741277058919
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,10240,3072,0.8910847981770834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,10240,10240,1.302239990234375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,3072,0.7043413162231446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,3072,0.4419359842936198
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,10240,2560,0.7470752080281575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,10240,8192,1.0015157063802085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,2560,0.5883967717488606
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,2560,0.36352640787760415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,10240,2048,0.5998442967732747
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,10240,7168,0.8585077285766601
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,2048,0.4599733352661133
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,2048,0.2881663958231608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,10240,1536,0.4553962707519531
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,10240,6144,0.7443402608235676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,1536,0.3944202740987142
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,1536,0.22336959838867188
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,10240,1024,0.31603412628173827
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,10240,5120,0.6177600224812825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,1024,0.2551232020060221
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,1024,0.16501973470052084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,10240,768,0.2451178709665934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,10240,3584,0.448525873819987
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,768,0.19286079406738282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,768,0.13207039833068848
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,10240,512,0.17356692949930827
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,10240,2560,0.3323946634928385
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,512,0.14181653658548993
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,512,0.10628480116526287
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,10240,256,0.12687466939290365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,10240,4096,0.5011391957600911
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,256,0.0979594628016154
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,256,0.08502933184305826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,10240,128,0.10885866483052571
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,10240,1536,0.21385812759399414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,128,0.07501973311106364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,128,0.075818665822347
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,10240,64,0.10516053040822346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,64,0.06890347003936767
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,64,0.07960000038146972
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,10240,32,0.10464213689168293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,32,0.06789866288503012
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,32,0.08037119706471761
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,10240,3072,0.3969663937886556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,8192,65536,17.485308837890624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,65536,13.04808349609375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,65536,8.761829630533855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,8192,16384,4.189312998453776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,10240,2048,0.27288745244344076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,16384,2.983043162027995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,16384,1.9928746541341145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,8192,12288,3.1973503112792967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,10240,768,0.13404906590779622
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,12288,2.253318277994792
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,12288,1.4786997477213542
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,8192,10240,2.6679242451985674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,10240,1024,0.16202239990234374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,10240,1.9230677286783855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,10240,1.2586090087890625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,8192,8192,2.144766998291016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,10240,256,0.07997439702351888
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,8192,1.5638591766357421
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,8192,0.9888533274332681
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,8192,7168,1.8751360575358074
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,10240,128,0.07622186342875162
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,7168,1.3000148773193358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,7168,0.8652533213297525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,8192,6144,1.6098431905110675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,10240,512,0.10721813042958576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,6144,1.1027732849121095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,6144,0.730725351969401
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,8192,5120,1.3167349497477212
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,8192,65536,7.022714742024739
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,5120,0.9637066523234049
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,5120,0.621671485900879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,8192,4096,1.0684405008951823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,8192,16384,1.6217226664225262
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,4096,0.7337034861246745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,4096,0.4623231887817383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,8192,3584,0.9126976013183594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,8192,12288,1.206445821126302
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,3584,0.6486687978108724
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,3584,0.39584213892618814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,8192,3072,0.7771658579508464
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,8192,10240,1.013307762145996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,3072,0.5500234603881836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,3072,0.3327701250712077
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,8192,2560,0.6495157241821289
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,8192,8192,0.8122869491577148
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,2560,0.46856212615966797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,2560,0.27011518478393554
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,8192,2048,0.5336010615030925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,8192,7168,0.7127754847208659
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,2048,0.37981974283854164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,2048,0.20726186434427896
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,8192,1536,0.39559040069580076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,8192,6144,0.6059722900390625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,1536,0.2862090746561686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,1536,0.16195732752482098
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,8192,1024,0.27866878509521487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,8192,5120,0.5105632146199544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,1024,0.19437440236409503
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,1024,0.11860693295796712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,8192,768,0.2118783950805664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,8192,3584,0.36060587565104163
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,768,0.15037546157836915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,768,0.10044159889221191
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,8192,512,0.153438933690389
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,8192,512,0.0860362688700358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,512,0.10941119988759358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,512,0.08135039806365967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,8192,256,0.11046720345815023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,8192,4096,0.4239797274271647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,256,0.06833279927571614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,256,0.06824639638264975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,8192,128,0.09094399611155192
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,8192,3072,0.3163893381754557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,128,0.059330133597056064
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,128,0.061193601290384925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,8192,64,0.08695893287658692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,64,0.05407679875691732
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,64,0.06138026714324951
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,8192,32,0.08771519660949707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,32,0.05393280188242594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,32,0.06379839976628622
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,7168,65536,16.378697713216148
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,8192,2560,0.26744852066040037
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,65536,10.597894287109375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,65536,8.439027404785156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,7168,16384,3.9419446309407555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,7168,16384,1.5126613616943358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,16384,2.809552001953125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,16384,1.7274410247802734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,7168,12288,2.9868202209472656
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,8192,2048,0.22208959261576333
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,8192,1536,0.17720534006754557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,12288,2.00740483601888
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,12288,1.286948267618815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,7168,10240,2.518353017171224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,10240,1.6031829833984375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,10240,1.060696538289388
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,7168,8192,2.009234110514323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,7168,8192,0.717742919921875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,8192,1.352669906616211
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,8192,768,0.11009813149770101
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,8192,0.8636277516682943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,7168,7168,1.734182357788086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,7168,1.1178816477457683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,7168,0.7921013514200846
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,7168,6144,1.5075242360432943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,8192,256,0.06435840129852295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,6144,0.9386688232421875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,6144,0.6353557586669922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,7168,5120,1.2581109364827474
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,8192,128,0.06056640148162842
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,5120,0.7965845108032227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,5120,0.5267008145650227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,7168,4096,1.0059605280558268
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,8192,1024,0.13119253317515056
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,4096,0.6236469268798828
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,4096,0.42196372350056965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,7168,4096,0.37864319483439124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,7168,3584,0.8770634969075521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,3584,0.5458154678344727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,3584,0.3685077349344889
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,7168,3072,0.7588597615559896
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,7168,65536,6.26230723063151
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,3072,0.4823818524678548
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,3072,0.30120000839233396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,7168,2560,0.6242762883504231
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,7168,12288,1.0572906494140626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,2560,0.4112277348836263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,2560,0.24834772745768227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,7168,2048,0.49871040980021164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,7168,2048,0.19867092768351238
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,2048,0.3152607917785645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,2048,0.18993172645568848
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,7168,1536,0.38411200841267906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,7168,10240,0.9195178349812826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,1536,0.25776960055033366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,1536,0.1527242660522461
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,7168,1024,0.26920318603515625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,7168,7168,0.6290943781534831
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,1024,0.17255147298177081
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,1024,0.10864960352579753
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,7168,768,0.2031925360361735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,7168,6144,0.5420501073201497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,768,0.13745066324869792
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,768,0.09800106684366862
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,7168,512,0.1528330643971761
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,7168,3584,0.3342432022094727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,512,0.10015573501586914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,512,0.07871253490447998
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,7168,256,0.10686506430308025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,7168,5120,0.44814399083455403
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,256,0.06465386549631755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,256,0.06665066480636597
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,7168,128,0.08914453188578288
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,7168,3072,0.2799477259318034
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,128,0.05252480109532674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,128,0.055335466066996256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,7168,64,0.08367679913838705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,64,0.04769706726074219
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,64,0.058620798587799075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,7168,32,0.08005332946777344
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,32,0.04768639802932739
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,32,0.05716906785964966
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,6144,65536,15.28485310872396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,7168,2560,0.2415583928426107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,65536,9.49951883951823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,65536,6.215180969238281
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,6144,16384,3.882401021321615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,7168,1536,0.1578506628672282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,16384,2.3908587137858075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,16384,1.4485215504964193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,6144,12288,2.8362111409505206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,7168,128,0.05394453207651774
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,12288,1.60502192179362
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,12288,1.2233951568603516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,6144,10240,2.388025665283203
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,7168,256,0.057650132973988855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,10240,1.4803306579589843
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,10240,0.9559359868367514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,6144,8192,1.8905269622802734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,7168,512,0.07779733339945474
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,8192,1.103664016723633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,8192,0.7278133392333984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,6144,7168,1.6507775624593097
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,7168,768,0.0986741304397583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,7168,0.9783466974894205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,7168,0.6531797409057617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,6144,6144,1.4265300750732421
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,7168,1024,0.11850666999816895
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,6144,0.8278271993001303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,6144,0.5579818725585938
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,6144,5120,1.1990100860595703
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,6144,65536,5.2583063761393225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,5120,0.707481575012207
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,5120,0.4651530583699544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,6144,4096,0.9600565592447916
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,6144,16384,1.283685302734375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,4096,0.595303471883138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,4096,0.3572650591532389
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,6144,3584,0.8506624221801757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,6144,12288,0.9309439977010092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,3584,0.49081172943115237
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,3584,0.3109504063924154
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,6144,3072,0.7145109176635742
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,6144,10240,0.7668448130289713
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,3072,0.4111071904500325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,3072,0.2644543965657552
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,6144,2560,0.598363749186198
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,6144,8192,0.6437546412150066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,2560,0.35792318979899085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,2560,0.21872320175170898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,6144,2048,0.47742185592651365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,6144,7168,0.5483221054077149
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,2048,0.29519147872924806
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,2048,0.17089279492696124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,6144,1536,0.3724010785420736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,6144,1536,0.1388319969177246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,1536,0.21519145965576172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,1536,0.1221845308939616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,6144,1024,0.2520991961161295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,6144,6144,0.46932051976521805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,1024,0.14932907422383626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,1024,0.09541227022806803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,6144,768,0.18976960182189942
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,6144,5120,0.3967402776082357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,768,0.11590720017751058
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,768,0.08055253028869629
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,6144,512,0.13549119631449383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,6144,4096,0.31719786326090493
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,512,0.08447679678599039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,512,0.06644266843795776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,6144,256,0.09713280200958252
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,6144,3584,0.28425280253092444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,256,0.05286933183670044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,256,0.052970667680104576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,6144,128,0.0784554640452067
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,6144,2560,0.2081450621287028
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,128,0.04589440027872722
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,128,0.04783360163370769
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,6144,64,0.07442773183186849
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,64,0.04137173493703206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,64,0.04888319969177246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,6144,32,0.07165013154347738
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,32,0.04225279887517293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,32,0.04889920155207316
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,5120,65536,14.602226765950522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,6144,3072,0.2429642677307129
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,65536,7.836151631673177
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,65536,5.653086853027344
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,5120,16384,3.5381823221842446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,6144,768,0.08597119649251303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,16384,1.8464522043863933
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,16384,1.297540283203125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,6144,2048,0.17693440119425458
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,5120,12288,2.685716247558594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,12288,1.4462079366048177
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,12288,0.9636074701944987
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,6144,512,0.06881600220998128
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,5120,10240,2.2603274027506512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,10240,1.1729515075683594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,10240,0.8133792241414388
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,5120,8192,1.826259231567383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,6144,1024,0.10339199701944987
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,8192,0.9105386734008789
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,8192,0.6360544204711914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,5120,7168,1.5519797007242837
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,6144,256,0.051319468021392825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,7168,0.7837535858154296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,7168,0.5612202962239583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,5120,6144,1.351149876912435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,6144,128,0.0482367992401123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,6144,0.6742442448933919
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,6144,0.4805973370869954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,5120,5120,1.1194229125976562
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,5120,65536,4.606403096516927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,5120,0.5630666732788085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,5120,0.3851231892903646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,5120,4096,0.9008117039998373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,5120,16384,1.0464277267456055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,4096,0.4589856147766113
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,4096,0.3050634702046712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,5120,3584,0.7848288218180339
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,5120,12288,0.8170719782511393
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,3584,0.40815146764119464
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,5120,10240,0.6697418848673503
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,3584,0.2662559986114502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,5120,3072,0.6731157302856445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,3072,0.34641599655151367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,3072,0.22783892949422202
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,5120,2560,0.5681664148966472
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,5120,8192,0.5332703908284505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,2560,0.29005120595296224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,2560,0.19010880788167317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,5120,2048,0.45382080078125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,5120,7168,0.4704799969991048
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,2048,0.23644372622172036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,2048,0.14579946200052898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,5120,1536,0.34861866633097327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,5120,6144,0.4004554748535156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,1536,0.17841173807779948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,1536,0.11045973300933838
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,5120,1024,0.2400767962137858
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,5120,5120,0.3402186711629232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,1024,0.12551999886830648
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,1024,0.08127146561940511
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,5120,768,0.17718399365743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,5120,3584,0.240830930074056
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,768,0.09767786661783853
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,768,0.0698037306467692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,5120,512,0.12766506671905517
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,5120,4096,0.27198721567789713
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,512,0.07292693456013997
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,512,0.05820586681365967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,5120,256,0.08930453459421793
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,5120,2560,0.1815872033437093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,256,0.0494762659072876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,256,0.04718826611836751
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,5120,128,0.0735157330830892
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,5120,1536,0.12029226620992024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,128,0.03959146738052368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,128,0.04219626585642497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,5120,64,0.0686346689860026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,64,0.03519893487294515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,64,0.04348586797714234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,5120,32,0.0651797334353129
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,32,0.03580480019251506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,32,0.04372373421986898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,5120,3072,0.2159477392832438
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,4096,65536,13.506520589192709
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,65536,6.399287414550781
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,65536,4.4498743693033855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,5120,2048,0.150545072555542
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,4096,16384,3.329229990641276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,16384,1.5980820973714194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,16384,1.0470325469970703
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,4096,12288,2.5268458048502604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,5120,768,0.0733407974243164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,12288,1.0841451009114584
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,12288,0.8086144129435221
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,4096,10240,2.118429819742839
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,5120,1024,0.08950613339742025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,10240,0.8816917419433594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,10240,0.6758517583211263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,4096,8192,1.6873152414957684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,5120,128,0.04179946581522624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,8192,0.7248149236043294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,8192,0.5093802769978841
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,4096,7168,1.4610869089762368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,5120,256,0.04383680025736491
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,7168,0.6117183685302734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,7168,0.4552735964457194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,4096,6144,1.2655818939208985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,4096,7168,0.3902453422546387
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,6144,0.5368480046590169
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,6144,0.39245227177937825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,4096,5120,1.0517056147257486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,5120,512,0.058950400352478026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,5120,0.440121587117513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,5120,0.3135690689086914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,4096,4096,0.846011734008789
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,4096,65536,3.7390121459960937
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,4096,0.37098986307779946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,4096,0.2526111920674642
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,4096,3584,0.7402069091796875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,4096,16384,0.8713589350382487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,3584,0.3301930745442708
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,3584,0.21915413538614908
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,4096,3072,0.6363882700602214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,4096,12288,0.6586687723795573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,3072,0.2741152127583822
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,3072,0.19063572883605956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,4096,2560,0.5338890711466472
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,4096,10240,0.5517759958902995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,2560,0.2346677303314209
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,2560,0.15488959948221842
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,4096,2048,0.43259518941243486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,4096,8192,0.43898560206095383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,2048,0.1842591921488444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,2048,0.11988373597462972
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,4096,1536,0.3301386515299479
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,4096,6144,0.32979094187418617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,1536,0.1403264045715332
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,1536,0.08919146855672201
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,4096,1024,0.22352426846822104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,4096,5120,0.28111785252889
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,1024,0.1002079963684082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,1024,0.0679370641708374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,4096,768,0.16440000534057617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,4096,4096,0.2256330649058024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,768,0.07759466965993246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,768,0.056866133213043214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,4096,512,0.11672746340433757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,4096,512,0.04949013392130534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,512,0.05745813449223837
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,512,0.046817068258921304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,4096,256,0.08069653511047363
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,4096,3584,0.2061002731323242
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,256,0.03773119846979777
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,256,0.03908586502075195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,4096,128,0.06372373501459758
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,4096,2560,0.14991893768310546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,128,0.03236906727155049
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,128,0.034626134236653644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,4096,64,0.059913599491119386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,64,0.029397332668304445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,64,0.03554133176803589
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,4096,32,0.05823253393173218
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,32,0.030044800043106078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,32,0.035556264718373615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3584,65536,12.94018758138021
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,4096,3072,0.17399999300638835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,65536,5.607332356770834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,65536,3.9064832051595055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3584,16384,3.2227327982584635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,4096,1536,0.09907840092976888
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,16384,1.2966037750244142
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,16384,0.9488224029541016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3584,12288,2.4561846415201822
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,4096,2048,0.1246613343556722
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,12288,0.9258080164591471
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,12288,0.7219445546468098
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3584,10240,2.061578623453776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,4096,256,0.036549333731333414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,10240,0.759555180867513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,10240,0.6068682352701823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3584,8192,1.640772247314453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,4096,768,0.06198826630910238
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,8192,0.6234826405843099
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,8192,0.4714005470275879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3584,7168,1.423925272623698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,4096,128,0.03402773141860962
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,7168,0.5564597447713215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,4096,1024,0.07404586474100748
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,7168,0.41370452245076494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3584,6144,1.2229493459065757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,6144,0.46151892344156903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,6144,0.35803839365641277
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3584,5120,1.01888427734375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3584,65536,3.3283968607584633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,5120,0.43119147618611653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,5120,0.2950613339742025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3584,4096,0.8319434483846029
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3584,16384,0.7825632095336914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,4096,0.31969385147094725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,4096,0.23643840154012047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3584,3584,0.7227402369181315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3584,12288,0.5888533274332682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,3584,0.27603626251220703
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,3584,0.20688959757486977
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3584,3072,0.6208181381225586
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3584,10240,0.49642025629679365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,3072,0.24165546099344887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,3072,0.17694719632466632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3584,2560,0.5218101183573405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3584,8192,0.39437974294026695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,2560,0.19957440694173176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,2560,0.1469173272450765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3584,2048,0.4223477363586426
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3584,7168,0.35153385798136394
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,2048,0.1642805258433024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,2048,0.11239573160807292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3584,1536,0.32107839584350584
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3584,5120,0.2507712046305338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,1536,0.12378133138020833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,1536,0.08392746448516845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3584,1024,0.21954026222229003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3584,6144,0.30722665786743164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,1024,0.0897877295811971
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,1024,0.060959998766581217
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3584,768,0.16113492647806804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3584,3584,0.18259305953979493
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,768,0.06969707012176514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,768,0.053641601403554284
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3584,512,0.11278399626413982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3584,4096,0.2050538698832194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,512,0.0512714664141337
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,512,0.043569068113962814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3584,256,0.07821333408355713
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3584,2560,0.13664746284484863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,256,0.03413653373718262
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,256,0.03634986480077108
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3584,128,0.062363731861114505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3584,3072,0.15835626920064289
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,128,0.029102933406829835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,128,0.03287573258082072
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3584,64,0.05826239983240763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,64,0.026983465751012164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,64,0.03310826619466146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3584,32,0.05532266696294149
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,32,0.026697599887847902
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,32,0.03300053278605143
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3072,65536,12.314317830403645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3584,2048,0.11305812994639079
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,65536,4.274116516113281
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,65536,3.318421427408854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3072,16384,3.115061442057292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3584,1536,0.09085973103841147
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,16384,1.088332748413086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,16384,0.834562110900879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3072,12288,2.3722078959147135
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3584,768,0.05525546471277872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,12288,0.8072757085164388
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,12288,0.6634464263916016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3072,10240,1.9995423634847005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3584,1024,0.0671135981877645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,10240,0.6500991821289063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,10240,0.5370474497477213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3072,8192,1.578009541829427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3584,256,0.033048532406489056
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,8192,0.5281930605570475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,8192,0.41884374618530273
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3072,7168,1.367427190144857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3584,512,0.042903467019399004
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,7168,0.4612746556599935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,7168,0.36758400599161783
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3072,6144,1.1912960052490233
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3584,128,0.02954026659329732
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,6144,0.40095678965250653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,6144,0.3200245221455892
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3072,5120,0.9869344075520834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3072,65536,2.8463371276855467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,5120,0.3363391876220703
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,5120,0.2683605194091797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3072,4096,0.8084469477335612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3072,16384,0.689962641398112
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,4096,0.26656853357950844
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,4096,0.21166826883951823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3072,3584,0.6980298360188801
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3072,12288,0.52609281539917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,3584,0.23917013804117837
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,3584,0.18539412816365558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3072,3072,0.6010186513264973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3072,10240,0.44778025945027666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,3072,0.2108469327290853
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,3072,0.15696533521016437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3072,2560,0.5184821446736654
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3072,8192,0.3484277407328288
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,2560,0.1703349272410075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,2560,0.13067200183868408
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3072,2048,0.4083999951680501
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3072,7168,0.30943145751953127
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,2048,0.13991146087646483
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,2048,0.09865813255310059
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3072,1536,0.31322345733642576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3072,1536,0.08114559650421142
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,1536,0.1078762690226237
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,1536,0.07375146547953287
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3072,1024,0.2129653294881185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3072,6144,0.2653834660847982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,1024,0.07591573397318521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,1024,0.0542794664700826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3072,768,0.15560213724772137
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3072,5120,0.22535360654195152
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,768,0.060568531354268394
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,768,0.04840426842371623
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3072,512,0.1086517333984375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3072,3584,0.16190080642700194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,512,0.04487893184026082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,512,0.03841600020726522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3072,256,0.07563520272572835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3072,4096,0.1818720022837321
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,256,0.029994666576385498
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,256,0.03249599933624268
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3072,128,0.059178666273752845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3072,2560,0.1235584020614624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,128,0.02588160037994385
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,128,0.02916266719500224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3072,64,0.05506453514099121
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,64,0.023475199937820435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,64,0.02879253427187602
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,3072,32,0.051634132862091064
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,32,0.024834134181340537
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,32,0.0291157325108846
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2560,65536,11.916629028320312
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3072,3072,0.14122452735900878
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,65536,3.8527712504069007
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,65536,2.903429412841797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2560,16384,3.0243914286295572
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3072,512,0.03869333267211914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,16384,0.8766645431518555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,16384,0.7328458786010742
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2560,12288,2.278185526529948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3072,2048,0.10044586658477783
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,12288,0.6499818801879883
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,12288,0.561846415201823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2560,10240,1.9172373453776044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3072,768,0.05021013418833414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,10240,0.5461343765258789
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,10240,0.47147626876831056
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2560,8192,1.543502934773763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3072,1024,0.06052800019582113
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,8192,0.4400469462076823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,8192,0.3689375877380371
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2560,7168,1.3265717824300132
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3072,256,0.029549866914749146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,7168,0.381547737121582
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,7168,0.3214837392171224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2560,6144,1.1489706675211588
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3072,128,0.026261333624521894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,6144,0.32995627721150717
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,6144,0.2801055908203125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2560,5120,0.9570581436157226
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2560,65536,2.5725674947102863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,5120,0.3292863845825195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,5120,0.23252479235331217
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2560,4096,0.7812992095947265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2560,16384,0.6109642664591471
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,4096,0.22745280265808104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,4096,0.18589013417561848
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2560,3584,0.6815957387288412
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2560,10240,0.37839679718017577
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,3584,0.19656532605489094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,3584,0.16211519241333008
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2560,3072,0.5871605555216471
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2560,12288,0.4578890800476074
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,3072,0.1728490670522054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,3072,0.14034026463826496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2560,2560,0.4925855954488118
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2560,8192,0.3068810780843099
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,2560,0.144870392481486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,2560,0.11770773728688558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2560,2048,0.39905920028686526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2560,7168,0.2709909439086914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,2048,0.11827840010325115
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,2048,0.08792959849039714
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2560,1536,0.30613654454549155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2560,6144,0.23207146326700845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,1536,0.08981760342915854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,1536,0.06579946676890056
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2560,1024,0.2068117300669352
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2560,5120,0.19622613588968912
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,1024,0.06416853268941244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,1024,0.04781653483708699
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2560,768,0.15104640324910482
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2560,3584,0.14306666056315104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,768,0.05189973513285319
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,768,0.042713598410288496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2560,512,0.10522987047831218
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2560,4096,0.1639413356781006
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,512,0.04002026716868083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,512,0.03441386620203654
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2560,256,0.0716426690419515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2560,2560,0.10650026798248291
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,256,0.02805440028508504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,256,0.028174932797749835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2560,128,0.05521920124689737
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2560,3072,0.12289280096689861
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,128,0.022061866521835328
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,128,0.0254037340482076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2560,64,0.04979626735051473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,64,0.019594667355219524
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,64,0.025621332724889118
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2560,32,0.048128000895182294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,32,0.020579200983047486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,32,0.02613760034243266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2048,65536,11.463155110677082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2560,2048,0.08892373243967691
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,65536,3.0192639668782553
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,65536,2.5251200358072916
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2048,16384,2.9222015380859374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2560,1536,0.0720362663269043
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,16384,0.7048896153767903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,16384,0.6233130772908528
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2048,12288,2.200140889485677
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2560,768,0.04342293341954549
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,12288,0.5248671849568685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,12288,0.4820831934611003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2048,10240,1.8484383900960286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2560,1024,0.0529749313990275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,10240,0.43657493591308594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,10240,0.4066133181254069
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2048,8192,1.4886517842610678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2560,256,0.02592960000038147
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,8192,0.36397012074788415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,8192,0.3167178789774577
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2048,7168,1.2801952362060547
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2560,512,0.03415253162384033
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,7168,0.31492268244425453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,7168,0.27844692866007487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2048,6144,1.1086762746175132
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2560,128,0.022923733790715536
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,6144,0.2704352060953776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,6144,0.24139413833618165
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2048,5120,0.9297130584716797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2048,65536,2.0837664286295574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,5120,0.22675627072652182
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,5120,0.20021440188090006
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2048,4096,0.7485951741536458
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2048,16384,0.5128309249877929
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,4096,0.18342612584431967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,4096,0.16079893112182617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2048,3584,0.6590911865234375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2048,12288,0.38680960337320963
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,3584,0.16108800570170084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,3584,0.14277226130167645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2048,3072,0.5687349319458008
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2048,10240,0.3222304026285807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,3072,0.1382047971089681
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,3072,0.1216597318649292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2048,2560,0.47810773849487304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2048,8192,0.25942932764689125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,2560,0.11651519934336345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,2560,0.10267732938130696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2048,2048,0.3875199953715006
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2048,7168,0.2343445301055908
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,2048,0.09798719882965087
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,2048,0.07764906883239746
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2048,1536,0.2977781295776367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2048,6144,0.19914986292521158
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,1536,0.07443412939707438
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,1536,0.057241598765055336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2048,1024,0.2010752042134603
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2048,5120,0.16752427419026691
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,1024,0.05246506532033285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,1024,0.041484800974527995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2048,768,0.14495147069295247
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2048,4096,0.13552319208780925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,768,0.04197546641031901
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,768,0.03713279962539673
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2048,512,0.09978880087534586
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2048,3584,0.12123306592305501
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,512,0.03213866750399272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,512,0.028550400336583452
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2048,256,0.06702293554941813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2048,2560,0.09041813214619955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,256,0.022332799434661866
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,256,0.02348053256670634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2048,128,0.05149333477020264
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2048,3072,0.10537280241648357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,128,0.017358932892481485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,128,0.021132800976435342
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2048,64,0.04581760168075562
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,64,0.015255467096964518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,64,0.0219050665696462
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,2048,32,0.043859199682871504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,32,0.015454933047294617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,32,0.02203306754430135
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1536,65536,11.051732381184895
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2048,768,0.03710720141728719
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,65536,2.2076927185058595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,65536,2.1478645324707033
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1536,16384,2.820357259114583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2048,1536,0.06223146518071493
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,16384,0.534069315592448
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,16384,0.5333845138549804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1536,12288,2.115233103434245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2048,2048,0.07530453205108642
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,12288,0.3879925409952799
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,12288,0.40742292404174807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1536,10240,1.7875284830729168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2048,1024,0.04618346691131592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,10240,0.3317184130350749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,10240,0.34005759557088217
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1536,8192,1.4345600128173828
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2048,256,0.020965333779652914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,8192,0.26395840644836427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,8192,0.27002773284912107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1536,7168,1.242202631632487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2048,512,0.028619732459386187
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,7168,0.2296895980834961
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,7168,0.23468267122904457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1536,6144,1.0685279846191407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2048,128,0.018326399723688762
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,6144,0.19909440676371257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,6144,0.20588906606038412
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1536,5120,0.899291737874349
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1536,65536,1.7050933837890625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,5120,0.17368213335673016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,5120,0.16955520311991373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1536,4096,0.7207349141438801
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1536,16384,0.425820795694987
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,4096,0.13743573824564617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,4096,0.13670399983723958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1536,3584,0.6359445571899414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1536,12288,0.3259722709655762
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,3584,0.12297386328379314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,3584,0.12178133328755696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1536,3072,0.5505205154418945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1536,10240,0.2671797434488932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,3072,0.10317333539326985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,3072,0.10498773256937663
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1536,2560,0.4647040049235026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1536,8192,0.21786880493164062
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,2560,0.08946986993153891
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,2560,0.08782933553059896
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1536,2048,0.37715412775675455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1536,7168,0.1909877300262451
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,2048,0.07244586944580078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,2048,0.06651626825332642
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1536,1536,0.28873812357584633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1536,6144,0.16435093879699708
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,1536,0.05573013226191202
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,1536,0.05018133322397868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1536,1024,0.19489386876424153
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1536,5120,0.13961812655131023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,1024,0.04018986622492472
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,1024,0.03556053241093953
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1536,768,0.13882880210876464
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1536,4096,0.1143839995066325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,768,0.03324693242708842
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,768,0.03179946740468343
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1536,512,0.09534293015797933
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1536,3584,0.1015658696492513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,512,0.025333333015441894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,512,0.02342933416366577
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1536,256,0.06346666812896729
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1536,2560,0.07689812978108725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,256,0.018011732896169027
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,256,0.018926932414372762
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1536,128,0.04751466512680054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1536,3072,0.09051093260447184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,128,0.013613866766293845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,128,0.01693013310432434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1536,64,0.040957868099212646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,64,0.011916800340016683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,64,0.017562667528788246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1536,32,0.038186665376027426
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,32,0.011991467078526814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,32,0.01805973251660665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1024,65536,10.631589762369792
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1536,2048,0.06475520133972168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,65536,1.4914250691731772
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,65536,1.7326165517171224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1024,16384,2.7146527608235678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1536,1536,0.05235306819279989
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,16384,0.35344638824462893
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,16384,0.43227411905924484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1536,768,0.03158826629320781
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1024,12288,2.045675786336263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,12288,0.2670090675354004
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,12288,0.32938451766967775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1024,10240,1.722826639811198
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1536,1024,0.03846720059712728
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,10240,0.22052480379740397
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,10240,0.2748319943745931
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1024,8192,1.3878421783447266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1536,256,0.016923733552296958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,8192,0.18058773676554363
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1536,128,0.013755733768145243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,8192,0.22171413103739418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1024,7168,1.199948755900065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,7168,0.1564095973968506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,7168,0.1926688035329183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1024,6144,1.0333621342976889
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1536,512,0.024243199825286867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,6144,0.13683627446492513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,6144,0.16782506306966144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1024,5120,0.8671509424845377
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1024,65536,1.3906773885091146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,5120,0.11982826391855876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,5120,0.1409663995107015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1024,4096,0.7016597112019857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1024,16384,0.3378442764282227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,4096,0.09315093358357748
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,4096,0.11292266845703125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1024,3584,0.6171317418416341
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1024,12288,0.253548796971639
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,3584,0.0811626672744751
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,3584,0.10157653490702312
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1024,10240,0.21426879564921059
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1024,3072,0.5340480168660482
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,3072,0.07044586340586344
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,3072,0.08706346352895102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1024,2560,0.4499658584594727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1024,8192,0.1715722719828288
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,2560,0.05910506645838419
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,2560,0.07199146747589111
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1024,2048,0.36482346852620445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1024,7168,0.15316799481709797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,2048,0.049294932683308916
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,2048,0.05597013235092163
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1024,1536,0.2799306551615397
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1024,6144,0.1318933327992757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,1536,0.03778453270594279
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,1536,0.04103786547978719
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1024,1024,0.1891285260518392
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1024,5120,0.11167999903361003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,1024,0.02696533401807149
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,1024,0.02911146680514018
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1024,768,0.13337599436442057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1024,4096,0.09351680278778077
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,768,0.022268799940745036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,768,0.026316799720128375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1024,512,0.08895573616027833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1024,3584,0.08202239672342936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,512,0.016544000307718913
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,512,0.01842026710510254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1024,256,0.0584330677986145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1024,2560,0.06246933142344156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,256,0.011686399579048157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,256,0.015915733575820924
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1024,128,0.04042773246765137
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1024,3072,0.07260479927062988
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,128,0.009272533655166625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,128,0.01409386694431305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1024,64,0.03593493302663167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,64,0.00817493349313736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,64,0.014402133226394654
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,1024,32,0.033958399295806886
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,32,0.008469333251317341
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,32,0.014660267035166421
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,768,65536,10.431195068359376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1024,2048,0.052781867980957034
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,65536,1.0629589080810546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,65536,1.5269643147786458
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,768,16384,2.6653483072916666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1024,768,0.024476800362269083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,16384,0.2726890563964844
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,16384,0.3822666803995768
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,768,12288,2.0061055501302083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1024,1536,0.04255359967549642
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,12288,0.2058794657389323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,12288,0.2939818700154623
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,768,10240,1.6838751475016278
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1024,1024,0.03094293276468913
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,10240,0.1698581377665202
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,10240,0.24687360127766927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,768,8192,1.3627285003662108
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1024,128,0.011693867047627766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,8192,0.1368234634399414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,8192,0.19532160758972167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,768,7168,1.1767765045166017
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1024,256,0.013502933581670127
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,7168,0.1218997319539388
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,7168,0.17100586891174316
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,768,6144,1.0151573181152345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1024,512,0.018745599190394084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,6144,0.10663893222808837
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,6144,0.1488810698191325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,768,5120,0.8509717305501303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,768,65536,1.1664107004801432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,5120,0.0894431988398234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,5120,0.12452479998270671
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,768,4096,0.687548828125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,768,16384,0.289686393737793
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,4096,0.07025173505147299
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,4096,0.10141119956970215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,768,3584,0.6044298807779949
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,768,12288,0.22179733912150063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,3584,0.06379520098368327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,3584,0.08987092971801758
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,768,3072,0.5235775947570801
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,768,10240,0.18519679705301922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,3072,0.054907735188802084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,3072,0.07660586833953857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,768,2560,0.44103358586629227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,768,8192,0.14865919748942058
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,2560,0.045764267444610596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,2560,0.06530346473058066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,768,2048,0.3584896087646484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,768,7168,0.13194026947021484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,2048,0.03739519913991292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,2048,0.05059306621551514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,768,1536,0.2757781346638998
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,768,6144,0.1143946647644043
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,1536,0.029084799687067668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,1536,0.0366922656695048
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,768,1024,0.18454079627990722
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,768,5120,0.09827199776967367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,1024,0.02109973430633545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,1024,0.02502506573994954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,768,768,0.13214293320973713
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,768,3584,0.07121280034383139
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,768,0.01680319905281067
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,768,0.021947733561197915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,768,512,0.08834880193074544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,768,4096,0.07973972956339517
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,512,0.01317759950955709
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,512,0.015826132893562318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,768,256,0.05603839953740438
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,768,2560,0.05422506729761759
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,256,0.009248000383377076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,256,0.013528533776601157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,768,128,0.038378667831420896
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,768,3072,0.0626261313756307
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,128,0.0075573335091273
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,128,0.012157866358757019
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,768,64,0.03368853330612183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,64,0.006493866443634033
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,64,0.012531200051307678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,768,32,0.03155840039253235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,32,0.006679466863473256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,32,0.0126720001300176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,512,65536,10.233980305989583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,768,2048,0.045764267444610596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,65536,0.7087530771891276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,65536,1.328167470296224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,512,16384,2.614435323079427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,768,768,0.021169066429138184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,16384,0.18552640279134114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,16384,0.33393494288126624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,512,12288,1.9707061767578125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,768,1536,0.03688960075378418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,12288,0.1356895923614502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,12288,0.2572906653086344
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,512,10240,1.6563167572021484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,768,1024,0.026195200284322102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,10240,0.11609493096669514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,10240,0.2131253401438395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,512,8192,1.3359371185302735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,768,512,0.01618133286635081
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,8192,0.09176320234934489
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,8192,0.1694709300994873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,512,7168,1.1545684814453125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,768,256,0.012168533603350322
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,7168,0.08145706653594971
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,7168,0.14937920570373536
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,512,6144,0.9976544062296548
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,768,128,0.010233599940935771
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,6144,0.07114453315734863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,6144,0.12926613489786784
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,512,5120,0.8345973332722982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,512,65536,0.9919658660888672
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,5120,0.06166826486587525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,5120,0.10893759727478028
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,512,4096,0.6764661153157552
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,512,16384,0.25093013445536294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,4096,0.04745813210805257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,4096,0.08896746635437011
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,512,3584,0.5941237131754558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,512,12288,0.19232959747314454
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,3584,0.043680000305175784
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,3584,0.07797973155975342
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,512,3072,0.5133397420247395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,512,10240,0.16037333806355794
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,3072,0.03711893161137898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,3072,0.06899733543395996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,512,2560,0.4339743932088216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,512,8192,0.13000960350036622
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,2560,0.03186239997545878
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,2560,0.056850131352742514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,512,2048,0.3524383862813314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,512,7168,0.11497386296590169
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,2048,0.026184533039728803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,2048,0.04304639895757039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,512,1536,0.2699626604715983
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,512,6144,0.10034666856129963
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,1536,0.020382932821909585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,1536,0.03065066734949748
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,512,1024,0.1820042610168457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,512,5120,0.08485226631164551
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,1024,0.014522666732470194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,1024,0.019061332941055296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,512,768,0.12788266340891521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,512,4096,0.06981226603190104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,768,0.012286933263142903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,768,0.018761599063873292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,512,512,0.08510826428731283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,512,3584,0.06291626691818238
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,512,0.009807999928792317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,512,0.013078400492668152
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,512,256,0.05366506576538086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,512,2560,0.04739840030670166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,256,0.007485866546630859
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,256,0.011422933141390482
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,512,128,0.036909866333007815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,512,3072,0.05531946818033854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,128,0.006305066744486491
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,128,0.01067626674969991
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,512,64,0.03213546673456828
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,64,0.005514666438102722
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,64,0.01104960044225057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,512,32,0.029626667499542236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,32,0.005724800129731497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,32,0.011183999975522359
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,256,65536,10.022981770833333
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,512,2048,0.040148266156514484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,65536,0.4222741444905599
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,65536,1.130776596069336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,256,16384,2.5645856221516925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,512,1536,0.03186773260434468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,16384,0.1050101359685262
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,16384,0.28966185251871746
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,256,12288,1.9331925710042317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,512,768,0.017854932943979898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,12288,0.08270826339721679
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,12288,0.22027413050333658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,256,10240,1.6251519521077473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,512,1024,0.022317866484324135
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,10240,0.06908160050710042
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,10240,0.18584426244099933
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,256,8192,1.3147445678710938
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,512,256,0.011211733023325603
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,8192,0.05523093144098917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,8192,0.15026987393697103
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,512,128,0.00939520001411438
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,256,7168,1.135040028889974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,7168,0.04975039958953857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,7168,0.1323754628499349
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,256,6144,0.9787786483764649
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,512,512,0.01430506706237793
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,6144,0.044158931573232016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,6144,0.11535147031148274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,256,5120,0.81986452738444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,256,65536,0.8877685546875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,5120,0.0428330659866333
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,5120,0.09656746387481689
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,256,4096,0.6624533335367839
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,256,16384,0.22462612787882485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,4096,0.031402667363484696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,4096,0.07775039672851562
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,256,3584,0.5826346715291341
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,256,12288,0.1705685297648112
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,3584,0.028017065922419232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,3584,0.06811199982961019
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,256,3072,0.5044309298197429
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,256,10240,0.14360000292460123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,3072,0.024825600783030193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,3072,0.05852373441060384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,256,2560,0.4240757306416829
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,256,8192,0.11541439692179363
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,2560,0.02108479936917623
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,2560,0.04883946577707927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,256,2048,0.34445759455362956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,256,7168,0.10394986470540364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,2048,0.015593600273132325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,2048,0.03577066659927368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,256,1536,0.26494827270507815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,256,6144,0.08751893043518066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,1536,0.012854400277137756
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,1536,0.023524266481399537
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,256,1024,0.17789653142293294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,256,5120,0.07465173403422037
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,1024,0.009201066692670186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,1024,0.015204266707102457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,256,768,0.12576746940612793
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,256,4096,0.06050560077031454
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,768,0.008062933385372163
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,768,0.015602133671442666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,256,512,0.08353066444396973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,256,3584,0.05376960039138794
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,512,0.0065610667069753
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,512,0.01090666651725769
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,256,256,0.05220053195953369
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,256,2560,0.04025280078252157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,256,0.005411200225353241
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,256,0.009890133142471313
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,256,128,0.03508586486180623
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,256,3072,0.04707093238830566
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,128,0.004599466423193614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,128,0.009226666887601216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,256,64,0.030197334289550782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,64,0.004152533411979675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,64,0.00955839951833089
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,256,32,0.028398933013280232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,32,0.004392533500989278
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,32,0.009875200192133586
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,128,65536,10.028587849934896
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,256,2048,0.033763198057810466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,65536,0.41630080540974934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,65536,1.1196853637695312
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,128,16384,2.5524085998535155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,256,1536,0.026532266537348432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,16384,0.10962879657745361
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,16384,0.28476266860961913
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,128,12288,1.932512029012044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,256,768,0.015185067057609558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,12288,0.07459839979807535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,12288,0.21742399533589682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,128,10240,1.6241141001383466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,256,1024,0.018013866742451985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,10240,0.06365546782811483
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,10240,0.18289920488993328
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,128,8192,1.3116650899251303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,256,256,0.009544533491134644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,8192,0.05083306630452474
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,8192,0.1465354601542155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,128,7168,1.1322176615397135
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,256,512,0.01230400005976359
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,7168,0.045311999320983884
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,7168,0.1291551987330119
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,128,6144,0.9761621475219726
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,256,128,0.007699200014273326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,6144,0.04042666753133138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,6144,0.11151253382364909
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,128,5120,0.8172970453898112
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,128,65536,0.8147104263305665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,5120,0.03890346686045329
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,5120,0.09363839626312256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,128,4096,0.6620202382405599
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,128,16384,0.2111168066660563
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,4096,0.028431999683380126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,4096,0.07619626522064209
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,128,3584,0.582206916809082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,128,12288,0.16030720074971516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,3584,0.025226666529973345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,3584,0.06634666522343954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,128,3072,0.5018303871154786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,128,10240,0.13517759641011556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,3072,0.02226026654243469
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,3072,0.05766613483428955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,128,2560,0.42415253321329754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,128,8192,0.1092351992925008
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,2560,0.01689599951108297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,2560,0.048230401674906415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,128,2048,0.3442805290222168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,128,7168,0.09670186837514241
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,2048,0.009447466333707173
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,2048,0.03374826510747274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,128,1536,0.26405973434448243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,128,6144,0.08356160322825114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,1536,0.007916800181070964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,1536,0.020722132921218873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,128,1024,0.17600320180257162
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,128,5120,0.07012373606363932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,1024,0.006135466694831848
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,1024,0.014844800035158793
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,128,768,0.12411306699117024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,128,4096,0.05648213227589925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,768,0.0056096002459526065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,768,0.015526400009791056
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,128,512,0.08312319914499919
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,128,3584,0.05012693405151367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,512,0.004863999783992767
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,512,0.010778666536013285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,128,256,0.05260053475697836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,128,2560,0.037596801916758224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,256,0.004182399809360504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,256,0.009672533472379048
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,128,128,0.03487679958343506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,128,3072,0.04376213153203328
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,128,0.003751466671625773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,128,0.009038933118184407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,128,64,0.030077866713205975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,64,0.003437866767247518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,64,0.00925546685854594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,128,32,0.028036266565322876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,32,0.0038506666819254553
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,32,0.009329066673914591
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,64,65536,10.009523518880208
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,65536,0.4568309466044108
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,65536,1.1139456431070962
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,64,16384,2.5521759033203124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,16384,0.09940480391184489
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,16384,0.2852106730143229
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,64,12288,1.932916259765625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,12288,0.07319786548614501
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,12288,0.21678506533304848
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,64,10240,1.6216191609700519
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,10240,0.06014506816864014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,10240,0.18299946784973145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,64,8192,1.3112927754720052
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,8192,0.04967466592788696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,8192,0.14604159990946453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,64,7168,1.1302581787109376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,7168,0.0437610665957133
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,7168,0.12863893508911134
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,64,6144,0.9759573618570963
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,6144,0.03832639853159587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,6144,0.11158080101013183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,64,5120,0.8170560201009115
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,5120,0.03290133277575175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,5120,0.09356266657511393
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,64,4096,0.6608021418253581
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,4096,0.02704213261604309
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,4096,0.07563199996948242
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,64,3584,0.5813983917236328
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,3584,0.02417280077934265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,3584,0.0667253335316976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,64,3072,0.5022485415140788
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,3072,0.021449599663416544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,3072,0.057436800003051756
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,64,2560,0.4228373209635417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,2560,0.016641066471735636
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,2560,0.047147734959920244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,64,2048,0.3438847859700521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,2048,0.010393599669138592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,2048,0.032662399609883624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,64,1536,0.26286826133728025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,1536,0.008661333719889324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,1536,0.019939200083414713
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,64,1024,0.17590293884277344
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,1024,0.0067210664351781205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,1024,0.014839466412862143
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,64,768,0.1250432014465332
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,768,0.005866666634877523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,768,0.015265066425005594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,64,512,0.08303253650665283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,512,0.004983466863632202
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,512,0.010455466310183207
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,64,256,0.052254935105641685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,256,0.00407679999868075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,256,0.009593600034713745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,64,128,0.034976001580556235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,128,0.003606399893760681
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,128,0.009080533186594646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,64,64,0.03006400068600972
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,64,0.0034261333445707956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,64,0.009212799866994222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,64,32,0.027833600838979084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,32,0.003509333233038584
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,32,0.009267200032869976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,32,65536,10.005856323242188
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,65536,0.449941349029541
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,65536,1.1133759816487632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,32,16384,2.5523775736490886
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,16384,0.09909226894378662
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,16384,0.28671894073486326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,32,12288,1.9319146474202473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,12288,0.07162986596425375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,12288,0.21748800277709962
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,32,10240,1.6230026245117188
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,10240,0.060071468353271484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,10240,0.18292694091796874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,32,8192,1.311131795247396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,8192,0.04907093445460002
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,8192,0.14579946200052898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,32,7168,1.1302239735921225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,7168,0.04322453339894612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,7168,0.12838826974232992
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,32,6144,0.9746880213419595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,6144,0.038159998257954915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,6144,0.11103040377298992
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,32,5120,0.817084821065267
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,5120,0.032178133726119995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,5120,0.09390613238016764
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,32,4096,0.6605066935221354
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,4096,0.026525866985321046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,4096,0.07570772965749105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,32,3584,0.5817354838053386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,3584,0.023602133989334105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,3584,0.067303466796875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,32,3072,0.5018848101298015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,3072,0.021048533916473388
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,3072,0.056789334615071616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,32,2560,0.4231146812438965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,2560,0.01439466675122579
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,2560,0.04702613353729248
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,32,2048,0.34398508071899414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,2048,0.009085866808891296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,2048,0.032782934109369916
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,32,1536,0.2630005359649658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,1536,0.007501866420110066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,1536,0.02044266661008199
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,32,1024,0.17634347279866536
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,1024,0.005906133353710175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,1024,0.014662399888038635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,32,768,0.12467947006225585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,768,0.00517546683549881
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,768,0.015253333250681558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,32,512,0.08310826619466147
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,512,0.0043488000830014546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,512,0.010523733496665955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,32,256,0.05226346651713053
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,256,0.003738666574160258
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,256,0.00976746678352356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,32,128,0.035062400499979655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,128,0.003419733295838038
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,128,0.009153067072232565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,32,64,0.030059732993443805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,64,0.003159466634194056
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,64,0.009224533041318258
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8192,32,32,0.0277621328830719
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,32,0.00325546662012736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,32,0.009192533294359843
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,65536,16384,8.578052775065105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,65536,16384,6.644600423177083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,65536,16384,13.189747111002603
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,65536,16384,8.386549377441407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,65536,12288,6.608414713541667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,65536,12288,5.016014099121094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,65536,12288,9.416673787434895
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,65536,12288,7.482329813639323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,65536,10240,5.452425638834635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,65536,10240,4.120367940266927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,65536,10240,7.915931701660156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,65536,10240,5.667128499348959
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,65536,8192,4.509016418457032
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,65536,8192,3.3235382080078124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,65536,8192,6.259142557779947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,65536,8192,3.822406514485677
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,65536,7168,4.014779663085937
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,65536,7168,3.0149012247721356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,65536,7168,5.469762166341146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,65536,7168,3.301777140299479
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,65536,6144,3.050731658935547
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,65536,6144,2.550452168782552
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,65536,6144,4.644598388671875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,65536,6144,2.8076713562011717
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,65536,5120,2.5416330973307293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,65536,5120,2.019202168782552
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,65536,5120,3.9222442626953127
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,65536,5120,2.154610188802083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,65536,4096,2.0292522430419924
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,65536,4096,1.6515040079752605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,65536,4096,3.121564737955729
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,65536,4096,1.718828837076823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,65536,3584,1.7784299214680988
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,65536,3584,1.4182090759277344
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,65536,3584,2.9134198506673177
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,65536,3584,1.5055765787760416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,65536,3072,1.5324597676595053
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,65536,3072,1.2331274668375651
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,65536,3072,2.5109174092610678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,65536,3072,1.3115381876627603
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,65536,2560,1.2705023447672525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,65536,2560,1.0116256078084311
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,65536,2560,2.0188106536865233
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,65536,2560,1.076788330078125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,65536,2048,1.015941301981608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,65536,2048,0.8001994450887044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,65536,2048,1.5899754842122396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,65536,2048,0.854741350809733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,65536,1536,0.7574581146240235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,65536,1536,0.625766372680664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,65536,1536,1.170966339111328
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,65536,1536,0.6441674550374349
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,65536,1024,0.5736117045084635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,65536,1024,0.45093758900960285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,65536,1024,0.792471440633138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,65536,1024,0.47717971801757814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,65536,768,0.4501152038574219
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,65536,768,0.3644864082336426
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,65536,768,0.6003936131795247
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,65536,768,0.37150932947794596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,65536,512,0.37224321365356444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,65536,512,0.27955093383789065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,65536,512,0.4338058789571126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,65536,512,0.32103681564331055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,65536,256,0.2955146789550781
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,65536,256,0.23885439236958822
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,65536,256,0.27297172546386717
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,65536,256,0.24974187215169272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,65536,128,0.2676085472106934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,65536,128,0.22605226834615072
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,65536,128,0.2254570643107096
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,65536,128,0.2348639965057373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,65536,64,0.27447360356648765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,65536,64,0.20886507034301757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,65536,64,0.2413813273111979
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,65536,32,0.27915306091308595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,65536,32,0.20311573346455894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,65536,32,0.2483445326487223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,16384,65536,13.823656209309897
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,16384,65536,6.723944600423176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,65536,14.039248657226562
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,65536,9.991848754882813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,16384,16384,2.9625526428222657
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,128,2048,0.030909866094589233
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,16384,3.1090858459472654
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,16384,1.8575402577718099
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,16384,12288,2.2694976806640623
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,16384,16384,1.5826101938883463
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,12288,2.4976341247558596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,12288,1.4060629526774089
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,16384,10240,1.912462870279948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,16384,10240,1.0077418645222982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,16384,12288,1.1844874064127604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,10240,2.0750016530354816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,10240,1.1714560190836587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,16384,8192,1.4904523213704428
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,8192,1.5939637502034505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,8192,0.8875455856323242
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,16384,7168,1.3419882456461587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,16384,8192,0.7646453221638997
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,7168,1.2642485300699868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,7168,0.7739871978759766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,16384,6144,1.1361749013264975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,16384,7168,0.6754485448201497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,6144,1.146448008219401
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,6144,0.6977984110514324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,16384,5120,0.9594485600789389
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,16384,6144,0.5761429468790691
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,5120,0.9259946823120118
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,5120,0.5594741185506185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,16384,4096,0.7637429555257161
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,16384,5120,0.4889994621276855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,4096,0.7371775945027669
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,4096,0.4488383928934733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,16384,3584,0.6762912114461263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,16384,4096,0.3928160031636556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,3584,0.6392757415771484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,3584,0.3917226791381836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,16384,3072,0.5786730448404949
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,16384,3584,0.3513248125712077
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,3072,0.5589087804158528
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,3072,0.3386773427327474
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,16384,2560,0.48764692942301435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,16384,2560,0.26123520533243816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,2560,0.46778132120768234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,2560,0.2833056131998698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,16384,2048,0.38896748224894206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,16384,3072,0.30159145991007485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,2048,0.3769866625467936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,2048,0.2257610638936361
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,16384,1536,0.2799743970235189
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,16384,2048,0.21355199813842773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,1536,0.28454825083414714
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,1536,0.16916693051656087
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,16384,1024,0.18865386644999188
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,16384,1536,0.16917227109273275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,1024,0.1932106653849284
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,1024,0.1194111982981364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,16384,768,0.15369067192077637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,16384,1024,0.12264320055643718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,768,0.15037333170572917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,768,0.09674452940622966
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,16384,512,0.12522026697794597
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,16384,512,0.07656213442484537
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,512,0.11047253608703614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,512,0.08180800278981527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,16384,256,0.09864106973012289
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,16384,256,0.06012266476949056
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,256,0.06842026710510254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,16384,768,0.10034346580505371
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,256,0.0685312032699585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,16384,128,0.09072746435801188
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,128,0.060679467519124355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,128,0.06481279929478964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,16384,64,0.09085120360056559
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,64,0.056346666812896726
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,64,0.06711146831512452
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,16384,32,0.09104426701863608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,32,0.056202665964762366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,32,0.06819626490275064
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,16384,128,0.05895466804504394
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,12288,65536,11.38922119140625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,65536,10.009295654296874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,12288,65536,5.16753184000651
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,12288,16384,2.531433614095052
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,65536,6.949978637695312
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,16384,2.327509307861328
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,16384,1.3664170583089192
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,12288,12288,1.9543445587158204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,12288,16384,1.2072500864664713
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,12288,1.7775904337565105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,12288,1.048085339864095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,12288,10240,1.6184319814046222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,12288,12288,0.8746517181396485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,10240,1.3826165517171225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,10240,0.9039157231648763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,12288,8192,1.3028629302978516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,12288,10240,0.7243061065673828
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,8192,1.067428207397461
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,8192,0.6805898666381835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,12288,7168,1.134216562906901
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,12288,8192,0.5853066762288411
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,7168,0.9738688151041666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,7168,0.6321130752563476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,12288,6144,0.9924885431925455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,12288,7168,0.5121514638264973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,6144,0.7901813507080078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,6144,0.5221834818522135
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,12288,5120,0.8281983693440755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,12288,5120,0.3713450749715169
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,5120,0.6602207819620769
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,5120,0.4287530581156413
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,12288,4096,0.6645984013875326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,12288,6144,0.4431424140930176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,4096,0.5365280151367188
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,4096,0.3486421267191569
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,12288,3584,0.5852863947550456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,12288,4096,0.30051307678222655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,3584,0.4670463879903157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,3584,0.3009546597798666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,12288,3072,0.5028095881144206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,12288,3584,0.2688960075378418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,3072,0.4052042643229167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,3072,0.26309226353963217
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,12288,2560,0.42951145172119143
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,12288,2560,0.2053290685017904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,2560,0.3391722679138184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,2560,0.21781867345174155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,12288,2048,0.3354880015055338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,12288,3072,0.23085974057515463
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,2048,0.2712640126546224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,2048,0.1729365348815918
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,12288,1536,0.24037973086039224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,12288,2048,0.1644704023996989
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,1536,0.2094282627105713
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,1536,0.1326090653737386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,12288,1024,0.1632522741953532
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,12288,1536,0.13018346627553304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,1024,0.14459625879923504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,1024,0.0913973331451416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,12288,768,0.13173120021820067
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,12288,1024,0.09590933322906495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,768,0.11501973470052082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,768,0.07765440146128336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,12288,512,0.10688532988230388
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,12288,768,0.07805866400400797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,512,0.08429439862569174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,512,0.06514240105946859
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,12288,256,0.08363839785257975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,12288,256,0.047413333257039385
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,256,0.05321066776911417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,256,0.054194132486979164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,12288,128,0.07613226572672525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,12288,512,0.0606879989306132
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,128,0.047275733947753903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,128,0.050836265087127686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,12288,64,0.07561813195546468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,64,0.04330666859944661
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,64,0.052169601122538246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,12288,32,0.07544213136037191
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,32,0.043391998608907065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,32,0.053179732958475744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,12288,128,0.046431998411814376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,10240,65536,10.033383178710938
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,65536,7.8719019571940105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,10240,65536,4.12103042602539
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,10240,16384,2.3231358846028645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,65536,6.170063781738281
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,16384,1.9679498036702472
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,16384,1.159393056233724
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,10240,12288,1.8421663920084637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,10240,16384,0.9883210500081381
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,12288,1.3442303975423178
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,12288,0.8781653086344401
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,10240,10240,1.4958250681559244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,10240,12288,0.7355552037556966
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,10240,1.203873062133789
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,10240,0.7820650736490886
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,10240,8192,1.1929536183675131
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,10240,10240,0.6095562616984049
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,8192,0.8858101526896158
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,8192,0.5758442560831706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,10240,7168,1.0377983729044595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,10240,8192,0.49446827570597335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,7168,0.7821919759114583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,7168,0.5432181040445964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,10240,6144,0.9334431966145834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,10240,7168,0.43184213638305663
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,6144,0.6637407938639324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,6144,0.44429759979248046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,10240,5120,0.7509653091430664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,10240,5120,0.31380160649617517
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,5120,0.5574645360310873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,5120,0.36422932942708336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,10240,4096,0.6077184041341146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,10240,6144,0.374124813079834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,4096,0.4517770767211914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,4096,0.2968853314717611
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,10240,3584,0.53963197072347
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,10240,4096,0.2545130729675293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,3584,0.39454933802286785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,3584,0.2558656056722005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,10240,3072,0.4640917460123698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,10240,3584,0.2273728052775065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,3072,0.3431040128072103
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,3072,0.22026880582173666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,10240,2560,0.38764266967773436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,10240,3072,0.1960085391998291
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,2560,0.28707733154296877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,2560,0.18183253606160482
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,10240,2048,0.306172784169515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,10240,2560,0.16903893152872723
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,2048,0.23432106971740724
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,2048,0.14834346771240234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,10240,1536,0.22298026084899902
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,10240,2048,0.13909866015116373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,1536,0.17756907145182294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,1536,0.1116096019744873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,10240,1024,0.14828267097473144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,10240,1536,0.11098986466725666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,1024,0.12511680126190186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,1024,0.07918079694112143
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,10240,768,0.1217685302098592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,10240,1024,0.08161919911702474
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,768,0.09823253154754638
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,768,0.06572159926096598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,10240,512,0.0977450688680013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,10240,768,0.06732693513234457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,512,0.07351893583933512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,512,0.056246399879455566
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,10240,256,0.07585813204447428
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,10240,512,0.05166186491648356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,256,0.04981013139088948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,256,0.04700160026550293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,10240,128,0.06819519996643067
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,10240,256,0.04049280087153117
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,128,0.04090346495310466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,128,0.04346666733423869
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,10240,64,0.0678442637125651
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,64,0.036645332972208663
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,64,0.0446559985478719
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,10240,32,0.06781013011932373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,32,0.03642880121866862
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,32,0.04547839959462484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,10240,128,0.04009600083033244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,8192,65536,8.809239705403645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,65536,6.077715047200521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,8192,65536,3.465692901611328
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,8192,16384,2.0840192159016926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,65536,4.290885416666667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,16384,1.473077392578125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,16384,0.9286752065022787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,8192,12288,1.6173130035400392
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,8192,16384,0.8082741419474283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,12288,1.1058816274007162
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,12288,0.719755744934082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,8192,10240,1.3640640258789063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,8192,10240,0.5002762794494628
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,10240,0.8977418899536133
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,10240,0.6149759928385417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,8192,12288,0.6012671788533528
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,8192,8192,1.0828938802083334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,8192,0.7051872253417969
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,8192,0.4718282699584961
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,8192,7168,0.9488735834757487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,8192,8192,0.40346453984578445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,7168,0.622109858194987
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,7168,0.43266986211140945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,8192,6144,0.8311029434204101
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,8192,6144,0.3067765235900879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,8192,7168,0.3533568064371745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,6144,0.5310346603393554
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,6144,0.36497599283854165
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,8192,5120,0.6959744135538737
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,5120,0.4448992093404134
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,5120,0.3001215934753418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,8192,4096,0.5612085342407227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,8192,5120,0.2559648036956787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,4096,0.3558890660603841
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,4096,0.23864320119222007
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,8192,3584,0.48425919214884444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,8192,4096,0.20941227277119956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,3584,0.3150858561197917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,3584,0.21120853424072267
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,8192,3072,0.4221813201904297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,8192,3584,0.18559892972310382
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,3072,0.2700970649719238
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,3072,0.17996586163838704
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,8192,2560,0.35326080322265624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,8192,2560,0.13944320678710936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,2560,0.22688746452331543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,2560,0.15029120445251465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,8192,3072,0.1613642692565918
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,8192,2048,0.2811584154764811
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,2048,0.1833024024963379
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,2048,0.12177279790242512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,8192,1536,0.20174506505330406
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,8192,2048,0.11490133603413899
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,1536,0.14084374109903972
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,1536,0.09322559833526611
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,8192,1024,0.13381013870239258
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,8192,1024,0.06747519969940186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,8192,1536,0.09080426692962647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,1024,0.09926506678263346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,1024,0.0633407990137736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,8192,768,0.10988159974416097
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,768,0.07854932943979899
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,768,0.05539520184199015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,8192,512,0.08853653271993002
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,8192,768,0.054679465293884275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,512,0.057557332515716556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,512,0.046826668580373126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,8192,256,0.06829333305358887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,8192,512,0.04197440147399902
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,256,0.03812266588211059
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,256,0.03915733496348063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,8192,128,0.06115093231201172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,8192,256,0.033291733264923094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,128,0.03284693360328674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,128,0.036473600069681804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,8192,64,0.060078934828440345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,64,0.029806933800379437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,64,0.03726079861323039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,8192,128,0.03253440062204997
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,8192,32,0.05928959846496582
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,32,0.030264532566070555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,32,0.03748266696929932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,7168,65536,8.285227966308593
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,7168,65536,3.0970250447591146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,65536,5.64691416422526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,7168,16384,1.9944149017333985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,65536,4.065075174967448
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,16384,1.35524050394694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,16384,0.8423487981160482
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,7168,12288,1.5160011291503905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,7168,12288,0.5309365272521973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,12288,0.943506113688151
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,12288,0.6527701059977213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,7168,10240,1.299311955769857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,7168,16384,0.7070154825846354
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,10240,0.7899349212646485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,10240,0.549608548482259
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,7168,8192,1.0365514755249023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,7168,10240,0.44354880650838213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,8192,0.611572265625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,8192,0.41942612330118817
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,7168,7168,0.9199530919392904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,7168,8192,0.3590016047159831
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,7168,0.5358954747517903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,7168,0.375767453511556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,7168,6144,0.7802752176920573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,7168,7168,0.3156458536783854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,6144,0.46406186421712237
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,6144,0.3289823849995931
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,7168,5120,0.6550399780273437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,7168,6144,0.27102505366007484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,5120,0.38691733678181967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,5120,0.2644277254740397
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,7168,4096,0.5247530619303385
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,7168,5120,0.22942186991373698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,4096,0.3137855847676595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,4096,0.21346240043640136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,7168,3584,0.47098347345987956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,7168,4096,0.18597973187764485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,3584,0.2744789441426595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,3584,0.18550400733947753
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,7168,3072,0.40055360794067385
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,7168,3584,0.16594986915588378
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,3072,0.23833492596944175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,3072,0.16082666714986166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,7168,2560,0.3374549229939779
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,7168,3072,0.14377387364705402
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,2560,0.19887572924296062
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,2560,0.13379519780476887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,7168,2048,0.2670175870259603
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,7168,2048,0.10217599868774414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,2048,0.160971736907959
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,2048,0.1084554672241211
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,7168,1536,0.19276053110758465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,7168,1536,0.08122453689575196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,1536,0.12372266451517741
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,1536,0.0826527992884318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,7168,1024,0.12744853496551514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,7168,2560,0.12401386896769206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,1024,0.08764159679412842
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,1024,0.058238931496938075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,7168,768,0.10497813224792481
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,7168,1024,0.060377601782480875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,768,0.06885759830474854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,768,0.05023786624272665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,7168,512,0.08429226875305176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,7168,768,0.04989226659138997
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,512,0.051210665702819826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,512,0.043144532044728595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,7168,256,0.06440320014953613
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,7168,512,0.03895893494288127
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,256,0.03388479948043823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,256,0.03535039822260539
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,7168,128,0.057148798306783044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,7168,256,0.031489066282908124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,128,0.029657600323359172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,128,0.03296320041020711
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,7168,64,0.05672640005747477
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,64,0.027147734165191652
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,64,0.03401493231455485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,7168,32,0.056301867961883544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,32,0.027034666140874224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,32,0.034434131781260174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,7168,128,0.029579732815424604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,6144,65536,7.7074462890625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,65536,4.7369847615559895
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,65536,3.183366394042969
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,6144,16384,1.9126079559326172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,6144,65536,2.6258944193522136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,16384,1.1261547088623047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,16384,0.7327477137247722
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,6144,12288,1.4771007537841796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,6144,12288,0.4625109354654948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,12288,0.786571756998698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,12288,0.5604703903198243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,6144,10240,1.2232896169026692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,6144,16384,0.6178922653198242
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,10240,0.6781600316365559
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,10240,0.47978668212890624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,6144,8192,0.9834645589192709
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,6144,10240,0.3852160135904948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,8192,0.5237770716349284
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,8192,0.36835947036743166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,6144,7168,0.850220807393392
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,6144,8192,0.3120448112487793
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,7168,0.46248960494995117
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,7168,0.32901439666748045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,6144,6144,0.7393760045369466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,6144,7168,0.2736469268798828
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,6144,0.3952778816223145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,6144,0.284114138285319
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,6144,5120,0.6211903889973958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,6144,6144,0.23691306114196778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,5120,0.333517869313558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,5120,0.2349407990773519
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,6144,4096,0.5053344090779622
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,6144,4096,0.16535040537516277
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,4096,0.2694101333618164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,4096,0.18576745986938475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,6144,3584,0.44150400161743164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,6144,5120,0.19916906356811523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,3584,0.235371732711792
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,3584,0.16225813229878744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,6144,3584,0.14524374008178711
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,6144,3072,0.3803338686625163
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,3072,0.20319679578145347
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,3072,0.13992212613423666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,6144,2560,0.32015679677327474
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,6144,3072,0.12646933396657306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,2560,0.1718762715657552
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,2560,0.11729919910430908
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,6144,2048,0.25423787434895834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,6144,2560,0.10860586961110433
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,2048,0.1384042739868164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,2048,0.09547733465830485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,6144,1536,0.1824512004852295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,6144,2048,0.09013653596242269
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,1536,0.10711466471354167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,1536,0.07306773662567138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,6144,1024,0.12021013100941975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,6144,1536,0.0718015988667806
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,1024,0.07523413499196371
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,1024,0.05057173172632853
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,6144,768,0.09868160088857016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,6144,1024,0.05328319867451986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,768,0.06018133163452148
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,768,0.04448320070902507
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,6144,512,0.07911787033081055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,6144,768,0.04371093511581421
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,512,0.044742401440938315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,512,0.03829866647720337
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,6144,256,0.059944534301757814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,6144,512,0.03519253333409627
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,256,0.030289065837860108
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,256,0.031465599934260055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,6144,128,0.053356798489888516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,6144,256,0.02868266701698303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,128,0.02622399926185608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,128,0.029469867547353108
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,6144,64,0.052194134394327796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,64,0.02434879938761393
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,64,0.030322132507960002
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,6144,32,0.05161386728286743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,32,0.025256532430648803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,32,0.030036266644795733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,5120,65536,7.1865580240885425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,6144,128,0.02653653422991435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,65536,4.006904602050781
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,65536,2.762451171875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,5120,16384,1.7970826466878254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,5120,65536,2.19486821492513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,16384,0.9276608149210611
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,16384,0.6387850443522136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,5120,12288,1.391428248087565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,5120,16384,0.5258421262105306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,12288,0.66954771677653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,12288,0.4896405220031738
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,5120,10240,1.1584949493408203
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,5120,12288,0.3941802660624186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,10240,0.5490037282307942
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,10240,0.41075731913248703
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,5120,8192,0.9354954401652018
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,5120,10240,0.33193705876668295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,8192,0.4441503842671712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,8192,0.31747519175211586
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,5120,7168,0.8067925135294596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,5120,8192,0.2668959935506185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,7168,0.3869866689046224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,7168,0.2797173182169596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,5120,6144,0.6969727834065755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,5120,7168,0.2360223929087321
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,6144,0.3357301394144694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,6144,0.24823466936747232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,5120,5120,0.5892010370890299
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,5120,6144,0.20314133961995445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,5120,0.2797034581502279
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,5120,0.19816746711730956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,5120,4096,0.4714837392171224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,5120,5120,0.17137173016866047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,4096,0.22704960505167643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,4096,0.1570485273996989
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,5120,3584,0.41714134216308596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,5120,4096,0.1403125286102295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,3584,0.19903146425882975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,3584,0.14018880526224772
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,5120,3072,0.3599967956542969
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,5120,3584,0.12534399827321369
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,3072,0.17279359499613445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,3072,0.1212608019510905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,5120,2560,0.3038688023885091
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,5120,3072,0.10854399998982747
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,2560,0.1450655937194824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,2560,0.10137920379638672
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,5120,2048,0.2409663995107015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,5120,2560,0.0940224011739095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,2048,0.11885440349578857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,2048,0.08130880196889242
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,5120,1536,0.17258772850036622
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,5120,2048,0.0784063975016276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,1536,0.09100586573282878
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,1536,0.06286613146464029
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,5120,1024,0.11413546403249104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,5120,1536,0.06188906828562418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,1024,0.06514559984207154
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,1024,0.045094398657480876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,5120,768,0.09379519621531168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,5120,1024,0.046529066562652585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,768,0.05213760137557984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,768,0.039104000727335615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,5120,512,0.07487893104553223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,5120,512,0.03129813273747762
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,512,0.039738667011260984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,512,0.03337599833806355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,5120,256,0.056731732686360684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,5120,768,0.038523733615875244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,256,0.027754666407903035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,256,0.02805333336194356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,5120,128,0.04967253208160401
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,5120,256,0.025094399849573772
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,128,0.022379734118779502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,128,0.025678932666778564
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,5120,64,0.04806400140126546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,64,0.01971093416213989
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,64,0.026341332991917925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,5120,32,0.0474613348642985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,32,0.02004693349202474
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,32,0.026146133740743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,5120,128,0.022372267643610635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,4096,65536,6.595593770345052
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,65536,3.1745941162109377
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,65536,2.0687508900960285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,4096,65536,1.8652032216389973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,4096,16384,1.6973492940266925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,16384,0.6923168182373047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,16384,0.49628267288208006
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,4096,12288,1.2758613586425782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,4096,16384,0.433458137512207
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,12288,0.5237141291300456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,12288,0.3869226773579916
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,4096,10240,1.0978421529134115
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,4096,12288,0.3254165331522624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,10240,0.43550825119018555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,10240,0.33284266789754235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,4096,8192,0.8599562962849936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,4096,10240,0.2740405400594076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,8192,0.3515455881754557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,8192,0.25533547401428225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,4096,7168,0.7528181076049805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,4096,8192,0.2204906622568766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,7168,0.30720640818277994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,7168,0.22512426376342773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,4096,6144,0.6477696100870769
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,4096,7168,0.1950559933980306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,6144,0.2658026695251465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,6144,0.1932202657063802
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,4096,5120,0.5410485585530599
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,4096,6144,0.1688192049662272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,5120,0.22192853291829429
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,5120,0.15806293487548828
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,4096,4096,0.4367061297098796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,4096,4096,0.11699626445770264
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,4096,0.17928427060445148
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,4096,0.1215669314066569
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,4096,3584,0.38175894419352213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,4096,5120,0.14317545890808106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,3584,0.15891092618306477
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,3584,0.10418986479441325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,4096,3072,0.33276265462239585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,4096,3584,0.1044927994410197
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,3072,0.13681920369466144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,3072,0.08504319985707601
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,4096,2560,0.2798656145731608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,4096,3072,0.09098133246103922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,2560,0.11610559622446696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,2560,0.07696533203125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,4096,2048,0.2203125317891439
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,4096,2560,0.07913173039754232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,2048,0.09408746560414633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,2048,0.06272106568018596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,4096,1536,0.1559658686319987
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,4096,2048,0.06551253398259481
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,1536,0.0729429324467977
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,1536,0.049132800102233885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,4096,1024,0.10451412995656331
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,4096,1536,0.052452266216278076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,1024,0.05155626535415649
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,1024,0.03579306602478027
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,4096,768,0.08599680264790853
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,4096,1024,0.03944746653238933
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,768,0.041544532775878905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,768,0.031557333469390866
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,4096,512,0.0679626703262329
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,4096,768,0.033053867022196454
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,512,0.03168639938036601
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,512,0.025729066133499144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,4096,256,0.050819198290506996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,4096,256,0.019063466787338258
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,256,0.021953066190083824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,256,0.021666133403778078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,4096,128,0.04414079984029134
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,4096,512,0.02626986702283223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,128,0.017017600933710735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,4096,128,0.016777600844701132
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,128,0.01947093407313029
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,4096,64,0.04205333391825358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,64,0.014425599575042724
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,64,0.020167466004689535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,4096,32,0.04150826533635457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,32,0.015251200397809347
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,32,0.020772266387939452
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3584,65536,6.3625239054361975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,65536,2.9018145243326825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,65536,1.833350372314453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3584,65536,1.8729119618733723
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3584,16384,1.6273216247558593
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,16384,0.6427498499552409
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,16384,0.45409921010335286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3584,12288,1.2410388946533204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3584,16384,0.44164053599039715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,12288,0.48756160736083987
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,12288,0.35505174001057943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3584,10240,1.046665573120117
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3584,12288,0.33259731928507485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,10240,0.40506451924641923
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,10240,0.29999252955118816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3584,8192,0.8336085637410482
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3584,8192,0.22665813763936363
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,8192,0.3254495938618978
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,8192,0.23065706888834633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3584,7168,0.7248853047688801
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3584,10240,0.27944533030192054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,7168,0.287659740447998
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,7168,0.20382293065388998
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3584,6144,0.6325600306193034
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3584,7168,0.19905279477437338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,6144,0.24586025873819986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,6144,0.17703253428141277
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3584,5120,0.5248170534769694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3584,6144,0.17278720537821451
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,5120,0.20807466506958008
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,5120,0.1425535996754964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3584,4096,0.4267210642496745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3584,5120,0.14573012987772624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,4096,0.166702938079834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,4096,0.11275200049082439
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3584,3584,0.374398930867513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3584,4096,0.1194048007329305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,3584,0.14805760383605956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,3584,0.09682133197784423
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3584,3072,0.3251615842183431
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3584,3072,0.09278186957041422
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,3072,0.12844693660736084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,3072,0.08314666748046876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3584,2560,0.2749631881713867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3584,3584,0.1075711965560913
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,2560,0.1077898661295573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,2560,0.07362346649169922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3584,2048,0.21615893046061196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3584,2048,0.06712426344553629
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,2048,0.08786453406016032
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,2048,0.05847040017445883
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3584,1536,0.15329705874125163
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3584,2560,0.08138559659322103
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,1536,0.06818133195241292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,1536,0.04497493505477905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3584,1024,0.10272426605224609
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3584,1536,0.05356160004933676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,1024,0.049160532156626385
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,1024,0.033266133069992064
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3584,768,0.08423999945322672
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3584,1024,0.042114134629567465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,768,0.03952320019404094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,768,0.029288534323374433
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3584,512,0.06611093282699584
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3584,768,0.03384213447570801
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,512,0.029523199796676634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,512,0.02355626622835795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3584,256,0.04943360090255737
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3584,512,0.027508266766866046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,256,0.020087466637293497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,256,0.02026453415552775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3584,128,0.04287039836247762
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3584,256,0.020283732811609903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,128,0.015020799636840821
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,128,0.017909334103266398
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3584,64,0.04009493192036946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,64,0.013904000322024027
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,64,0.01841813325881958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3584,32,0.039408000310262044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,32,0.015597866972287497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,32,0.01880319913228353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3072,65536,6.1408330281575525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3584,128,0.017992534240086875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,65536,2.266057586669922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,65536,1.6574442545572918
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3072,16384,1.580575942993164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3072,65536,1.4507520039876303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,16384,0.5492917378743489
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,16384,0.4021002769470215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3072,12288,1.2068608601888022
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3072,16384,0.3464511871337891
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,12288,0.3931839942932129
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,12288,0.3179306666056315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3072,10240,1.010582415262858
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3072,12288,0.26104639371236166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,10240,0.3262751897176107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,10240,0.2660906632741292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3072,8192,0.8068362553914389
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3072,10240,0.21958506902058922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,8192,0.2644320011138916
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,8192,0.2050645351409912
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3072,7168,0.7028032302856445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3072,8192,0.17655146916707357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,7168,0.23031466801961265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,7168,0.1797226587931315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3072,6144,0.6085226694742839
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3072,7168,0.15719146728515626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,6144,0.20027947425842285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,6144,0.1583295981089274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3072,5120,0.5110368092854818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3072,6144,0.13550186157226562
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,5120,0.16671147346496581
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,5120,0.12703359921773275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3072,4096,0.4152938524881999
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3072,5120,0.11545706590016683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,4096,0.13620799382527668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,4096,0.10185919602711994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3072,3584,0.36720641454060876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3072,4096,0.09481493631998697
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,3584,0.11919466654459636
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,3584,0.08552213509877524
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3072,3072,0.316483211517334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3072,3584,0.08416319688161214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,3072,0.10434026718139648
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,3072,0.0740885337193807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3072,2560,0.26707305908203127
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3072,3072,0.07426986694335938
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,2560,0.08742506504058838
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,2560,0.06361813147862752
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3072,2048,0.21080212593078612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3072,2560,0.06377706527709961
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,2048,0.071833602587382
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,2048,0.052558934688568114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3072,1536,0.14921174049377442
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3072,2048,0.053875199953715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,1536,0.05536533196767172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,1536,0.03998719851175944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3072,1024,0.1001973311106364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3072,1536,0.0425439993540446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,1024,0.04017173449198405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,1024,0.029986133178075153
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3072,768,0.0812490701675415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3072,1024,0.03223573366800944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,768,0.03277226686477661
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,768,0.025960532824198405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3072,512,0.06396480003992716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3072,768,0.026213334004084272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,512,0.025016532341639204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,512,0.020716800292332967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3072,256,0.04692906538645426
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3072,512,0.021028266350428263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,256,0.017784533898035686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,256,0.017318399747212727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3072,128,0.040267733732859294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3072,256,0.014944000045458474
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,128,0.013556266824404398
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,128,0.015557333827018738
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3072,64,0.03750186761220296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,64,0.011993599931399028
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,64,0.01653439998626709
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,3072,32,0.035640533765157065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,32,0.0125109334786733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,32,0.017001599073410034
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3072,128,0.013247999548912048
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2560,65536,5.955619303385417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,65536,1.8172320048014323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,65536,1.4333216349283853
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2560,16384,1.5313461303710938
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2560,65536,1.3620928446451823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,16384,0.43210026423136394
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,16384,0.36012051900227865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2560,12288,1.1537269592285155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2560,16384,0.35465920766194664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,12288,0.32552318572998046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,12288,0.278983465830485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2560,10240,0.9869557062784831
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2560,12288,0.26692800521850585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,10240,0.27079572677612307
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,10240,0.233189328511556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2560,8192,0.7838997522989909
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2560,10240,0.22346240679423013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,8192,0.21881386439005532
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,8192,0.1834389368693034
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2560,7168,0.6832106908162434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2560,8192,0.181549866994222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,7168,0.1910410722096761
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,7168,0.15799892743428548
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2560,6144,0.5907626469930013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2560,6144,0.1375989278157552
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,6144,0.1651626745859782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,6144,0.13766825993855794
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2560,5120,0.4991349220275879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2560,7168,0.15990079243977864
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,5120,0.1391146659851074
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,5120,0.11374293168385823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2560,4096,0.40285971959431965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2560,5120,0.11745279630025227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,4096,0.11201600233713786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,4096,0.0884341319402059
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2560,3584,0.35512959162394203
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2560,4096,0.09528106848398844
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,3584,0.0991701364517212
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,3584,0.07789653142293294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2560,3072,0.3079327901204427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2560,3584,0.0851733366648356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,3072,0.08563306331634521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,3072,0.06603093147277832
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2560,2560,0.260702927907308
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2560,3072,0.07453546524047852
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,2560,0.07266133626302083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,2560,0.056995201110839847
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2560,2048,0.20489706993103027
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2560,2560,0.0650762677192688
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,2048,0.058906666437784826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,2048,0.046510934829711914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2560,1536,0.14346453348795574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2560,2048,0.05390826861063639
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,1536,0.045976531505584714
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,1536,0.0356607993443807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2560,1024,0.09637760321299235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2560,1536,0.043613866964975996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,1024,0.03299306631088257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,1024,0.02585386633872986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2560,768,0.07875093619028727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2560,1024,0.033429332574208576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,768,0.025947733720143633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,768,0.022900267442067464
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2560,512,0.06118506590525309
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2560,768,0.027811199426651
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,512,0.019927465915679933
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,512,0.018359466393788656
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2560,256,0.04170346657435099
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2560,512,0.0221781333287557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,256,0.013823999961217245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,256,0.015554133057594299
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2560,128,0.035656531651814774
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2560,256,0.015899733702341715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,128,0.010626133282979329
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,128,0.014042666554450989
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2560,64,0.0346346656481425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,64,0.009287466605504353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,64,0.015063466628392539
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2560,32,0.03294826745986938
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,32,0.009562666217486065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,32,0.014633599917093912
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2560,128,0.014648532867431641
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2048,65536,5.737370808919271
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,65536,1.4726282755533853
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,65536,1.2445109049479166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2048,16384,1.4742229461669922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2048,65536,1.0176629384358724
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,16384,0.34825493494669596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,16384,0.3150282541910807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2048,12288,1.12325439453125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2048,16384,0.2565120061238607
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,12288,0.2610325336456299
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,12288,0.23772692680358887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2048,10240,0.9393632253011068
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2048,12288,0.19490346908569336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,10240,0.21921599706014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,10240,0.20138346354166664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2048,8192,0.7564704259236653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2048,10240,0.1639680067698161
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,8192,0.17535893122355145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,8192,0.15708373387654623
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2048,7168,0.6631914774576824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2048,8192,0.13253546555836995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,7168,0.1550165335337321
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,7168,0.1398303985595703
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2048,6144,0.5753013610839843
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2048,7168,0.11711466312408447
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,6144,0.13295893669128417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,6144,0.11995946566263835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2048,5120,0.4830581347147624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2048,6144,0.101909335454305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,5120,0.11230826377868652
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,5120,0.10107946395874023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2048,4096,0.39269332885742186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2048,5120,0.0869653304417928
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,4096,0.09033919970194498
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,4096,0.07842559814453125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2048,3584,0.34491093953450525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2048,4096,0.0714581330617269
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,3584,0.0804533322652181
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,3584,0.06720533370971679
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2048,3072,0.2995509465535482
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2048,3584,0.06495466629664103
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,3072,0.06910080115000407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,3072,0.05860480070114136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2048,2560,0.25387733777364097
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2048,3072,0.0566922664642334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,2560,0.058628265062967935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2048,2560,0.04998933474222819
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,2560,0.04966293176015218
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2048,2048,0.20062185923258463
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,2048,0.047890134652455646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,2048,0.040827735265096025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2048,1536,0.14187946319580078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2048,2048,0.0413045326868693
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,1536,0.03709760109583537
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,1536,0.030333866675694782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2048,1024,0.09183039665222167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2048,1536,0.032708267370859787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,1024,0.026650667190551758
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,1024,0.021967999140421548
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2048,768,0.07587412993113199
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2048,1024,0.0240064005057017
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,768,0.021156267325083414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,768,0.02026559909184774
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2048,512,0.05821333328882853
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2048,768,0.019875200589497884
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,512,0.016335999965667723
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,512,0.016334933042526246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2048,256,0.03986453215281169
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2048,512,0.016140799721082053
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,256,0.011687466502189636
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2048,256,0.012121599912643433
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,256,0.013699199755986533
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2048,128,0.03374293247858683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2048,128,0.01113813320795695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,128,0.009121066331863404
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,128,0.012264532844225566
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2048,64,0.03197866678237915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,64,0.007930666704972585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,64,0.01285653313000997
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,2048,32,0.03139413396517436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,32,0.008172800143559773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,32,0.012997333208719888
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1536,65536,5.533043416341146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,65536,1.1304351806640625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,65536,1.0550912221272788
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1536,16384,1.4368181864420573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1536,65536,0.8543338775634766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,16384,0.2672032038370768
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,16384,0.2683178583780924
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1536,12288,1.0861493428548177
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1536,16384,0.2112895965576172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,12288,0.20031466484069824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,12288,0.20392319361368813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1536,10240,0.9104127883911133
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1536,10240,0.13347519238789876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,10240,0.1675722599029541
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,10240,0.1703829288482666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1536,8192,0.7370943705240885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1536,12288,0.16027414004007975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,8192,0.13549760182698567
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,8192,0.1358261267344157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1536,7168,0.6424095789591472
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1536,8192,0.10945813655853272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,7168,0.11840853691101075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,7168,0.1178752024968465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1536,6144,0.5564938863118489
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1536,7168,0.09615253607432048
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,6144,0.10279253323872883
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,6144,0.1041866699854533
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1536,5120,0.469052791595459
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1536,5120,0.0713045358657837
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,5120,0.08647786776224772
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,5120,0.0861525297164917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1536,4096,0.38124478658040367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,128,1536,0.024000000953674317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,4096,0.07087893486022949
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,4096,0.0661525328954061
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1536,6144,0.08446826934814453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1536,3584,0.3360533396402995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,3584,0.062459735075632725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,3584,0.0587007999420166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1536,3072,0.2917322794596354
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1536,4096,0.059507199128468836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,3072,0.053311999638875326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,3072,0.05028586784998575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1536,2560,0.24755733807881675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1536,3584,0.05405120054880778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,2560,0.0449504017829895
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,2560,0.04349759817123413
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1536,2048,0.19488213857014974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1536,3072,0.047933868567148846
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,2048,0.036823467413584395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,2048,0.03369919856389363
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1536,1536,0.137500794728597
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1536,2560,0.04107306798299153
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,1536,0.02765120069185893
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,1536,0.02542720039685567
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1536,1024,0.09002453486124674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1536,2048,0.033877333005269364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,1024,0.020629332462946574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,1024,0.018709333737691243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1536,768,0.07201066811879477
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1536,1536,0.02688213388125102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,768,0.01676693360010783
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,768,0.017732266585032144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1536,512,0.056036265691121426
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1536,1024,0.019742933909098308
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,512,0.013054933150609335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,512,0.014062933127085366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1536,256,0.037939198811848956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1536,768,0.0168778657913208
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,256,0.009315199653307597
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,256,0.011779200037320454
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1536,128,0.03202986717224121
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1536,128,0.009799466530481974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,128,0.007533866663773854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,128,0.010711466272672016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1536,64,0.03006613254547119
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,64,0.006466133395830791
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,64,0.011099732915560405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1536,32,0.0295306662718455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,32,0.006775466601053874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,32,0.011220266421635944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1536,512,0.014225066701571146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1536,256,0.011076266566912334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1024,65536,5.336568705240885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,65536,0.6977898915608723
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,65536,0.8617653528849283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1024,16384,1.39333864847819
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,16384,0.17698559761047364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,16384,0.2166133403778076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1024,12288,1.0488864262898763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1024,65536,0.665506108601888
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,12288,0.13262399832407634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,12288,0.16968960762023927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1024,10240,0.8807039896647135
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1024,16384,0.17065812746683756
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,10240,0.10990613301595051
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,10240,0.1383072058359782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1024,8192,0.713590431213379
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1024,12288,0.13004266421000163
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,8192,0.0902079979578654
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,8192,0.11219093004862468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1024,7168,0.6219104131062825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1024,10240,0.1085642655690511
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,7168,0.07873067061106363
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,7168,0.10075626373291016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1024,6144,0.5405984242757161
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1024,8192,0.0893280029296875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,6144,0.06770239671071371
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,6144,0.08577280044555664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1024,5120,0.453003724416097
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1024,7168,0.0793557325998942
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,5120,0.058100267251332605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,5120,0.07088853518168131
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1024,4096,0.36869866053263345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1024,6144,0.06886080106099447
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,4096,0.04667093356450398
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,4096,0.05687573353449503
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1024,3584,0.32623573939005535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1024,5120,0.05962986548741659
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,3584,0.04094719886779785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,3584,0.04853973388671875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1024,3072,0.28219305674235023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1024,4096,0.04999573230743408
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,3072,0.03628693421681722
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,3072,0.0412778655687968
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1024,2560,0.23914987246195474
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1024,3584,0.044810668627421064
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,2560,0.030289065837860108
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,2560,0.0354912002881368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1024,2048,0.18977386156717937
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1024,3072,0.039350398381551105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,2048,0.02502506573994954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,2048,0.02564479907353719
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1024,1536,0.13385814030965168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1024,2560,0.03380906581878662
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,1536,0.019718400637308755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,1536,0.020483199755350748
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1024,1024,0.08727893034617105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1024,2048,0.027321600914001466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,1024,0.014602667093276978
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,1024,0.016458666324615477
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1024,768,0.07101866404215494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1024,1536,0.021592533588409422
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,768,0.012232533097267151
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,768,0.015306666493415833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1024,512,0.05358826716740926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1024,1024,0.016747732957204185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,512,0.009949866930643718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,512,0.011777066191037496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1024,256,0.03603733380635579
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1024,256,0.010167466600735982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,256,0.00749120016892751
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,256,0.010170666376749675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1024,128,0.030261333783467608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1024,768,0.014631467064221701
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,128,0.006287999947865804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,128,0.009226666887601216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1024,64,0.028293333450953168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,64,0.005489066739877065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,64,0.009516800443331402
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,1024,32,0.028032000859578448
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,32,0.005721599857012431
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,32,0.009563733140627544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1024,512,0.012446932991345723
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1024,128,0.008962133526802063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,768,65536,5.289748128255209
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,65536,0.5332522710164388
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,65536,0.8265141169230142
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,768,16384,1.364461898803711
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,16384,0.13531732559204102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,16384,0.2062837282816569
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,768,12288,1.029853884379069
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,768,65536,0.5782047907511394
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,12288,0.10356372992197674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,12288,0.1599605401357015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,768,10240,0.8703861236572266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,768,16384,0.14702612559000652
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,10240,0.08512319723765055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,10240,0.1328757365544637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,768,8192,0.7047050476074219
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,768,8192,0.07694933414459229
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,8192,0.0686954657236735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,8192,0.1067914644877116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,768,12288,0.11201706727345784
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,768,7168,0.06870400110880534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,768,7168,0.6152704238891602
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,7168,0.06115946769714355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,7168,0.09463679790496826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,768,6144,0.5343786875406902
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,6144,0.0526090661684672
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,6144,0.08151893615722657
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,768,5120,0.4485450744628906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,768,10240,0.09490559895833334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,5120,0.04492479960123698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,5120,0.06792746384938558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,768,4096,0.36563199361165366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,768,6144,0.06047039826711019
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,4096,0.03602773348490397
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,4096,0.05458240111668905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,768,3584,0.3230069478352865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,768,5120,0.05186880032221476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,3584,0.03185706734657288
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,3584,0.04593066771825154
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,768,3072,0.2811530749003092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,768,4096,0.04358506600062052
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,3072,0.02759466568628947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,3072,0.038925866285959884
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,768,2560,0.23753172556559243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,768,3584,0.039367465178171794
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,2560,0.023260800043741862
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,2560,0.03233493367830913
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,768,2048,0.18800427118937174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,768,2048,0.024086399873097738
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,2048,0.01932906707127889
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,2048,0.023963733514149984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,768,3072,0.0343722661336263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,768,1536,0.019151999553044637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,768,1536,0.13234346707661945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,1536,0.01537493367989858
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,1536,0.01966080069541931
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,768,1024,0.08727146784464518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,1024,0.011643733580907185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,1024,0.015729066729545594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,768,768,0.07045546372731527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,768,2560,0.02929813265800476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,768,0.01000853379567464
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,768,0.014820266763369241
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,768,512,0.05361066659291586
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,768,1024,0.014964266618092855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,512,0.008035199840863545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,512,0.011606400211652119
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,768,768,0.013215999801953635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,768,256,0.036106665929158524
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,256,0.006298666695753734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,256,0.010033067067464192
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,768,128,0.030747733513514203
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,768,512,0.011497599879900615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,128,0.005397333204746247
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,128,0.009139200051625569
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,768,64,0.028373332818349202
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,64,0.004745600124200186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,64,0.009355733791987102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,768,32,0.027514666318893433
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,32,0.004954666892687479
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,32,0.009292800227801006
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,512,65536,5.131754557291666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,768,256,0.009222400188446046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,768,128,0.008272000153859456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,65536,0.3701653480529785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,65536,0.660813840230306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,512,16384,0.13291626771291096
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,16384,0.09173546632130941
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,512,16384,1.324511973063151
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,16384,0.16715946197509765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,512,12288,1.0067520141601562
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,12288,0.07013866901397706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,12288,0.13005440235137938
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,512,10240,0.8493024190266928
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,512,10240,0.08955519994099935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,10240,0.05910933415095011
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,10240,0.11004373232523601
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,512,8192,0.6888490676879883
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,512,65536,0.5205183982849121
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,8192,0.049250133832295734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,512,12288,0.10375040372212727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,8192,0.0865290641784668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,512,7168,0.5985621134440104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,7168,0.04239893356959025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,7168,0.07689493497212728
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,512,6144,0.5179605484008789
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,512,8192,0.07068053086598715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,6144,0.03647146622339885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,6144,0.06798720359802246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,512,5120,0.4382677396138509
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,128,768,0.01334933340549469
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,5120,0.03129600087801616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,5120,0.05554133256276449
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,512,4096,0.355890146891276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,512,7168,0.06269226471583048
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,4096,0.025140267610549927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,4096,0.042827733357747394
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,512,3584,0.31448214848836265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,512,3584,0.035293865203857425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,3584,0.021640533208847047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,3584,0.03582186698913574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,512,3072,0.27325973510742185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,512,6144,0.05500906705856323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,3072,0.01881600022315979
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,3072,0.02927359938621521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,512,2560,0.23083839416503907
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,512,5120,0.04783573150634766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,2560,0.015947733322779337
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,2560,0.023829332987467446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,512,2048,0.1835114638010661
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,512,4096,0.03904213507970174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,2048,0.013460266590118408
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,2048,0.0187285323937734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,512,1536,0.12810880343119305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,512,3072,0.03067306677500407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,1536,0.011283199985822041
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,1536,0.015782399972279867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,512,1024,0.08441600004831949
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,512,2560,0.02577386697133382
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,1024,0.008787199854850769
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,1024,0.012645333011945089
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,512,768,0.0679136037826538
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,512,2048,0.021303466955820718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,768,0.007505066692829132
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,768,0.012714667121569314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,512,512,0.05206079880396525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,512,1536,0.016641066471735636
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,512,0.006277333199977875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,512,0.009511466821034749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,128,1024,0.016574933131535848
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,512,256,0.03471999963124593
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,256,0.005205333232879639
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,256,0.008496000369389852
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,512,128,0.028907734155654907
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,512,1024,0.013272533814112345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,128,0.004509866734345754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,128,0.007798400024573009
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,512,64,0.026654932896296186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,64,0.004084266722202301
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,64,0.007982933521270752
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,512,32,0.025872000058492023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,32,0.004275199770927429
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,512,768,0.011813333630561829
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,32,0.008038400113582611
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,512,512,0.010284800330797832
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,65536,0.22201919555664062
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,65536,0.607046381632487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,256,16384,1.3065364837646485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,256,65536,5.08897705078125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,16384,0.07248000303904215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,16384,0.15885653495788574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,256,12288,0.9963530858357748
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,512,256,0.008222933113574981
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,12288,0.05340480009714762
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,12288,0.12177600065867107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,512,128,0.007339733342329661
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,256,10240,0.07762026786804199
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,10240,0.044640000661214194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,256,10240,0.8390975952148437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,10240,0.10232106844584148
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,256,8192,0.6808106740315755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,8192,0.035905067125956217
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,8192,0.08237866560618082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,256,7168,0.5938688278198242
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,256,65536,0.4528181393941243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,7168,0.03356159925460815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,7168,0.07345600128173828
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,256,6144,0.5141791979471843
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,256,6144,0.04869759877522786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,6144,0.02676266630490621
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,6144,0.06382186810175577
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,256,5120,0.43376318613688153
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,256,16384,0.11739947001139324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,5120,0.024606933196385704
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,5120,0.05223893324534098
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,256,4096,0.35168425242106116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,256,12288,0.09032639662424723
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,4096,0.01990506649017334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,4096,0.039443198839823404
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,256,3584,0.31139411926269533
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,256,3584,0.03115306695302327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,256,8192,0.06177173455556234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,3584,0.017460266749064125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,3584,0.03208319942156474
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,256,3072,0.27018772761027016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,3072,0.013955199718475341
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,3072,0.025831466913223265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,256,2560,0.22881919542948403
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,256,7168,0.05500693321228027
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,2560,0.011475200454394024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,2560,0.020526933670043945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,256,2048,0.1804682731628418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,256,5120,0.04206506808598836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,2048,0.009948800007502239
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,2048,0.01768640081087748
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,256,1536,0.12822720209757488
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,256,4096,0.03488213221232096
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,1536,0.008610133330027263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,1536,0.014109866817792258
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,256,1024,0.08353493213653565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,256,3072,0.026842667659123735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,1024,0.006353066861629486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,1024,0.01184213360150655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,256,2560,0.02220906615257263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,256,768,0.06744000116984049
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,768,0.00565226674079895
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,768,0.011845333377520244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,256,512,0.05176639954249064
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,256,2048,0.018355200688044228
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,512,0.004955733319123586
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,512,0.009325866897900898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,256,256,0.03406080007553101
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,256,1536,0.014941866199175516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,256,0.004159999887148539
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,256,0.00830506682395935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,256,128,0.02846933404604594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,256,1024,0.011949867010116577
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,128,0.0037546666959921518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,128,0.00755626658598582
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,256,64,0.026245333751042682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,64,0.0034442665676275887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,64,0.0076885332663853955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,256,32,0.025656533241271973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,32,0.003651199986537298
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,32,0.007879466811815897
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,256,768,0.010315733154614766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,256,512,0.008861866593360902
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,128,65536,5.072767130533854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,65536,0.20541973114013673
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,256,256,0.007110400001207988
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,65536,0.606822395324707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,16384,0.061027201016743984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,16384,0.15627199808756512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,128,16384,1.3087679545084634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,256,128,0.006397866706053417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,128,12288,0.9938602447509766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,12288,0.04532693227132161
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,12288,0.11982080141703289
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,128,10240,0.8363199869791667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,10240,0.03444799979527791
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,10240,0.1004970630009969
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,128,8192,0.6786506652832032
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,128,8192,0.058899199962615965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,8192,0.02832319935162862
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,8192,0.08244907061258952
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,128,7168,0.5924138387044271
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,128,65536,0.4536896069844564
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,7168,0.028625067075093585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,7168,0.07249279816945395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,128,6144,0.5128896077473958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,128,16384,0.11359573205312093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,6144,0.022899200518925987
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,6144,0.06283520062764486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,128,5120,0.43280534744262694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,128,12288,0.08657279809316
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,5120,0.019308799505233766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,5120,0.051940266291300455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,128,4096,0.35149440765380857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,128,10240,0.07495679855346679
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,4096,0.01508799990018209
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,4096,0.03665706713994344
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,128,3584,0.31031573613484703
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,128,7168,0.051888000965118405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,3584,0.012938666343688964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,3584,0.028797866900761922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,128,3072,0.2697663942972819
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,128,6144,0.045848532517751055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,3072,0.011774933338165284
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,3072,0.022770132621129355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,128,2560,0.2285312016805013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,128,5120,0.0389514684677124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,2560,0.010175999999046326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,2560,0.019332265853881835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,128,2048,0.1805034637451172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,128,4096,0.03224640091260274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,2048,0.008669867118199667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,2048,0.01643519997596741
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,128,1536,0.12685866355895997
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,128,3072,0.024910932779312132
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,128,3584,0.028930133581161498
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,1536,0.007490133245786031
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,1536,0.014330666263898215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,128,1024,0.08260052998860677
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,1024,0.005763199925422668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,1024,0.011382399996121725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,128,768,0.06747306982676188
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,128,2560,0.020533333222071327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,768,0.005076266825199127
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,768,0.011478400230407715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,128,512,0.05151253143946329
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,128,2048,0.01699733336766561
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,512,0.004305066665013632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,512,0.009080533186594646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,128,256,0.03441386620203654
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,128,1024,0.01072746713956197
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,256,0.003770666569471359
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,256,0.008122666676839193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,128,128,0.0284768005212148
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,128,1536,0.013757866621017457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,128,0.003458133339881897
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,128,0.007463466624418895
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,128,64,0.025915733973185223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,64,0.003260799994071325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,64,0.007623466849327088
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,128,32,0.025227733453114826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,32,0.003509333233038584
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,32,0.007570133109887441
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,128,768,0.00965333382288615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,65536,0.19772693316141765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,64,65536,5.078058878580729
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,65536,0.6047669092814127
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,16384,0.05618666807810465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,16384,0.155949862798055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,64,16384,1.3091157277425132
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,12288,0.04307839870452881
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,64,12288,0.9920576095581055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,12288,0.12010880311330159
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,10240,0.036296534538269046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,10240,0.100437331199646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,64,10240,0.8364778518676758
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,8192,0.030728532870610552
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,64,8192,0.6795562744140625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,8192,0.08106239636739096
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,7168,0.027896533409754436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,7168,0.0719925324122111
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,64,7168,0.5923168182373046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,6144,0.02185493310292562
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,64,6144,0.5126218795776367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,6144,0.06272213459014893
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,5120,0.016691199938456216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,5120,0.05181653499603271
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,64,5120,0.43161598841349286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,4096,0.010243200262387593
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,64,4096,0.3519189198811849
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,4096,0.03627520004908244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,3584,0.008943999807039898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,3584,0.027718400955200194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,64,3584,0.31087465286254884
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,3072,0.007970133423805236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,3072,0.0225983997186025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,64,3072,0.2694517453511556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,2560,0.00802346666653951
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,64,2560,0.22849493026733397
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,2560,0.019067732493082683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,2048,0.007145600020885467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,64,2048,0.1806997299194336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,2048,0.01627626617749532
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,1536,0.006026666859785716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,64,1536,0.12678613662719726
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,1536,0.013798399766286214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,1024,0.004925866425037384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,64,1024,0.08269333044687907
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,1024,0.01134399970372518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,768,0.00439573327700297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,64,768,0.06742613315582276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,768,0.011272533734639486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,64,512,0.0513152003288269
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,512,0.004012800008058548
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,64,256,0.03367893298467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,512,0.008999466896057129
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,256,0.0036277333895365395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,256,0.008113066852092742
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,128,0.0033183999359607695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,64,128,0.028408533334732054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,128,0.007494399944941203
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,64,64,0.026158932844797773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,64,0.0031487998863061273
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,64,0.007484800120194752
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,64,32,0.025123200813929242
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,32,0.003221333275238673
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,32,0.007525333265463512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,65536,0.19076159795125325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,65536,0.6092117309570313
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,32,16384,1.3106048583984375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,16384,0.052496000130971276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,16384,0.15772159894307453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,32,12288,0.9941631952921549
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,12288,0.04158506790796916
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,12288,0.12009387016296387
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,32,65536,5.072562154134115
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,32,10240,0.8373621622721353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,10240,0.035176531473795576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,10240,0.09992640018463135
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,32,8192,0.6791221618652343
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,8192,0.029287467400232952
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,8192,0.08166826566060384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,32,7168,0.5922613143920898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,7168,0.02659519910812378
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,7168,0.07217600345611572
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,6144,0.026653865973154705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,6144,0.062295468648274746
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,32,6144,0.5133429209391276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,5120,0.01691306630770365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,5120,0.05162453254063924
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,32,5120,0.4318527857462565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,4096,0.009307733178138733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,32,4096,0.3517535845438639
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,4096,0.03624000151952107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,3584,0.008189866443475087
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,3584,0.02760106722513835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,32,3584,0.310972785949707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,3072,0.007122133175532024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,3072,0.02235306700070699
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,32,3072,0.2694133440653483
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,2560,0.006833066542943318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,32,2560,0.22813332875569664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,2560,0.018977065881093345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,2048,0.006134399771690368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,2048,0.016385066509246825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,32,2048,0.18040639559427898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,1536,0.005085866649945577
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,32,1536,0.12664213180541992
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,1536,0.014522666732470194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,1024,0.004363733530044556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,32,1024,0.08237226804097494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,1024,0.011691733201344808
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,768,0.004035199930270513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,32,768,0.06680213610331218
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,768,0.01172266701857249
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,512,0.0037237333754698435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,32,512,0.05129493474960327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,512,0.009050666292508443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,32,256,0.03373333215713501
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,256,0.0033749334514141084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,256,0.008056533336639405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,32,128,0.028104533751805622
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,128,0.003173333406448364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,128,0.007438933352629344
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,32,64,0.026178133487701417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,64,0.0029472000896930695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,64,0.007539199789365132
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4096,32,32,0.025141332546869916
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,32,0.003155199935038885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,32,0.007490133245786031
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,65536,16384,3.9289014180501303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,65536,16384,6.381190490722656
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,65536,16384,3.289454905192057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,65536,16384,3.4671541849772134
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,65536,12288,2.9061269124348956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,65536,12288,2.5094581604003907
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,65536,12288,3.2359583536783854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,65536,12288,4.681845092773438
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,65536,10240,2.473089090983073
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,65536,10240,4.077144622802734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,65536,10240,2.067286427815755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,65536,10240,2.0713301340738934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,65536,8192,1.9896874745686848
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,65536,8192,3.308562215169271
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,65536,8192,1.7599786122639973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,65536,7168,1.8110431671142577
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,65536,7168,1.4204811096191405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,65536,8192,1.617254384358724
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,65536,7168,1.420748774210612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,65536,7168,2.9455958048502606
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,65536,6144,1.2180309295654297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,65536,6144,1.5095167795817057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,65536,6144,1.2101749420166015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,65536,6144,2.4483286539713545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,65536,5120,1.250488535563151
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,65536,5120,1.0340725580851235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,65536,5120,1.0546293258666992
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,65536,5120,1.9889141082763673
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,65536,4096,1.0153077443440757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,65536,4096,1.5876309712727865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,65536,4096,0.7750837326049804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,65536,3584,0.8767701466878256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,65536,3584,0.7031274795532226
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,65536,3584,1.282317860921224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,65536,3584,0.7105706532796223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,65536,3072,0.7362602869669597
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,65536,3072,0.6115711847941081
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,65536,4096,0.7600959777832031
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,65536,3072,1.1951658884684244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,65536,3072,0.6194581349690755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,65536,2560,0.6142314910888672
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,65536,2560,0.9663338979085287
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,65536,2560,0.517029349009196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,65536,2048,0.48130985895792644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,65536,2048,0.40702505111694337
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,65536,2048,0.7542741139729817
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,65536,2048,0.41100587844848635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,65536,2560,0.4955626805623372
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,65536,1536,0.3675584157307943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,65536,1536,0.5604330698649089
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,65536,1536,0.30764268239339193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,65536,1024,0.2827989260355631
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,65536,1024,0.23573226928710939
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,65536,1024,0.38620586395263673
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,65536,1024,0.2246335983276367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,65536,768,0.22864960034688314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,65536,768,0.19461013476053873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,65536,768,0.30154240926106773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,65536,768,0.18944533665974933
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,65536,512,0.19249599774678547
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,65536,512,0.13942186037699383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,65536,512,0.22061759630839028
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,65536,512,0.16340053876241048
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,65536,256,0.16676799456278485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,65536,256,0.11589120229085285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,65536,256,0.14332373936971027
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,65536,256,0.13776426315307616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,65536,128,0.15634454091389974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,65536,128,0.11836053530375164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,65536,128,0.11777280171712239
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,65536,128,0.1312490701675415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,65536,64,0.158297602335612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,65536,64,0.11545706590016683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,65536,64,0.13440640767415363
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,65536,32,0.15865813891092936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,65536,32,0.11261546611785889
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,65536,1536,0.32053867975870765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,65536,32,0.13444159825642904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,16384,65536,3.4730880737304686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,16384,65536,6.137735493977865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,65536,6.32702382405599
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,65536,3.560110982259115
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,16384,16384,1.468493906656901
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,16384,1.4426016489664713
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,16384,0.8143338521321615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,16384,12288,1.1312981923421224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,16384,16384,0.7730047861735027
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,12288,1.094811757405599
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,12288,0.62205015818278
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,16384,10240,0.9527690887451172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,16384,12288,0.5728981018066406
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,10240,0.8758453369140625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,10240,0.5323861440022786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,16384,8192,0.755567995707194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,16384,10240,0.4829589207967122
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,8192,0.7071701049804687
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,8192,0.40404373804728194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,16384,7168,0.6781226476033528
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,16384,8192,0.38586133321126304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,7168,0.617955207824707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,7168,0.37237974802652996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,16384,6144,0.587498664855957
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,16384,7168,0.342305056254069
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,6144,0.5338517506917317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,6144,0.32463893890380857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,16384,5120,0.4880618731180827
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,16384,5120,0.24839146931966147
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,5120,0.44724585215250656
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,5120,0.2605013370513916
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,16384,4096,0.3887711842854818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,16384,6144,0.2916341463724772
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,4096,0.3568533261617025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,4096,0.21910932858784996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,16384,3584,0.33261226018269857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,16384,3584,0.17801599502563475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,3584,0.31467841466267904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,3584,0.18269972801208495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,16384,3072,0.282204786936442
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,16384,4096,0.2008789380391439
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,3072,0.27333332697550455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,3072,0.16114133199055988
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,16384,2560,0.2250805377960205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,16384,3072,0.15239146550496419
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,2560,0.22661333084106444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,2560,0.13667093912760417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,16384,2048,0.17845013936360676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,16384,2560,0.13230400085449218
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,2048,0.1852245330810547
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,2048,0.10719679991404216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,16384,1536,0.1407039960225423
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,16384,2048,0.10838080247243245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,1536,0.14080106417338054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,1536,0.08399360179901123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,16384,1024,0.10304640134175617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,16384,1536,0.08707093397776286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,1024,0.10011946360270182
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,1024,0.06292266845703125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,16384,768,0.0903434673945109
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,16384,768,0.05428373416264852
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,768,0.07907946904500326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,768,0.053220268090566
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,16384,512,0.07633919715881347
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,16384,1024,0.0642197330792745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,512,0.058050131797790526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,512,0.04810133377710978
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,16384,256,0.0655349334081014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,16384,512,0.041060264905293783
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,256,0.038228265444437665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,256,0.041682132085164386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,16384,128,0.06050560077031454
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,16384,256,0.03386666774749756
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,128,0.03310719927151998
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,128,0.03932906786600749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,16384,64,0.060908798376719156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,64,0.030370134115219116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,64,0.040643199284871416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,16384,32,0.06165973345438639
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,32,0.030906667311986286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,32,0.04054613510767619
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,16384,128,0.03309333324432373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,12288,65536,4.989969889322916
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,65536,4.720386250813802
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,65536,2.656920623779297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,12288,65536,2.5860565185546873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,12288,16384,1.253228759765625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,16384,1.0735883076985677
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,16384,0.6692458470662435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,12288,12288,0.9768853505452475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,12288,12288,0.4420949300130208
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,12288,0.8580106735229492
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,12288,0.547212791442871
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,12288,10240,0.8344928105672201
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,12288,10240,0.38268906275431314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,10240,0.65469970703125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,10240,0.40906454722086594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,12288,8192,0.6597663879394531
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,12288,8192,0.3010549227396647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,12288,16384,0.5813717524210612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,8192,0.5264447848002116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,8192,0.32107413609822594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,12288,7168,0.58024107615153
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,7168,0.462277348836263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,7168,0.28761920928955076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,12288,6144,0.5064117431640625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,12288,6144,0.2241663932800293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,6144,0.39796905517578124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,6144,0.25162240664164226
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,12288,5120,0.4225461324055989
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,12288,7168,0.2617258707682292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,5120,0.3362314542134603
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,5120,0.20501972834269205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,12288,4096,0.33555520375569664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,12288,5120,0.18825066884358724
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,4096,0.268233585357666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,4096,0.16819732983907063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,12288,3584,0.2872309366861979
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,12288,3584,0.1363200028737386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,3584,0.23643840154012047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,3584,0.1448106606801351
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,12288,3072,0.24064532915751138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,12288,4096,0.15416426658630372
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,3072,0.20576106707255043
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,3072,0.12427732944488526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,12288,2560,0.194649600982666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,12288,3072,0.11748586495717366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,2560,0.17110293706258137
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,2560,0.10580159823099773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,12288,2048,0.15450453758239746
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,12288,2048,0.08413973649342855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,2048,0.1389813264211019
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,2048,0.08458027044932047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,12288,1536,0.12203733126322429
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,12288,2560,0.10236053466796875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,1536,0.1082581361134847
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,1536,0.06530773242314657
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,12288,1024,0.09098986784617105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,12288,1536,0.06743146578470866
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,1024,0.07562239964803061
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,1024,0.04904106855392456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,12288,768,0.0790719985961914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,12288,1024,0.05076906681060791
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,768,0.06030826568603516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,768,0.04315199851989746
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,12288,512,0.06700692971547445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,12288,768,0.04297600189844768
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,512,0.044701866308848065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,512,0.03889706532160441
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,12288,256,0.05703039964040121
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,12288,512,0.03390613396962484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,256,0.030178133646647138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,256,0.03441813389460246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,12288,128,0.053447465101877846
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,12288,256,0.027500800291697186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,128,0.025891200701395674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,128,0.032264532645543416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,12288,64,0.053437864780426024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,64,0.023942399024963378
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,64,0.03330346743265788
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,12288,32,0.0534005324045817
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,32,0.02541973392168681
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,32,0.033370665709177655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,12288,128,0.02678080002466838
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,10240,65536,4.52218017578125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,65536,4.057145690917968
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,65536,2.297108205159505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,10240,16384,1.1591808319091796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,10240,65536,2.0502068837483725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,16384,0.8928746541341146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,16384,0.5334069569905598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,10240,12288,0.9111818949381509
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,10240,12288,0.3715146700541178
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,12288,0.663926378885905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,12288,0.4225599924723308
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,10240,10240,0.753053855895996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,10240,16384,0.49103145599365233
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,10240,0.56506773630778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,10240,0.3497354825337728
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,10240,8192,0.6005546569824218
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,10240,8192,0.25170772870381675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,8192,0.451800537109375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,8192,0.27256107330322266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,10240,7168,0.5356778462727865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,10240,10240,0.30845438639322914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,7168,0.3923541386922201
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,7168,0.24346240361531576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,10240,6144,0.4660650571187337
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,10240,6144,0.18948373794555665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,6144,0.33708906173706055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,6144,0.21104106903076172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,10240,5120,0.39103358586629233
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,10240,7168,0.22141013145446778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,5120,0.286408519744873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,5120,0.17218666076660155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,10240,4096,0.31181119283040365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,10240,5120,0.15932693481445312
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,4096,0.2278538703918457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,4096,0.14072853724161785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,10240,3584,0.26362773577372234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,10240,3584,0.11544640064239502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,3584,0.201417605082194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,3584,0.12336640357971192
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,10240,3072,0.22175787289937338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,10240,4096,0.13085866769154866
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,3072,0.17580159505208331
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,3072,0.10528533458709717
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,10240,2560,0.1802837371826172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,10240,3072,0.0999338706334432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,2560,0.14706560770670574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,2560,0.08943466345469156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,10240,2048,0.14160000483194987
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,10240,2560,0.08698346614837646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,2048,0.12152213255564373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,2048,0.07120640277862549
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,10240,1536,0.11214613119761149
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,10240,2048,0.0717408021291097
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,1536,0.09250880082448323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,1536,0.05597546497980753
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,10240,1024,0.08406399885813395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,10240,1536,0.05781439940134684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,1024,0.06585280100504556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,1024,0.04258879820505778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,10240,768,0.0732917308807373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,10240,1024,0.043687466780344644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,768,0.052086400985717776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,768,0.03745493491490682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,10240,512,0.06173226833343506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,10240,768,0.03738986651102702
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,512,0.0394538680712382
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,512,0.034244267145792644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,10240,256,0.053115733464558924
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,10240,512,0.03025813301404317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,256,0.02809813419977824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,256,0.02977173328399658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,10240,128,0.04946879943211873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,10240,128,0.022346667448679605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,10240,256,0.024338134129842124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,128,0.022420267264048256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,128,0.028500266869862872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,10240,64,0.04952213366826375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,64,0.019709867238998414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,64,0.029501867294311524
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,10240,32,0.04927146832148234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,32,0.020759467283884683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,32,0.029308799902598066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,8192,65536,4.094967397054036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,65536,3.1697174072265626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,65536,1.7666613260904949
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,8192,16384,1.0851946512858073
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,8192,65536,1.7394357045491535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,16384,0.7054933547973633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,16384,0.42620480855305987
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,8192,12288,0.8221408208211264
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,8192,12288,0.3266709327697754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,12288,0.5278016090393066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,12288,0.3364341417948405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,8192,10240,0.6848842620849609
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,8192,16384,0.39707520802815754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,10240,0.4361514727274577
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,10240,0.27462294896443684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,8192,8192,0.5538538614908854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,8192,8192,0.20410240491231285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,8192,0.3513717333475749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,8192,0.21896106402079263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,8192,7168,0.49449707667032883
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,8192,10240,0.25457599957784016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,7168,0.3117429415384928
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,7168,0.19476373990376789
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,8192,6144,0.4234005292256673
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,8192,7168,0.17967999776204427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,6144,0.2653813362121582
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,6144,0.17255892753601074
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,8192,5120,0.35649814605712893
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,8192,6144,0.1564341386159261
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,5120,0.22493759791056314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,5120,0.14154772758483886
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,8192,4096,0.28482879002889
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,8192,5120,0.13061439990997314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,4096,0.1797013282775879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,4096,0.1161845366160075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,8192,3584,0.24144426981608072
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,8192,4096,0.1076032002766927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,3584,0.16031999588012696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,3584,0.09986133575439453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,8192,3072,0.20115200678507486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,8192,3072,0.08330453236897786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,3072,0.13866880734761555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,3072,0.08519252936045328
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,8192,2560,0.1647029399871826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,8192,3584,0.09469546476999918
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,2560,0.11560959815979004
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,2560,0.07238186995188395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,8192,2048,0.12951253255208334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,8192,2048,0.05959999958674113
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,2048,0.09433920383453369
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,2048,0.058107733726501465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,8192,1536,0.10253760019938152
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,8192,2560,0.07177920341491699
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,1536,0.07348159948984781
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,1536,0.04601813157399495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,8192,1024,0.07709759871164958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,8192,1536,0.04803306659062703
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,1024,0.05179839928944906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,1024,0.03565226793289185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,8192,768,0.06712000370025635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,8192,1024,0.03755199909210205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,768,0.04194986820220947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,768,0.03166399995485942
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,8192,512,0.056600534915924074
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,8192,768,0.03194453318913777
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,512,0.03171839912732442
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,512,0.028569600979487103
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,8192,256,0.04808426698048909
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,8192,512,0.026572799682617186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,256,0.022087466716766358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,256,0.023838933308919272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,8192,128,0.04471360047658284
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,8192,256,0.01886720061302185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,128,0.017473065853118898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,128,0.021962666511535646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,8192,64,0.04325226545333862
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,64,0.014782933394114175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,64,0.022453333934148154
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,8192,32,0.042165335019429526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,32,0.015523200233777365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,32,0.02286720077196757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,7168,65536,3.8962773640950523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,8192,128,0.01648853321870168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,65536,2.5184160868326826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,65536,1.6015188852945965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,7168,16384,1.016261355082194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,7168,65536,1.7918474833170575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,16384,0.6149589538574218
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,16384,0.3950570742289225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,7168,12288,0.7735616048177083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,7168,16384,0.4075605392456055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,12288,0.46533867518107097
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,12288,0.2967594782511393
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,7168,10240,0.6554976145426432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,7168,12288,0.30514774322509763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,10240,0.38429333368937174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,10240,0.25517120361328127
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,7168,8192,0.5278421401977539
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,7168,8192,0.20881919860839843
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,8192,0.3084266662597656
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,8192,0.1972490628560384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,7168,7168,0.46434774398803713
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,7168,7168,0.18324480056762696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,7168,0.2709727923075358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,7168,0.17333226203918456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,7168,6144,0.4067263921101888
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,7168,10240,0.2589322725931803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,6144,0.23591252962748208
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,6144,0.14912320772806803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,7168,5120,0.3397706667582194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,7168,6144,0.1576032002766927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,5120,0.1960426648457845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,5120,0.12670079867045086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,7168,4096,0.2728234608968099
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,7168,5120,0.13363733291625976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,4096,0.16008426348368326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,4096,0.10230399767557781
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,7168,3584,0.22996160189310708
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,7168,4096,0.10878079732259113
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,3584,0.13852906227111816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,3584,0.09010346730550131
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,7168,3072,0.1929418722788493
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,7168,3584,0.09948693116505941
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,3072,0.11948373317718505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,3072,0.07772160371144612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,7168,2560,0.15722346305847168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,7168,2560,0.07403199672698975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,2560,0.10003413359324138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,2560,0.06424426635106405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,7168,2048,0.12340479691823322
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,7168,3072,0.0849951982498169
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,2048,0.08122560183207193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,2048,0.05243093172709147
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,7168,1536,0.09900053342183432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,7168,2048,0.06152106523513794
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,1536,0.061210668087005614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,1536,0.04220266739527385
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,7168,1024,0.07454506556193033
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,7168,1536,0.04952213366826375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,1024,0.04314560095469157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,1024,0.032526934146881105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,7168,768,0.06471360127131144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,7168,1024,0.039758932590484616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,768,0.03384960095087687
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,768,0.029127466678619384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,7168,512,0.05426559845606486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,7168,768,0.03228800098101298
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,512,0.024833067258199056
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,512,0.02564053336779277
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,7168,256,0.04581973155339559
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,7168,512,0.026689066489537554
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,256,0.016860800981521606
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,256,0.02175359924634298
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,7168,128,0.040703999996185306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,7168,256,0.020475733280181884
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,128,0.014254933595657349
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,128,0.019497599204381308
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,7168,64,0.03995946645736694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,64,0.013412266969680786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,64,0.020062933365503945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,7168,32,0.03737066586812337
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,32,0.014306132992108664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,32,0.019824000199635823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,7168,128,0.01796906590461731
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,6144,65536,1.219797388712565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,6144,65536,3.681708780924479
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,65536,1.4030634562174478
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,65536,2.3228212992350263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,6144,16384,0.35021654764811194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,6144,16384,0.9592085520426432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,16384,0.34285014470418296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,16384,0.536514155069987
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,6144,12288,0.7259679794311523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,12288,0.39558080037434895
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,12288,0.25874560674031577
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,6144,10240,0.6163839975992839
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,6144,12288,0.23253973325093588
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,10240,0.32846399943033855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,10240,0.2229834715525309
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,6144,8192,0.49646612803141277
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,6144,8192,0.15859200159708658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,8192,0.2642624060312907
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,6144,10240,0.1973632017771403
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,8192,0.16931093533833821
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,6144,7168,0.44048105875651045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,7168,0.234115203221639
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,7168,0.14990720748901368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,6144,6144,0.38531306584676106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,6144,7168,0.1398037274678548
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,6144,0.1997109254201253
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,6144,0.13240426381429035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,6144,5120,0.3237408002217611
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,6144,6144,0.12201066811879475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,5120,0.16942186355590821
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,5120,0.11069653034210206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,6144,4096,0.2612544059753418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,6144,5120,0.1021888017654419
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,4096,0.13593494097391765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,4096,0.08845866521199544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,6144,3584,0.21999893188476563
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,6144,3584,0.07463786602020264
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,3584,0.1198367993036906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,6144,4096,0.08414080142974853
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,3584,0.07785066763559977
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,6144,3072,0.06492160161336263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,6144,3072,0.18305813471476237
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,3072,0.10437333583831787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,3072,0.06732800006866455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,6144,2560,0.148251740137736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,2560,0.08784533341725667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,2560,0.057061334451039634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,6144,2048,0.11766826311747233
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,6144,2048,0.048069334030151366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,2048,0.07194026311238608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,2048,0.04638506571451823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,6144,1536,0.0942570686340332
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,6144,2560,0.05683199961980184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,1536,0.05605440139770508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,1536,0.03738666772842407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,6144,1024,0.07088426748911539
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,6144,1536,0.03886186679204305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,1024,0.040030932426452635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,1024,0.02955519954363505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,6144,768,0.06199893156687418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,6144,1024,0.02983786662419637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,768,0.03253440062204997
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,768,0.025612799326578777
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,6144,512,0.051582932472229004
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,6144,768,0.025006934007008867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,512,0.024954666694005333
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,512,0.022525866826375328
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,6144,256,0.043841067949930826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,6144,256,0.01478506624698639
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,256,0.017772799730300902
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,256,0.019377066691716512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,6144,128,0.03821546634038289
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,6144,128,0.013412266969680786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,128,0.013524267077445983
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,128,0.016833066940307617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,6144,64,0.03659520149230957
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,64,0.01213653286298116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,64,0.017614932854970296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,6144,32,0.03487573464711507
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,32,0.012518399953842163
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,32,0.017517866690953572
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,6144,512,0.01965120037396749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,5120,65536,3.490210215250651
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,65536,1.9101696014404297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,65536,1.1768224080403646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,5120,16384,0.910744539896647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,5120,16384,0.3182602564493815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,16384,0.4387317339579265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,16384,0.2861824035644531
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,5120,12288,0.6979040145874024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,5120,65536,1.2854026794433593
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,12288,0.3339647928873698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,12288,0.21938346227010092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,5120,10240,0.5802549362182617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,5120,12288,0.2350719928741455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,10240,0.27628587086995443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,10240,0.18679572741190592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,5120,8192,0.4705034573872884
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,5120,10240,0.19993599255879718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,8192,0.22449280420939127
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,8192,0.14350825945536297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,5120,7168,0.4202101389567058
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,5120,8192,0.16042879422505696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,7168,0.19475626945495605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,7168,0.1275274674097697
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,5120,6144,0.3659903844197591
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,5120,6144,0.1216096003850301
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,6144,0.16805334091186525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,6144,0.11410026550292969
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,5120,5120,0.3086933453877767
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,5120,5120,0.10303893089294433
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,5120,0.14121813774108888
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,5120,0.09546773433685303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,5120,4096,0.24639360109965006
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,5120,7168,0.1429759979248047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,4096,0.11529706319173176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,4096,0.07571199735005697
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,5120,3584,0.20844799677530923
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,5120,4096,0.08520853519439697
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,3584,0.1004640022913615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,3584,0.0672714630762736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,5120,3072,0.17272106806437176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,5120,3584,0.07772693634033204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,3072,0.08811413447062175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,3072,0.057547732194264734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,5120,2560,0.13957014083862304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,5120,3072,0.06694826285044352
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,2560,0.07357439994812012
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,2560,0.0497653325398763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,5120,2048,0.11094506581624348
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,5120,2560,0.059374932448069254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,2048,0.060653865337371826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,2048,0.04009386698404948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,5120,1536,0.0887658675511678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,5120,2048,0.04910720189412435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,1536,0.04661973317464192
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,1536,0.032792532444000246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,5120,1024,0.0677845319112142
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,5120,1024,0.031218133370081586
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,1024,0.033164799213409424
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,1024,0.025290666023890178
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,5120,768,0.058645331859588624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,5120,768,0.025131734212239583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,768,0.026260266701380413
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,5120,1536,0.0397546648979187
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,768,0.022376533349355063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,5120,512,0.044990932941436766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,512,0.01992959976196289
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,512,0.01918399930000305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,5120,256,0.03689066569010417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,5120,512,0.02067199945449829
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,256,0.013843199610710144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,5120,256,0.015715199708938598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,256,0.01635199983914693
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,5120,128,0.03429119984308879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,5120,128,0.014427733421325684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,128,0.010745599865913391
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,128,0.01495146652062734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,5120,64,0.03290346662203471
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,64,0.009269332885742188
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,64,0.015643733739852905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,5120,32,0.032414933045705155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,32,0.009578667084376017
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,32,0.015621333320935567
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,4096,65536,3.262159983317057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,4096,65536,0.8636053085327149
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,65536,1.478490702311198
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,65536,1.0231221516927085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,4096,16384,0.23993919690450033
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,4096,16384,0.8503199895222983
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,16384,0.3450144131978353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,16384,0.23609493573506674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,4096,12288,0.6553205490112305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,12288,0.2596522649129232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,12288,0.18538133303324383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,4096,10240,0.5514432271321614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,4096,10240,0.14043092727661133
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,10240,0.21761493682861327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,10240,0.15537066459655763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,4096,8192,0.44208641052246095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,4096,8192,0.11392319997151693
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,8192,0.17538560231526693
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,8192,0.12259413401285808
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,4096,7168,0.39905920028686526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,4096,7168,0.10064746538798015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,7168,0.1541312058766683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,7168,0.10742080211639404
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,4096,6144,0.3479562759399414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,4096,12288,0.16711254119873048
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,6144,0.1342037359873454
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,6144,0.09335253238677979
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,4096,5120,0.2915957450866699
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,4096,6144,0.08692479928334554
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,5120,0.11111573378245036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,5120,0.07959893544514975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,4096,4096,0.23369174003601073
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,4096,5120,0.07435306708017984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,4096,0.0908789316813151
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,4096,0.06248213450113932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,4096,3584,0.19594987233479816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,4096,4096,0.06109866698582968
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,3584,0.07962453365325928
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,3584,0.05588693221410116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,4096,3072,0.16443840662638348
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,4096,3072,0.04802240133285522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,3072,0.06890026728312174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,3072,0.04837119976679484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,4096,2560,0.13124159971872967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,4096,3584,0.05551573435465494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,2560,0.0587445338567098
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,2560,0.04190613428751628
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,4096,2048,0.10375146865844727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,4096,2048,0.03546346823374431
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,2048,0.04780373175938924
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,2048,0.03410666783650716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,4096,1536,0.08400639692942301
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,4096,2560,0.042124799887339276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,1536,0.03652373154958089
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,1536,0.02719680070877075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,4096,1024,0.0634112000465393
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,4096,1024,0.0215829332669576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,1024,0.025758934020996094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,1024,0.021104000012079873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,4096,768,0.05571200052897135
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,4096,1536,0.02789013385772705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,768,0.020968532562255858
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,768,0.018705066045125326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,4096,512,0.04255786736806234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,4096,768,0.01818559964497884
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,512,0.01625706652800242
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,512,0.01665386656920115
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,4096,256,0.03495253324508667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,4096,512,0.014983466267585755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,256,0.011633066336313884
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,256,0.014434132973353067
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,4096,128,0.031225599845250446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,4096,256,0.01190826694170634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,128,0.009115733702977498
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,128,0.013238400220870972
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,4096,64,0.03091946641604106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,64,0.007931733131408691
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,64,0.013814399639765421
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,4096,32,0.03057066599527995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,32,0.008167466521263123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,32,0.01388800044854482
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,4096,128,0.011062399546305338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3584,65536,3.2375862121582033
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,65536,1.2985365549723307
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,65536,1.001698112487793
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3584,16384,0.8450784047444662
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3584,16384,0.22749120394388833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,16384,0.3289141337076823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,16384,0.23737813631693522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3584,12288,0.6518496195475261
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3584,12288,0.1706816037495931
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,12288,0.24804479281107583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,12288,0.17636159261067708
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3584,10240,0.5464384078979492
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3584,65536,0.8883562723795573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3584,10240,0.1555445353190104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,10240,0.21042346954345703
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,10240,0.1514645258585612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3584,8192,0.4422143936157227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,8192,0.16659626960754395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,8192,0.11863893667856854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3584,7168,0.39594987233479817
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3584,8192,0.1170464038848877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,7168,0.14833173751831055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,7168,0.10322879950205485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3584,6144,0.3451946576436361
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3584,7168,0.10273706912994385
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,6144,0.1263797362645467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,6144,0.09167359670003256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3584,5120,0.2917269388834635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3584,6144,0.08922879695892334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,5120,0.10755306879679363
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,5120,0.07711893717447917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3584,4096,0.23024746576944985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3584,5120,0.07539520263671876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,4096,0.08601919809977213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,4096,0.06237333218256632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3584,3584,0.1953312079111735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3584,4096,0.06283413171768189
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,3584,0.07691520055135091
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,3584,0.054103465874989834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3584,3072,0.16286080678304035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3584,3584,0.05682986577351888
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,3072,0.06572693188985189
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,3072,0.04745066563288371
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3584,2560,0.13265066941579182
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3584,2560,0.04337706565856934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,2560,0.055308798948923744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,2560,0.04038399855295817
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3584,2048,0.10201386610666911
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,128,512,0.010782933235168457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,2048,0.04529386758804321
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,2048,0.03285333315531413
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3584,1536,0.08397973378499349
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3584,3072,0.04984106620152791
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,1536,0.03486293156941732
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,1536,0.026395734151204425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3584,1024,0.06350186665852865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3584,2048,0.03576960166295369
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,1024,0.02547946572303772
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,1024,0.020986666282018028
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3584,768,0.05542613267898559
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3584,1536,0.029397332668304445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,768,0.020901334285736085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,768,0.018900267283121743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3584,512,0.04209173520406087
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3584,1024,0.023496532440185548
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,512,0.01618666648864746
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,512,0.01648853321870168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3584,256,0.034593065579732254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3584,768,0.019423999389012656
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,256,0.01154026687145233
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,256,0.014360533157984415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3584,128,0.031269333759943646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3584,128,0.01183573305606842
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,128,0.009211732943852743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,128,0.013114666938781739
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3584,64,0.030962133407592775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,64,0.007998933394749958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,64,0.013424000144004822
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3584,32,0.03059413234392802
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,32,0.008182399968306223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,32,0.013571199774742127
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3072,65536,3.0784639994303387
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3072,65536,0.6929269154866536
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3584,512,0.016267733772595723
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,65536,1.0766826629638673
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,65536,0.8131093343098958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3072,16384,0.8124021530151367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,16384,0.27059412002563477
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,16384,0.19219093322753905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3584,256,0.012730666995048523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3072,12288,0.615221341451009
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,12288,0.20375466346740723
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,128,512,0.008266666531562805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,12288,0.14644266764322916
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3072,10240,0.11155412991841633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3072,10240,0.5183552106221516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,10240,0.17065812746683756
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,10240,0.12302827040354411
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3072,8192,0.4205898602803548
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,8192,0.1371829350789388
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,8192,0.09763519763946533
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3072,7168,0.3787893295288086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3072,7168,0.08022826512654622
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,7168,0.12046826680501302
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,7168,0.08560319741566977
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3072,6144,0.3300053278605143
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3072,16384,0.17277119954427084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,6144,0.10293760299682617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,6144,0.07589759826660156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3072,5120,0.27699305216471354
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3072,12288,0.13262826601664227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,5120,0.08807466824849447
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3072,8192,0.09098133246103922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,5120,0.06332373221715291
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3072,4096,0.22036266326904297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3072,6144,0.06995733578999838
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,4096,0.07065707047780355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,4096,0.05026559829711914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3072,3584,0.04476906855901082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3072,3584,0.18641600608825684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,3584,0.06208320061365763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,3584,0.04485546747843425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3072,3072,0.15605546633402506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,3072,0.05369919935862223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,3072,0.03928639888763428
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3072,2560,0.12613013585408528
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3072,2560,0.03365226586659749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,2560,0.04511679808298747
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,2560,0.03419306675593058
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3072,2048,0.09879466692606607
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3072,5120,0.06002560059229532
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,2048,0.03628480037053426
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,2048,0.02725013295809428
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3072,1536,0.07913493315378825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3072,4096,0.049709868431091306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,1536,0.02831466595331828
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,1536,0.021991467475891112
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3072,1024,0.06051413218180338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3072,3072,0.03926506837209066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,1024,0.020441599686940513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,1024,0.017864533265431724
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3072,768,0.05020800034205118
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3072,2048,0.028500266869862872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,768,0.01667733391125997
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,768,0.016361600160598753
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3072,512,0.04008106787999471
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3072,1536,0.023540266354878745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,512,0.013180800278981528
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,512,0.0142794668674469
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3072,256,0.03249066670735677
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3072,1024,0.01842986742655436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,256,0.009179733196894328
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,256,0.012501333157221475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3072,128,0.029676800966262816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3072,768,0.016059733430544534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,128,0.007420800129572551
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,128,0.011633066336313884
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3072,64,0.029330132404963176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,64,0.006513066589832306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,64,0.012001066406567892
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,3072,32,0.028880000114440918
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,32,0.006875733534495036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,32,0.011975466211636861
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3072,512,0.013648000359535218
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3072,256,0.010950400431950887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2560,65536,3.038923645019531
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,65536,0.8589749018351236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,65536,0.7825418472290039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2560,16384,0.7984042485555013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,16384,0.21404693921407064
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3072,128,0.009799466530481974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,16384,0.18383466402689616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2560,12288,0.6070453643798828
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,12288,0.16461119651794434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,12288,0.14027199745178223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2560,10240,0.5133674621582032
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2560,65536,0.6163434982299805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,10240,0.13764373461405438
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,10240,0.12028586864471436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2560,8192,0.41813974380493163
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2560,16384,0.15597012837727864
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,8192,0.11117013295491536
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,8192,0.09321599801381429
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2560,7168,0.37382933298746746
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2560,12288,0.11710506280263264
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,7168,0.09595519701639811
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,7168,0.08041706879933676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2560,6144,0.32465813954671224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2560,10240,0.0998207966486613
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,6144,0.08469440142313639
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,6144,0.07221546967824301
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2560,5120,0.2754314740498861
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2560,8192,0.08153279622395834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,5120,0.07052053610483805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,5120,0.06179733276367187
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2560,4096,0.21886080106099448
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2560,4096,0.044513066609700516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,4096,0.05734613339106241
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,4096,0.048027733961741134
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2560,3584,0.18522027333577473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2560,7168,0.07194773356119791
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,3584,0.05114666620890299
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,3584,0.04315413236618042
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2560,3072,0.15415679613749186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2560,6144,0.06291413307189941
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,3072,0.04358293215433757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,3072,0.03808426856994629
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2560,2560,0.12491412957509358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2560,5120,0.05380266507466634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,2560,0.03748906850814819
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,2560,0.03188693324724833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2560,2048,0.09850133260091146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2560,3584,0.04032426675160726
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,2048,0.030243200063705445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,2048,0.02648106614748637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2560,1536,0.0787285327911377
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2560,3072,0.03532586495081584
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,1536,0.023798400163650514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,1536,0.021668267250061036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2560,1024,0.06023893356323242
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2560,2560,0.030100266138712566
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,1024,0.017618133624394735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,1024,0.017548799514770508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2560,768,0.051113601525624594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2560,2048,0.025434666872024538
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,768,0.014633599917093912
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,768,0.016017066438992818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2560,512,0.03976320028305054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2560,512,0.012582400441169738
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,512,0.011618133385976155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,512,0.014110933740933737
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2560,256,0.032631466786066696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2560,1536,0.0211082657178243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,256,0.008567466338475545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,256,0.01225920021533966
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2560,128,0.029502934217453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2560,128,0.009171199798583985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,128,0.007101866602897644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,128,0.01149120032787323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2560,64,0.02898453275362651
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,64,0.006252799928188324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,64,0.011748266220092774
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2560,32,0.028988800446192425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,32,0.006482133269309997
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,32,0.011902933319409687
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2560,1024,0.01665493349234263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2560,768,0.014940800269444785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,65536,0.6709877649943035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,65536,0.6342506408691406
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2048,65536,2.8985761006673174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2048,16384,0.7596437454223632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,16384,0.1713258743286133
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,16384,0.14532267252604167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2048,12288,0.5805909474690755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2560,256,0.010294399658838908
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,12288,0.1314570665359497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,12288,0.10954773426055908
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2048,10240,0.48956158955891926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2048,65536,0.5107968012491862
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,10240,0.11010986963907879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,10240,0.09514453411102294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2048,8192,0.39567041397094727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2048,16384,0.13011093139648439
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,8192,0.08808746337890624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,8192,0.07187840143839518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2048,7168,0.3568810780843099
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2048,12288,0.09987413088480632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,7168,0.07682240009307861
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,7168,0.06415893236796061
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2048,6144,0.31177921295166017
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2048,10240,0.08506240049997965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,6144,0.06785386403401693
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,6144,0.05742293198903402
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2048,5120,0.2626826604207357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2048,8192,0.06966400146484375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,5120,0.057076267401377355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,5120,0.04846719900767009
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2048,4096,0.20693546930948892
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2048,7168,0.06172266801198324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,4096,0.04573440154393514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,4096,0.037674665451049805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2048,3584,0.17524693806966146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2048,6144,0.054826664924621585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,3584,0.04020373423894246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,3584,0.03383786678314209
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2048,3072,0.14709332784016926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2048,5120,0.04709759950637817
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,3072,0.03511893351872762
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,3072,0.029435733954111736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2048,2560,0.11921066443125408
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2048,4096,0.03860053221384684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,2560,0.02948586742083232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,2560,0.025594667593638105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2048,2048,0.09261013666788737
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2048,3584,0.033921066919962564
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,2048,0.02405653397242228
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,2048,0.020874667167663574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2048,1536,0.07508906523386637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2048,3072,0.029847466945648195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,1536,0.01962560017903646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,1536,0.017734400431315103
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2048,1024,0.056900266806284586
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2048,2560,0.026296534140904743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,1024,0.014751999576886495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,1024,0.014842666188875833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2048,768,0.04800320068995158
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2048,2048,0.022000000874201456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,768,0.012421333789825439
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,768,0.013271466890970866
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2048,512,0.036985599994659425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2048,1536,0.018289067347844443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,512,0.009754666686058044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2048,1024,0.015032533804575601
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,512,0.011866666873296102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2048,256,0.03054933349291484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2048,768,0.013448533415794373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,256,0.007411199808120728
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,256,0.010520533720652262
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2048,128,0.027900799115498857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2048,128,0.008719999591509502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,128,0.006183466811974844
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,128,0.00992746651172638
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2048,64,0.027302400271097822
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,64,0.005496533215045929
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,64,0.010318932930628459
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,2048,32,0.02720959981282552
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,32,0.005725866556167603
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,32,0.01036906639734904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1536,65536,0.4175039927164714
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,65536,0.5307637214660644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,65536,0.6128821055094401
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1536,65536,2.8513216654459637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1536,16384,0.7406869252522786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,16384,0.13379093805948894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,16384,0.13836159706115722
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2048,512,0.011271466811498005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1536,12288,0.5710965474446614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,12288,0.10162560145060222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,12288,0.10231253306070964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1536,10240,0.4832586606343587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2048,256,0.009709866841634114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,10240,0.08421866893768311
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,10240,0.08866346677144368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1536,8192,0.3911456108093262
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1536,16384,0.10859626928965252
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,8192,0.06952213446299235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,8192,0.06846720377604167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1536,7168,0.35305598576863606
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1536,12288,0.0825930674870809
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,7168,0.060754132270812986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,7168,0.06172266801198324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1536,6144,0.30743678410847985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1536,10240,0.06997866630554199
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,6144,0.05219626824061076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,6144,0.054946132500966395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1536,5120,0.2601045290629069
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1536,5120,0.038701868057250975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,5120,0.043884801864624026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,5120,0.04766186475753784
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1536,8192,0.057843200365702306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1536,4096,0.2061631997426351
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,4096,0.035626665751139326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,4096,0.03636053403218587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1536,3584,0.17404266993204753
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1536,7168,0.05235199928283692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,3584,0.03144960006078084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,3584,0.03247039914131165
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1536,3072,0.14547200202941896
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1536,3072,0.025050665934880572
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,3072,0.026850134134292603
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,3072,0.02847040096918742
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1536,6144,0.04556800127029419
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1536,2560,0.11844693024953205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,2560,0.023086933294932048
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,2560,0.024943999449412026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1536,2048,0.09332266648610434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1536,2048,0.019166932503382365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,2048,0.019230933984120686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,2048,0.02127573291460673
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1536,1536,0.07490666707356772
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1536,1536,0.01597760021686554
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,1536,0.015518933534622192
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,1536,0.017745065689086913
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1536,1024,0.05673066775004069
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1536,4096,0.03209813237190247
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,1024,0.011657599608103435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,1024,0.014510933558146158
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1536,768,0.04893120129903157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1536,3584,0.028410667181015016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,768,0.009842133522033692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,768,0.013396267096201578
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1536,512,0.0372320016225179
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1536,512,0.010339200496673584
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,512,0.008010666569073994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,512,0.011918933192888895
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1536,256,0.03074986735979716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1536,2560,0.021955200036366782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,256,0.006257066627343495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,256,0.010614400108655293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1536,128,0.02810773253440857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1536,1024,0.01330880026022593
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,128,0.00528959979613622
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,128,0.010054399569829304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1536,64,0.02720853288968404
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,64,0.004692266881465912
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1536,768,0.011913599570592244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,64,0.010225066542625427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1536,32,0.02684906721115112
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,32,0.004929066697756449
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,32,0.010249599814414978
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1536,256,0.008754133184750875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,65536,0.3555861473083496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1024,65536,2.683355712890625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,65536,0.4256095886230469
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1024,16384,0.7049823760986328
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1024,16384,0.09101653099060059
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,16384,0.09119786421457926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,16384,0.09876053333282471
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1024,12288,0.07561493714650472
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1024,12288,0.5402816136678059
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,12288,0.06849813461303711
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,12288,0.07270507017771402
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1024,10240,0.45808000564575196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,10240,0.057384534676869714
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,10240,0.06477440198262532
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1024,8192,0.3731754620869955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1536,128,0.00798399994770686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,8192,0.046828798453013104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,8192,0.05043733517328898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1024,7168,0.3370154698689779
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1024,7168,0.044072532653808595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,7168,0.04153279860814412
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,7168,0.0452128012975057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1024,6144,0.2921536127726237
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1024,65536,0.3618314743041992
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,6144,0.03525333404541016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,6144,0.04010239839553833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1024,5120,0.2482250690460205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1024,10240,0.06062186559041342
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,5120,0.030572799841562907
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,5120,0.03430399894714355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1024,4096,0.19678293863932292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1024,4096,0.026655999819437663
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,4096,0.023676800727844238
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,4096,0.026286933819452924
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1024,3584,0.1666613260904948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1024,3584,0.023795199394226075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1024,8192,0.049746131896972655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,3584,0.0210698664188385
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,3584,0.02418986757596334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1024,3072,0.1390794595082601
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,3072,0.018651733795801796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,3072,0.02132693330446879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1024,2560,0.11237226327260334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1024,2560,0.018245333433151247
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,2560,0.016149333119392394
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,2560,0.01872106591860453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1024,2048,0.08867626984914144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1024,6144,0.03918933471043905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,2048,0.013416533668835958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,2048,0.01629973351955414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1024,1536,0.07153920332590738
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1024,1536,0.013659733533859252
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,1536,0.011090133587519329
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,1536,0.014089600245157877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1024,1024,0.05436906814575195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1024,5120,0.033297065893809
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,1024,0.008674133817354839
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,1024,0.011959466338157653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1024,768,0.04563200076421102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1024,3072,0.021074134111404418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,768,0.0075082664688428235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,768,0.011532800396283467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1024,512,0.03633813460667928
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1024,2048,0.015794133146603904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,512,0.006265600025653839
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,512,0.010098133484522502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1024,256,0.028857600688934327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1024,256,0.007816533247629803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,256,0.00513919989267985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,256,0.00916373332341512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1024,128,0.026551467180252076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1024,128,0.007031466563542683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,128,0.00444160004456838
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,128,0.008646399776140849
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1024,64,0.0255786657333374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,64,0.004099199920892716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,64,0.008804266651471455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,1024,32,0.025272534290949507
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,32,0.004274133344491323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,32,0.008810666203498841
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1024,1024,0.011386666695276897
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1024,768,0.010385066270828247
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,65536,0.27363627751668296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,768,65536,2.6634676615397135
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,65536,0.3922389348347982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,768,16384,0.08964479764302571
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,16384,0.07145919799804687
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,768,16384,0.6976320266723632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,16384,0.08861546516418457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,768,12288,0.5365461349487305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,12288,0.05603520075480143
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,12288,0.06694186528523763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,768,10240,0.4532586733500163
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,768,10240,0.05957333246866862
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,10240,0.04604586760203044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,10240,0.05776106516520182
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,768,8192,0.3683765411376953
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1024,512,0.009142399827639262
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,8192,0.038328532377878824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,8192,0.046697600682576494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,768,7168,0.33441492716471355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,768,7168,0.043900799751281736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,7168,0.039817599455515544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,7168,0.04259413480758667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,768,6144,0.2913279851277669
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,768,65536,0.33782294591267903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,6144,0.02895039916038513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,6144,0.037393065293629964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,768,5120,0.2458869298299154
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,768,12288,0.06853226820627847
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,5120,0.028519467512766523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,5120,0.03156053423881531
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,768,4096,0.19487679799397786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,768,8192,0.0482474684715271
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,4096,0.019287467002868652
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,4096,0.023847466707229613
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,768,3584,0.16442346572875977
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,768,6144,0.038389333089192706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,3584,0.01775253415107727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,3584,0.02208426594734192
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,768,3072,0.1385536034901937
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,768,3072,0.020435200134913126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,3072,0.017089066902796428
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,3072,0.019898666938145956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,768,5120,0.0319541335105896
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,768,2560,0.11207679907480876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,768,2560,0.017871999740600587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,2560,0.014703999956448874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,2560,0.01731520096460978
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,768,2048,0.08769280115763346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,2048,0.01128000020980835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,2048,0.015727999806404113
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,768,1536,0.071615997950236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,768,1536,0.013943466544151305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,1536,0.009508267045021057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,1536,0.013765333096186319
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,768,1024,0.05425279935201009
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,768,4096,0.025333333015441894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,1024,0.007446399827798207
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,1024,0.011710932850837708
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,768,768,0.045915734767913816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,768,3584,0.022762666145960488
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,768,0.006499200065930684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,768,0.010961066683133442
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,768,512,0.03543039957682292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,768,512,0.008504533767700195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,512,0.005569066603978475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,512,0.009991466999053955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,768,256,0.029167999823888142
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,768,2048,0.015398400028546652
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,256,0.004625066618124644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,256,0.009165866176287334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,768,128,0.026294400294621784
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,768,128,0.006602666775385539
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,128,0.004105599969625473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,128,0.008583466211954754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,768,64,0.02539413372675578
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,64,0.0038005332152048744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,64,0.008682666222254436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,768,1024,0.011224533120791119
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,768,32,0.024977066119511924
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,32,0.003953066716591517
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,768,768,0.009949866930643718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,32,0.008787199854850769
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,65536,0.19363519350687664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,65536,0.3753120104471842
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,512,16384,0.6902634938557942
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,768,256,0.007249066730340321
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,512,65536,2.6275563557942707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,16384,0.05447359879811605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,16384,0.08749866485595703
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,512,12288,0.5295029322306315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,12288,0.050220799446105954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,12288,0.06273386478424073
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,512,10240,0.4479061444600423
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,512,65536,0.2933717409769694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,10240,0.04319039980570476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,10240,0.05501653353373209
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,512,8192,0.36584107081095374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,512,16384,0.0766325314839681
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,8192,0.032935466368993124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,8192,0.04437013467152913
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,512,7168,0.33122666676839196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,512,12288,0.06088320016860962
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,7168,0.03259306748708089
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,7168,0.0389792005221049
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,512,6144,0.28666133880615235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,512,10240,0.05254826545715332
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,6144,0.02490133245786031
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,6144,0.03514773448308309
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,512,5120,0.24386560122172035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,512,8192,0.042173866430918375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,5120,0.023973333835601806
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,5120,0.028663466374079388
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,512,4096,0.19382507006327312
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,512,7168,0.03754986524581909
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,4096,0.016500266393025716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,4096,0.02269866665204366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,512,3584,0.1639893372853597
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,512,6144,0.033291733264923094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,3584,0.01588373382886251
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,512,5120,0.027850667635599773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,3584,0.02062079906463623
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,512,3072,0.13672000567118328
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,3072,0.014507733782132468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,3072,0.018075732390085857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,512,2560,0.11035733222961426
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,512,2560,0.016004266341527303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,2560,0.01225386659304301
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,2560,0.01609813372294108
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,512,2048,0.08670933246612549
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,512,2048,0.013942399621009826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,2048,0.010724266370137531
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,2048,0.014295466740926108
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,512,1536,0.0701194683710734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,512,4096,0.022294400135676067
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,1536,0.008556800087292989
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,1536,0.01286720037460327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,512,1024,0.05392640034357706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,512,3584,0.020427733659744263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,1024,0.006454400221506755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,1024,0.011143466830253601
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,512,768,0.04447466532389323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,512,768,0.0088128000497818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,768,0.0058112000425656635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,768,0.010628267129262289
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,512,512,0.03501439889272054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,512,3072,0.018447999159495035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,512,0.004995200037956238
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,512,0.009686400492986042
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,512,256,0.028679466247558592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,512,1536,0.012095999717712403
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,256,0.004231466849644979
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,256,0.008904533584912618
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,512,128,0.02595306634902954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,512,128,0.006131199995676676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,128,0.0037834666669368743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,128,0.008354133367538452
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,512,64,0.02528426647186279
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,64,0.0034272000193595886
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,64,0.008541867136955261
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,512,32,0.0247381329536438
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,32,0.003589333345492681
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,32,0.008513066172599792
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,512,1024,0.009964799880981446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,256,65536,2.62760009765625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,65536,0.12298026879628499
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,65536,0.37256746292114257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,256,16384,0.6885279973347982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,256,16384,0.06978453000386556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,16384,0.03948800166447957
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,16384,0.08592426776885986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,512,512,0.007712000111738841
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,256,12288,0.5272053400675456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,512,256,0.006629333396752675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,12288,0.03267839948336283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,12288,0.0626688003540039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,256,10240,0.4468821207682292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,10240,0.030964267253875733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,10240,0.053674666086832676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,256,8192,0.36474453608194984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,256,65536,0.2651125272115072
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,8192,0.02147946755091349
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,8192,0.0433514674504598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,256,7168,0.32995945612589517
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,256,12288,0.05362879832585653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,7168,0.019774933656056724
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,7168,0.03888533512751262
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,256,10240,0.04669653177261353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,256,6144,0.287169075012207
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,6144,0.017298134167989095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,6144,0.033852799733479814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,256,5120,0.2438368002573649
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,256,8192,0.038534398873647055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,5120,0.01728106737136841
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,5120,0.02637973427772522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,256,4096,0.1929109255472819
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,256,4096,0.01998400092124939
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,4096,0.012276267011960346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,4096,0.021260799964269002
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,256,3584,0.1627072016398112
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,256,7168,0.033692800998687746
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,3584,0.011252267162005107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,3584,0.01939199964205424
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,256,3072,0.13534293174743653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,256,3072,0.015749333302179973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,3072,0.009852799773216247
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,3072,0.017972266674041747
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,256,2560,0.11013440291086833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,256,6144,0.02987946669260661
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,2560,0.008731733759244282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,2560,0.01616426706314087
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,256,2048,0.08679680029551187
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,256,5120,0.02468693256378174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,2048,0.007461333274841308
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,2048,0.014327466487884521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,256,1536,0.07037013371785482
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,256,3584,0.017869865894317626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,1536,0.006595199803511302
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,1536,0.012570666273434958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,256,1024,0.05358506838480631
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,256,1024,0.009058133761088053
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,1024,0.005239466826121012
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,1024,0.010904533664385478
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,256,768,0.04561706781387329
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,256,2560,0.014050133029619851
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,768,0.00459199994802475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,768,0.010404266913731893
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,256,512,0.034541865189870194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,256,2048,0.012264532844225566
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,512,0.004120533416668574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,512,0.009545600414276123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,256,256,0.028385066986083986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,256,256,0.006278400123119354
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,256,0.003697066754102707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,256,0.00881813367207845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,256,128,0.02573653260866801
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,256,1536,0.010714667042096455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,128,0.0034048000971476236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,128,0.008297599852085114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,256,64,0.025009065866470337
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,64,0.003235200047492981
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,64,0.008411733309427898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,256,32,0.024592000246047973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,32,0.0034175999462604523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,32,0.008474666873613994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,128,65536,2.6336041768391927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,256,768,0.008090666433175405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,65536,0.1085546652475993
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,65536,0.3727285385131836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,256,512,0.006962133447329204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,128,16384,0.6907349268595377
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,16384,0.03315946658452352
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,16384,0.08687679767608643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,128,12288,0.5284757296244303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,128,12288,0.047620264689127605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,12288,0.026680533091227216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,12288,0.06276799837748209
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,128,10240,0.44706134796142577
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,128,10240,0.040889600912729904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,10240,0.023357866207758586
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,10240,0.05368426640828451
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,128,8192,0.364849058787028
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,256,128,0.005724800129731497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,8192,0.017145599921544394
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,8192,0.04285120169321696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,128,7168,0.33120320638020834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,128,7168,0.029616000254948933
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,7168,0.01649386684099833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,7168,0.038024532794952395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,128,6144,0.2866719881693522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,128,6144,0.025406932830810545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,6144,0.013896532853444419
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,6144,0.0321727991104126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,128,5120,0.24331839879353842
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,128,65536,0.2320789337158203
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,5120,0.01145919958750407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,5120,0.02615999976793925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,128,4096,0.1932693322499593
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,128,16384,0.06143573522567749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,128,4096,0.017084799210230508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,4096,0.009454933802286784
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,4096,0.02123519976933797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,128,3584,0.16305920282999675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,3584,0.008654933174451191
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,3584,0.019378133614857993
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,128,3072,0.13528107007344564
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,128,8192,0.03354133367538452
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,3072,0.008125866452852886
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,3072,0.01759679913520813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,128,2560,0.11040000120798747
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,128,5120,0.021066667636235555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,2560,0.007436800003051758
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,2560,0.016074666380882265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,128,2048,0.08650346597035727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,128,3072,0.0139765332142512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,2048,0.00655680000782013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,2048,0.014165332913398743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,128,3584,0.015446399648984274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,128,1536,0.06986773014068604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,128,1536,0.009700266520182292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,1536,0.0059008002281188965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,1536,0.012556800246238708
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,128,1024,0.053445335229237875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,1024,0.004863999783992767
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,1024,0.010846933722496033
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,128,768,0.044939732551574706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,128,2560,0.012553600470225015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,768,0.004369066655635833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,768,0.010244266192118327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,128,512,0.03418986797332764
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,128,2048,0.011008000373840332
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,512,0.004023466755946477
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,512,0.009460266431172688
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,128,256,0.028565333286921187
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,128,768,0.0077354664603869125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,256,0.003570133447647095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,256,0.008850133419036866
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,128,128,0.025616000096003216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,128,1024,0.008384000261624653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,128,0.0032874666154384612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,128,0.008308266599973042
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,128,64,0.024946133295694985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,64,0.0031648000081380212
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,128,512,0.0067562664548556015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,64,0.008342400193214417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,128,32,0.02441920042037964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,32,0.0032074667513370516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,32,0.008277333279450735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,65536,0.10475306510925293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,65536,0.37210346857706705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,64,16384,0.6900032043457032
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,16384,0.0306005338827769
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,16384,0.08700799942016602
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,64,65536,2.632978057861328
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,64,12288,0.5284074783325196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,12288,0.02760426600774129
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,12288,0.06262400150299072
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,10240,0.021255467335383096
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,10240,0.05337599913279215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,64,10240,0.4476160049438477
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,8192,0.012316800157229106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,8192,0.04295039971669515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,64,8192,0.364902400970459
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,7168,0.011333333452542622
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,7168,0.03823893467585246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,64,7168,0.330734920501709
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,6144,0.012989866733551025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,6144,0.031768532594045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,64,6144,0.28663679758707683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,5120,0.009649067123730978
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,64,5120,0.24361066818237304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,5120,0.025603200991948443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,4096,0.008807466427485148
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,4096,0.021269333362579346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,64,4096,0.19304854075113934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,3584,0.007782400151093801
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,64,3584,0.16312959988911946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,3584,0.01978879968325297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,3072,0.007068799932797749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,64,3072,0.13589973449707032
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,3072,0.01786880095799764
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,2560,0.006759466727574666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,64,2560,0.1099669297536214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,2560,0.01613759994506836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,64,2048,0.08616213003794351
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,2048,0.005898666878541311
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,2048,0.01413653294245402
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,1536,0.005220266679922739
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,64,1536,0.07024746735890706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,1536,0.012461866935094197
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,64,1024,0.053649067878723145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,1024,0.004465066889921824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,1024,0.010924800237019857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,64,768,0.04486293395360311
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,768,0.004117333392302195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,768,0.010237866640090942
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,64,512,0.034289065996805826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,512,0.003802666564782461
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,512,0.009371733665466309
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,64,256,0.028334933519363403
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,256,0.0034175999462604523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,256,0.008707200487454731
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,128,0.003171200056870779
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,64,128,0.025663999716440837
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,128,0.008184533317883809
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,64,64,0.024844799439112344
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,64,0.0030634666482607523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,64,0.008293333152929943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,32,0.003068800022204717
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,64,32,0.024435200293858848
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,32,0.008306133250395458
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,65536,0.10499306519826253
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,65536,0.37204373677571617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,32,16384,0.6915327707926433
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,16384,0.029338665803273517
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,16384,0.0874837319056193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,32,65536,2.6265599568684896
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,32,12288,0.527945613861084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,12288,0.02635626594225566
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,12288,0.0625109314918518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,10240,0.018736000855763754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,10240,0.05326399803161621
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,32,10240,0.44730987548828127
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,8192,0.011159466703732808
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,8192,0.04261866807937622
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,32,8192,0.3643616040547689
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,7168,0.009962667028109233
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,32,7168,0.3301898638407389
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,7168,0.03792320092519124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,6144,0.009322667121887207
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,6144,0.031191466252009074
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,32,6144,0.2866304079691569
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,5120,0.008593066533406576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,32,5120,0.24283520380655924
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,5120,0.025493333737055462
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,4096,0.007905066510041555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,4096,0.021186133225758873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,32,4096,0.19220159848531088
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,3584,0.007593599955240886
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,32,3584,0.1630143960316976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,3584,0.019364267587661743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,32,3072,0.13556586901346843
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,3072,0.007357866565386455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,3072,0.01760639945665995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,32,2560,0.11019519964853923
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,2560,0.0062730665008227035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,2560,0.01607360045115153
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,32,2048,0.08654399712880453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,2048,0.005592533449331919
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,2048,0.014180266857147216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,32,1536,0.06997013092041016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,1536,0.005019733309745788
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,1536,0.012600533167521157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,32,1024,0.05338346560796102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,1024,0.004353066782156626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,1024,0.010818133751551311
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,32,768,0.04442453384399414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,768,0.004033066580692927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,768,0.010261332988739014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,32,512,0.03418453137079875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,512,0.003688533355792364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,512,0.009373866518338521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,32,256,0.028460800647735596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,256,0.0033930666744709016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,256,0.008706133564313252
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,32,128,0.025753599405288697
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,128,0.0031690667072931922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,128,0.008206933240095774
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,32,64,0.024770132700602212
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,64,0.008286933104197185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,64,0.003011200080315272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2048,32,32,0.024421334266662598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,32,0.003045333425203959
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,32,0.008255999783674877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,65536,16384,1.4880276997884114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,65536,16384,1.984759521484375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,65536,16384,1.580312474568685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,65536,16384,3.385376993815104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,65536,12288,1.5149823506673177
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,65536,12288,2.4763478597005206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,65536,12288,1.1606965382893881
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,65536,10240,1.2618709564208985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,65536,10240,0.9326826731363932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,65536,12288,1.1697930653889974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,65536,10240,2.089234161376953
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,65536,10240,0.968833096822103
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,65536,8192,0.9892650604248047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,65536,8192,0.8072330474853515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,65536,8192,0.726579221089681
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,65536,8192,1.7964575449625653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,65536,7168,0.8748287836710611
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,65536,7168,0.7256511688232422
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,65536,7168,0.6541173299153645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,65536,7168,1.4125471750895182
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,65536,6144,0.7617066701253254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,65536,6144,1.1814634958902994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,65536,6144,0.5660160064697266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,65536,5120,0.6191477457682292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,65536,5120,0.47815148035685223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,65536,5120,0.9489184061686198
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,65536,5120,0.4720778783162435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,65536,4096,0.4858240127563477
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,65536,4096,0.3919839859008789
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,65536,4096,0.7504639943440755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,65536,6144,0.5708757400512695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,65536,4096,0.3645279884338379
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,65536,3584,0.42701120376586915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,65536,3584,0.3556490580240885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,65536,3584,0.32540906270345055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,65536,3584,0.665075174967448
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,65536,3072,0.3683349291483561
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,65536,3072,0.5658655802408854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,65536,3072,0.2837525367736816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,65536,2560,0.30732587178548176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,65536,2560,0.25408639907836916
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,65536,2560,0.47463467915852864
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,65536,2560,0.23621546427408852
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,65536,2048,0.24503679275512696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,65536,2048,0.2091808001200358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,65536,2048,0.38667093912760414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,65536,2048,0.18992640177408854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,65536,1536,0.1911466598510742
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,65536,1536,0.16451306343078614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,65536,1536,0.29420267740885414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,65536,1536,0.14658986727396647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,65536,1024,0.14799893697102864
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,65536,1024,0.1212224006652832
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,65536,1024,0.2037504037221273
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,65536,1024,0.11595306396484376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,65536,768,0.12679466406504314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,65536,768,0.10170559883117676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,65536,768,0.15915733973185223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,65536,3072,0.29897918701171877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,65536,768,0.09990613460540772
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,65536,512,0.07821760177612305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,65536,512,0.11413866678873699
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,65536,512,0.11701866785685222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,65536,512,0.08893013000488281
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,65536,256,0.09882986545562744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,65536,256,0.06743040084838867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,65536,256,0.07126932938893636
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,65536,256,0.07514453728993734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,65536,128,0.09358399709065755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,65536,128,0.06384106477101645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,65536,128,0.06296000083287558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,65536,128,0.07115413347880045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,65536,64,0.09441920121510824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,65536,64,0.05773226817448934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,65536,64,0.07291413148244222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,65536,32,0.09472106297810873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,65536,32,0.05721173286437988
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,65536,32,0.07365866502126059
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,16384,65536,2.930907694498698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,65536,3.220880126953125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,65536,1.6031424204508464
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,16384,16384,0.7248714447021485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,16384,16384,0.39068800608317056
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,16384,0.7088042577107747
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,16384,0.39399468104044594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,16384,12288,0.5744319915771484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,16384,12288,0.2910186767578125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,16384,65536,1.6435999552408855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,12288,0.2949098587036133
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,12288,0.5302517255147298
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,16384,10240,0.25095252990722655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,16384,10240,0.4810965220133464
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,10240,0.44121707280476885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,10240,0.2470400015513102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,16384,8192,0.3745973269144694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,8192,0.35454079310099285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,8192,0.19877972602844238
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,16384,7168,0.3413344065348307
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,16384,7168,0.17475306193033854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,7168,0.31134185791015623
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,7168,0.16818559964497884
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,16384,6144,0.2920778592427572
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,16384,6144,0.15065280596415204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,6144,0.2677856127421061
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,6144,0.14529600143432617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,16384,5120,0.2344714641571045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,16384,5120,0.1274677356084188
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,5120,0.22466665903727212
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,5120,0.12034666538238525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,16384,4096,0.18423147201538087
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,16384,8192,0.19835093816121419
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,4096,0.18329280217488605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,4096,0.0966912031173706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,16384,3584,0.16325546900431315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,16384,4096,0.10406186580657958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,3584,0.1598709265391032
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,3584,0.08298346996307374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,16384,3072,0.14244267145792644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,16384,3072,0.08109119733174643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,3072,0.13852160771687824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,3072,0.0735157330830892
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,16384,2560,0.12287360032399494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,16384,2560,0.06994453271230062
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,2560,0.11725440025329589
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,2560,0.06445333162943521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,16384,2048,0.10067093372344971
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,16384,2048,0.058347733815511074
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,2048,0.09583360354105631
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,2048,0.05273600021998087
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,16384,1536,0.08281706968943278
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,16384,1536,0.047304534912109376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,1536,0.07424426873524984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,1536,0.04261120160420735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,16384,3584,0.09385173320770264
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,16384,1024,0.06487679878870646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,1024,0.05258133411407471
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,1024,0.03476053476333618
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,16384,768,0.05806293487548828
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,16384,768,0.031856000423431396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,768,0.04231893221537272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,768,0.031142399708429976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,16384,512,0.05244693358739218
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,16384,1024,0.03622080087661743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,512,0.031939200560251874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,512,0.02754559914271037
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,16384,256,0.04593706528345744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,16384,512,0.026661332448323565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,256,0.022638932863871256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,256,0.023036799828211465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,16384,128,0.04426346619923909
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,16384,256,0.019349332650502524
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,128,0.017245866854985557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,128,0.022846933205922446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,16384,64,0.0435370683670044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,64,0.01486186683177948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,64,0.022794665892918904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,16384,32,0.04400639931360881
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,16384,128,0.016590933005015053
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,32,0.015412267049153647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,32,0.02283946673075358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,12288,65536,2.4675615946451823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,65536,2.2504308064778646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,65536,1.197268295288086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,12288,16384,0.6362677256266276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,12288,16384,0.29592533111572267
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,16384,0.5301162719726562
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,12288,65536,1.157489013671875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,16384,0.2939242680867513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,12288,12288,0.4898058573404948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,12288,12288,0.22242026329040526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,12288,0.3953301429748535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,12288,0.2266453266143799
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,12288,10240,0.4138186772664388
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,12288,10240,0.18733654022216797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,10240,0.19576106071472169
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,10240,0.3330762545267741
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,12288,8192,0.1523189385732015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,12288,8192,0.32222506205240886
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,8192,0.14572906494140625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,8192,0.2685962677001953
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,12288,7168,0.134552526473999
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,12288,7168,0.29482240676879884
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,7168,0.12366293271382649
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,7168,0.2362112045288086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,12288,6144,0.2494368076324463
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,6144,0.2016437371571859
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,6144,0.11085440317789715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,12288,5120,0.20335572560628257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,12288,5120,0.09854826927185059
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,5120,0.1692245324452718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,5120,0.09326612949371338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,12288,4096,0.15886933008829754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,12288,6144,0.11653013229370117
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,4096,0.13819413185119628
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,4096,0.07599679629007974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,12288,3584,0.14193493525187176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,12288,4096,0.08048640092213949
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,3584,0.12226773103078206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,3584,0.06571520169576009
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,12288,3072,0.12355093161265056
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,12288,3072,0.0629472017288208
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,3072,0.10505279699961345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,3072,0.057911467552185056
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,12288,2560,0.10845226446787518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,12288,2560,0.054568533102671304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,2560,0.08912959893544516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,2560,0.05029653310775757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,12288,2048,0.08903786341349283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,12288,2048,0.045916799704233804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,2048,0.07281280358632405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,2048,0.04166613419850667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,12288,1536,0.07402453422546387
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,12288,1536,0.037869866689046225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,1536,0.05654720067977905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,1536,0.034254932403564455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,12288,1024,0.058723199367523196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,12288,1024,0.029291733105977373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,1024,0.04040640195210775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,1024,0.02746666669845581
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,12288,768,0.052406398455301914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,12288,768,0.02431679964065552
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,768,0.0331712007522583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,768,0.024678399165471397
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,12288,512,0.046987732251485184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,12288,512,0.02001813252766927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,512,0.0256117324034373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,512,0.02071466644605001
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,12288,256,0.04125333229700724
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,12288,256,0.01444586714108785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,256,0.018125865856806436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,256,0.017189333836237587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,12288,128,0.035447466373443606
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,12288,3584,0.07266879876454671
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,128,0.013704533378283182
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,128,0.01627946694691976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,12288,64,0.03500373363494873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,64,0.01220373312632243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,64,0.016636799772580466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,12288,32,0.03505386511484782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,32,0.012370133399963379
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,32,0.01723626653353373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,10240,65536,2.2432896931966146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,10240,65536,1.1976938883463542
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,65536,1.982546106974284
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,12288,128,0.013217066725095114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,10240,16384,0.5810111999511719
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,65536,0.9876192092895508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,16384,0.4492224057515462
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,16384,0.25217280387878416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,10240,12288,0.44933439890543625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,10240,16384,0.29365228017171224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,12288,0.3348640124003092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,12288,0.19064213434855143
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,10240,10240,0.37973651885986326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,10240,12288,0.22364479700724282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,10240,10240,0.18663360277811686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,10240,0.2832373301188151
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,10240,0.16404693921407063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,10240,8192,0.1509376049041748
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,10240,8192,0.2970730781555176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,8192,0.22508479754130045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,8192,0.12282666365305583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,10240,7168,0.1347061316172282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,10240,7168,0.27348159154256185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,7168,0.20056212743123375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,7168,0.10910080273946125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,10240,6144,0.2294794718424479
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,6144,0.17050239245096843
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,6144,0.09486400286356608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,10240,5120,0.18749119440714518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,10240,5120,0.09817066987355551
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,5120,0.14318079948425294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,5120,0.0778218666712443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,10240,4096,0.14559787114461262
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,10240,4096,0.08088640371958414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,4096,0.11566932996114095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,4096,0.06324373483657837
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,10240,3584,0.1304202636082967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,10240,3584,0.07392746607462565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,3584,0.1024821360905965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,3584,0.0573802669843038
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,10240,3072,0.1144330660502116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,10240,3072,0.06386026541392008
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,3072,0.08877973556518555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,3072,0.04919893344243367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,10240,2560,0.10083413124084473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,10240,2560,0.056517334779103603
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,2560,0.0750922679901123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,2560,0.04343893527984619
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,10240,2048,0.0828437328338623
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,10240,2048,0.047115735212961835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,2048,0.06110613346099854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,2048,0.03654400110244751
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,10240,1536,0.06923306783040364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,10240,1536,0.03836693366368611
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,1536,0.04785066843032837
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,1536,0.030002133051554365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,10240,6144,0.11659093697865804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,10240,1024,0.054416000843048096
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,1024,0.0340117335319519
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,1024,0.023511467377344768
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,10240,768,0.04800533453623454
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,10240,768,0.02455679972966512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,768,0.027129600445429485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,768,0.020514132579167683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,10240,512,0.04084800084431966
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,10240,512,0.02060479919115702
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,512,0.020322134097417198
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,512,0.017692800362904867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,10240,256,0.035973334312438966
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,10240,256,0.016167466839154564
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,256,0.014030933380126953
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,256,0.015018666783968607
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,10240,128,0.032706133524576825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,10240,128,0.014378666877746582
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,128,0.01071679989496867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,128,0.013933866222699483
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,10240,64,0.032737066348393754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,64,0.009333333373069764
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,64,0.01507306694984436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,10240,32,0.03323306639989217
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,32,0.009562666217486065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,32,0.014658133188883463
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,8192,65536,2.0566805521647136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,10240,1024,0.030576000610987347
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,65536,1.4522602081298828
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,65536,0.7987733205159505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,8192,16384,0.5291242599487305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,8192,65536,0.8007850646972656
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,16384,0.3931328137715658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,16384,0.22603306770324708
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,8192,12288,0.41666240692138673
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,8192,12288,0.15606187184651693
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,12288,0.26242666244506835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,12288,0.1528458595275879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,8192,10240,0.35226561228434244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,8192,16384,0.20482239723205567
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,10240,0.2210613409678141
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,10240,0.1270016034444173
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,8192,8192,0.27221333185831703
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,8192,10240,0.13000960350036622
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,8192,0.17629440625508624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,8192,0.10022613207499187
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,8192,7168,0.2508106708526611
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,8192,7168,0.09330666859944661
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,7168,0.15494613647460936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,7168,0.0881066640218099
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,8192,6144,0.21261866887410483
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,8192,6144,0.081059201558431
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,6144,0.13361387252807616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,6144,0.0752458651860555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,8192,5120,0.16949973106384278
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,8192,5120,0.06885120073954264
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,5120,0.11215786933898926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,5120,0.06399360100428263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,8192,4096,0.1345311959584554
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,8192,4096,0.05691306591033936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,8192,8192,0.10612906614939373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,4096,0.0908245325088501
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,4096,0.053199998537699376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,8192,3584,0.1197930653889974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,3584,0.08027093410491944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,3584,0.0473685344060262
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,8192,3072,0.1058282693227132
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,8192,3072,0.04547733465830485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,3072,0.06976319948832194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,3072,0.040644268194834396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,8192,2560,0.09284053643544515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,8192,3584,0.05175893306732178
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,2560,0.05947306553522745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,2560,0.03621973196665446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,8192,2048,0.07564266522725424
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,8192,2560,0.03992319901784261
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,2048,0.04830400149027507
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,2048,0.03014613389968872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,8192,1536,0.06445653438568115
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,8192,1536,0.026779733101526898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,1536,0.037274666627248126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,1536,0.024062933524449666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,8192,1024,0.04910080035527547
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,8192,1024,0.020424532890319824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,1024,0.02611946662267049
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,1024,0.019258666038513183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,8192,768,0.04312533140182495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,8192,768,0.01760639945665995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,768,0.020973867177963255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,768,0.016950400670369466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,8192,512,0.03705173333485921
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,8192,512,0.014631467064221701
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,512,0.01639573375384013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,512,0.014883200327555338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,8192,256,0.033419732252756754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,8192,256,0.011698133746782939
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,256,0.011532800396283467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,256,0.013041067123413085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,8192,128,0.03033493359883626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,8192,128,0.011025066177050274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,128,0.009020800391832989
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,128,0.012108799815177918
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,8192,64,0.030533333619435627
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,64,0.007929599781831106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,64,0.012939733266830445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,8192,2048,0.033801599343617754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,8192,32,0.030331732829411824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,32,0.008060800035794576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,32,0.012853333353996277
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,7168,65536,0.8051285425821939
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,7168,65536,2.0164724985758467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,65536,1.387153116861979
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,7168,16384,0.531222407023112
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,65536,0.8170890808105469
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,7168,16384,0.2167424043019613
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,16384,0.33560425440470376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,16384,0.19396799405415852
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,7168,12288,0.15741012891133627
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,7168,12288,0.40918827056884766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,12288,0.24913172721862792
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,12288,0.15077865918477376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,7168,10240,0.13278613090515137
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,7168,10240,0.34785493214925134
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,10240,0.21036052703857422
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,10240,0.12913386821746825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,7168,8192,0.10725013415018718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,7168,8192,0.26920960744222006
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,8192,0.1680405298868815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,8192,0.09784320195515951
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,7168,7168,0.2505023956298828
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,7168,0.1479317347208659
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,7168,0.0856661319732666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,7168,6144,0.211735471089681
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,7168,6144,0.08216959635416667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,6144,0.12746453285217285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,6144,0.07538987000783284
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,7168,5120,0.17140053113301595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,7168,5120,0.07006719907124838
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,5120,0.10717226664225261
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,5120,0.0639903982480367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,7168,4096,0.13122666676839193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,7168,4096,0.05832000176111857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,4096,0.0869493325551351
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,4096,0.051990401744842527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,7168,3584,0.11800106366475423
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,7168,3584,0.053497600555419925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,3584,0.07659733295440674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,3584,0.046607999006907146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,7168,3072,0.1041759967803955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,7168,3072,0.04676266511281331
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,3072,0.06675093173980713
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,3072,0.04100266695022583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,7168,2560,0.0925055980682373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,7168,2560,0.04169706503550212
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,2560,0.05635519822438558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,2560,0.03522986570994059
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,7168,2048,0.07589653333028158
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,7168,2048,0.034986666838328045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,2048,0.046060800552368164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,2048,0.029019733270009358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,7168,1536,0.06449386676152548
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,7168,7168,0.0955029328664144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,1536,0.035796264807383224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,1536,0.023218133052190146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,7168,1024,0.04848639965057373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,7168,1536,0.02760106722513835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,1024,0.025854933261871337
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,1024,0.019078399737675986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,7168,768,0.04316266775131226
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,7168,1024,0.02232746680577596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,7168,768,0.01856106718381246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,768,0.020872533321380615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,768,0.016910932461420693
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,7168,512,0.03699839909871419
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,512,0.016314666469891867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,512,0.015077333648999533
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,7168,256,0.032425600290298465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,7168,256,0.012713600198427835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,256,0.011664000153541566
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,256,0.013128532965977987
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,7168,128,0.030552534262339275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,7168,128,0.011821867028872172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,128,0.009190400441487629
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,128,0.01214400033156077
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,7168,64,0.030613332986831665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,64,0.007995733122030894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,64,0.012776533762613932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,7168,32,0.03041599988937378
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,32,0.008188800017038981
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,32,0.013075199723243714
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,6144,65536,1.851904042561849
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,6144,65536,0.6170986811319987
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,65536,1.0885108947753905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,65536,0.6553216298421224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,6144,16384,0.49046932856241865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,6144,16384,0.15830186208089192
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,16384,0.27325439453125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,7168,512,0.01593386630217234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,16384,0.15522133509318034
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,6144,12288,0.12003733317057293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,6144,12288,0.3717482566833496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,12288,0.20683199564615884
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,12288,0.11559147040049236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,6144,10240,0.1017568031946818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,6144,10240,0.315010134379069
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,10240,0.17419734001159667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,10240,0.10918933550516766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,6144,8192,0.08229119777679443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,6144,8192,0.24815893173217773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,8192,0.13881386121114095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,8192,0.07962666352589926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,6144,7168,0.23195947011311852
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,7168,0.12188906669616699
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,7168,0.06967573165893555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,6144,6144,0.1943893273671468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,6144,6144,0.06341866652170816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,6144,0.10529920260111492
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,6144,0.05943466822306315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,6144,5120,0.15653759638468426
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,6144,5120,0.05411200126012167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,5120,0.08838187058766683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,5120,0.05187733173370361
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,6144,4096,0.12217813332875568
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,6144,4096,0.04518506526947021
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,4096,0.07146026293436686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,4096,0.04206186532974243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,6144,3584,0.10973013242085775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,6144,3584,0.04082560141881307
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,3584,0.06319040060043335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,3584,0.03810240030288696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,6144,3072,0.09636159737904867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,6144,3072,0.03579946756362915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,3072,0.054746667544047035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,3072,0.03315839966138204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,6144,2560,0.08628906408945719
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,6144,7168,0.07339200178782145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,2560,0.04665280183156331
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,2560,0.02765760024388631
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,6144,2048,0.07067626317342122
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,6144,2560,0.031566933790842695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,2048,0.0376693328221639
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,2048,0.022711465756098427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,6144,1536,0.06038186550140381
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,6144,1536,0.022082134087880453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,1536,0.02882133324940999
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,1536,0.019407999515533448
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,6144,1024,0.047235198815663654
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,6144,1024,0.01723626653353373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,1024,0.02048106590906779
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,1024,0.01598186691602071
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,6144,768,0.04055146773656209
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,6144,768,0.015194666385650635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,768,0.01673813263575236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,768,0.014405333002408347
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,6144,512,0.03540906508763631
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,6144,512,0.013194666306177775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,512,0.012831999858220419
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,512,0.012596266468365988
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,6144,256,0.030077866713205975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,6144,256,0.010678399602572124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,256,0.009116799632708231
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,256,0.011127466956774395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,6144,128,0.028542933861414592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,6144,128,0.009666132926940917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,128,0.007359999914964039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,128,0.010474666953086853
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,6144,64,0.0286463995774587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,64,0.0064522668719291685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,64,0.01109333336353302
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,6144,2048,0.026798933744430542
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,6144,32,0.028139734268188478
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,32,0.006751999755700429
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,32,0.01115733285744985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,5120,65536,1.8103562672932942
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,5120,65536,0.538105583190918
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,65536,0.6333077112833659
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,65536,0.892089589436849
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,5120,16384,0.15546132723490397
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,5120,16384,0.478439458211263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,16384,0.24197759628295898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,16384,0.15199039777119955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,5120,12288,0.38018986384073894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,12288,0.16601813634236653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,12288,0.11758613586425781
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,5120,10240,0.31340373357137047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,5120,10240,0.09030186335245768
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,10240,0.13897600173950195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,10240,0.09680639902750651
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,5120,8192,0.24525866508483887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,5120,8192,0.07332159678141276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,8192,0.11191786924997967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,8192,0.07525226275126139
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,5120,7168,0.22832533518473305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,5120,7168,0.0652405341466268
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,7168,0.09864532947540283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,7168,0.06799893379211426
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,5120,6144,0.19218986829121906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,5120,6144,0.05676373243331909
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,6144,0.08547840118408204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,6144,0.05857280095418295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,5120,5120,0.1527509371439616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,5120,5120,0.0484661340713501
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,5120,0.07179093360900879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,5120,0.050970665613810216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,5120,4096,0.11980799833933513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,5120,4096,0.040712531407674155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,4096,0.05818560123443604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,4096,0.041092268625895184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,5120,3584,0.1083349307378133
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,5120,3584,0.036880000432332354
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,3584,0.05151253143946329
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,3584,0.03708159923553467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,5120,3072,0.09476799964904785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,5120,3072,0.03236266573270162
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,3072,0.04479999939600627
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,3072,0.031428267558415726
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,5120,2560,0.08444693088531494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,5120,2560,0.027752532561620073
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,2560,0.03801706631978353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,2560,0.02664320071538289
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,5120,2048,0.07055679957071939
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,5120,2048,0.023754666248957314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,2048,0.031268266836802165
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,2048,0.022426666816075642
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,5120,1536,0.059480532010396325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,5120,1536,0.01957119901974996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,1536,0.024251733223597208
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,1536,0.01918720006942749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,5120,12288,0.10645973682403564
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,5120,1024,0.04714560111363729
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,1024,0.017655466000239053
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,1024,0.01583466629187266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,5120,768,0.03956693410873413
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,5120,768,0.014085333546002707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,768,0.01471466620763143
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,768,0.014426666498184203
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,5120,512,0.03489813407262166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,5120,512,0.012169599533081055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,512,0.011565867066383361
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,512,0.01263146698474884
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,5120,256,0.030188800891240437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,5120,1024,0.015479466319084168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,256,0.00869546631971995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,256,0.01107199986775716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,5120,128,0.02844373385111491
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,5120,256,0.010051199793815612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,128,0.0072053333123524976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,128,0.010357333223025005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,5120,64,0.028843732674916585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,64,0.006423466900984447
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,64,0.01099733312924703
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,5120,32,0.028573866685231524
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,32,0.0066655998428662615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,32,0.011132799585660299
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,4096,65536,1.655421829223633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,4096,65536,0.43202667236328124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,65536,0.6824405034383138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,65536,0.46366399129231767
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,4096,16384,0.44040212631225584
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,4096,16384,0.11546239852905274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,16384,0.17325545946756998
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,16384,0.11100052992502848
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,4096,12288,0.33770987192789714
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,5120,128,0.009038933118184407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,12288,0.13180480003356934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,12288,0.0832373301188151
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,4096,10240,0.2858784039815267
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,4096,10240,0.07351893583933512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,10240,0.11057919661204021
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,10240,0.07081387042999268
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,4096,8192,0.22444693247477213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,4096,12288,0.08647680282592773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,8192,0.0887733300526937
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,8192,0.057918934027353916
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,4096,7168,0.21078294118245444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,4096,7168,0.05353173414866129
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,7168,0.0781877358754476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,7168,0.050970665613810216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,4096,6144,0.17547520001729328
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,4096,6144,0.04623253345489502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,6144,0.06747199694315592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,6144,0.043567999203999834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,4096,5120,0.14096105893452962
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,4096,5120,0.03980799913406372
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,5120,0.05676906506220499
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,5120,0.03686720132827759
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,4096,4096,0.10910186767578126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,4096,4096,0.03297599951426188
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,4096,0.04616533517837525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,4096,0.029576534032821657
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,4096,3584,0.09981760183970133
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,4096,8192,0.06022400061289469
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,3584,0.04118506511052449
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,3584,0.02664746642112732
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,4096,3072,0.08726186752319336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,4096,3584,0.030373332897822063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,3072,0.03552639881769816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,3072,0.023324799537658692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,4096,2560,0.07889173030853272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,4096,2560,0.023118933041890465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,2560,0.029726932446161907
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,2560,0.020834134022394816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,4096,2048,0.06539306640625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,4096,2048,0.020385066668192543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,2048,0.02421226700146993
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,2048,0.017681066195170084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,4096,1536,0.05533013343811035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,4096,1536,0.0170741339524587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,1536,0.019324799378712974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,1536,0.015037866433461508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,4096,1024,0.04294293324152629
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,4096,1024,0.013779200116793313
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,1024,0.01452906628449758
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,1024,0.0125109334786733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,4096,768,0.03658133347829183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,4096,768,0.012350933750470479
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,768,0.012180266777674358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,768,0.011476266384124755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,4096,512,0.032314666112263996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,4096,512,0.011050666371981304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,512,0.009701333443323771
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,512,0.010380799571673077
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,4096,256,0.028421332438786823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,4096,256,0.00950933297475179
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,256,0.007399466633796692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,256,0.00942186713218689
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,4096,128,0.02696853280067444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,4096,3072,0.026336000363032026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,128,0.0061365331212679545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,128,0.008887466788291932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,4096,64,0.026573866605758667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,64,0.005464533468087515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,64,0.009270399808883667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,4096,32,0.026497066020965576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,32,0.005691733459631602
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,32,0.009480533003807069
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3584,65536,1.622869364420573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3584,65536,0.44118401209513347
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,65536,0.6493045171101888
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,65536,0.46463572184244795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3584,16384,0.4391434669494629
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,4096,128,0.008727467060089112
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,16384,0.16360534032185872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,16384,0.10891306400299072
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3584,12288,0.3286698659261068
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3584,12288,0.08693866729736328
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,12288,0.12355733712514241
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,12288,0.08219093481699626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3584,10240,0.27769813537597654
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3584,10240,0.07353386878967286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,10240,0.10404160022735595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,10240,0.06980586846669515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3584,8192,0.21949440638224282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3584,16384,0.11578133106231689
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,8192,0.08471786975860596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,8192,0.056286935011545816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3584,7168,0.20689172744750978
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3584,8192,0.060850131511688235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,7168,0.07376533349355062
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,7168,0.04979413350423177
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3584,6144,0.17187946637471516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3584,6144,0.047483734289805096
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,6144,0.06387199958165488
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,6144,0.04290986855824788
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3584,5120,0.13839999834696454
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3584,5120,0.04096213181813558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,5120,0.05391146739323934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,5120,0.03656746546427409
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3584,4096,0.10823360284169514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3584,7168,0.05446293354034424
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3584,4096,0.03409813245137532
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,4096,0.04173333247502645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,4096,0.02911679943402608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3584,3584,0.0980842669804891
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,3584,0.03630186716715495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,3584,0.02619626720746358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3584,3072,0.08716800212860107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3584,3072,0.027265065908432008
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,3072,0.0313482662041982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,3072,0.023401600122451783
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3584,2560,0.07800426483154296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3584,2560,0.0241482675075531
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,2560,0.02603200078010559
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,2560,0.02067306637763977
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3584,2048,0.06492160161336263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3584,2048,0.021093332767486574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,2048,0.021303466955820718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,2048,0.017753599087397258
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3584,1536,0.055112532774607335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3584,3584,0.03127359946568807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,1536,0.01703146696090698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,1536,0.01493333379427592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3584,1024,0.042181332906087235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3584,1024,0.014683733383814493
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,1024,0.012434132893880208
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,1024,0.012549333771069846
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3584,768,0.03646293481190999
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3584,1536,0.01757226586341858
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,768,0.010151466727256775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,768,0.011522133151690166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3584,512,0.03219626744588216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3584,512,0.011575466394424439
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,512,0.007982933521270752
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,512,0.01036906639734904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3584,256,0.0284223993619283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3584,768,0.012947199741999307
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,256,0.006359466910362243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,256,0.009447466333707173
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3584,128,0.02679786682128906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3584,128,0.00899733304977417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,128,0.005377066632111868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,128,0.008889599641164144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3584,64,0.026539733012517292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,64,0.005122133096059163
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,64,0.00928106705347697
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3584,32,0.026503467559814455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,32,0.00543146679798762
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,32,0.009297066926956176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3584,256,0.010034132997194927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3072,65536,1.6127381642659504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3072,65536,0.3430239995320638
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,65536,0.5001898765563965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,65536,0.5422239939371745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3072,16384,0.4258144060770671
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,16384,0.13610347112019855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,16384,0.10779840151468914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3072,12288,0.3282538731892904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3072,16384,0.09045759836832681
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,12288,0.1037450631459554
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,12288,0.08131840229034423
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3072,10240,0.2800586700439453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3072,12288,0.06860053539276123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,10240,0.08719253540039062
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,10240,0.07029546896616617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3072,8192,0.21999786694844564
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3072,10240,0.058507732550303136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,8192,0.07028266588846842
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,8192,0.05538986523946127
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3072,7168,0.20607040723164877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3072,8192,0.04775360027949015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,7168,0.06097919940948486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,7168,0.04882986545562744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3072,6144,0.17074880599975586
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3072,6144,0.03738240003585815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,6144,0.05267626841862997
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,6144,0.042721064885457356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3072,5120,0.13766825993855794
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3072,5120,0.032202666997909545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,5120,0.04443519910176595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,5120,0.03545279900232951
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3072,4096,0.10835306644439698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3072,4096,0.02665173411369324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,4096,0.03622399965922038
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,4096,0.02831573287645976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3072,3584,0.09757333596547445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3072,3584,0.02437653342882792
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,3584,0.03182506759961446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,3584,0.02550400098164876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3072,3072,0.08599466482798258
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3072,7168,0.04302186568578084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,3072,0.02789013385772705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,3072,0.022886399428049722
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3072,2560,0.07731946309407553
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3072,3072,0.0218122661113739
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,2560,0.023116799195607503
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,2560,0.020542933543523153
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3072,2048,0.06468480030695598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3072,2048,0.016733866930007935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,2048,0.019059199094772338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,2048,0.01765226721763611
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3072,1536,0.05491733153661092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3072,1536,0.014669866363207499
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,1536,0.015421866377194723
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,1536,0.014803199966748556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3072,1024,0.041715200742085776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3072,2560,0.0195413331190745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,1024,0.011614933609962463
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,1024,0.012404266993204753
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3072,768,0.03635733524958293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3072,768,0.010897066195805867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,768,0.009868799646695455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,768,0.011390933394432068
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3072,512,0.03240639964739482
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3072,512,0.00993386705716451
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,512,0.00804799993832906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,512,0.010365866621335347
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3072,256,0.028590933481852217
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3072,1024,0.012036266922950744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,256,0.006355200211207073
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,256,0.009443199634552002
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3072,128,0.027091199159622194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3072,128,0.007841066519419352
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,128,0.005298133194446564
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,128,0.008825600147247314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3072,64,0.02688000003496806
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,64,0.004818133513132731
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,64,0.009369599819183349
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,3072,32,0.02709973255793254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,32,0.00491839994986852
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,32,0.009302399555842082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3072,256,0.008541867136955261
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2560,65536,1.6009930928548175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,65536,0.4503744125366211
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,65536,0.4544960021972656
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2560,16384,0.42592213948567703
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2560,16384,0.08145813147226968
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,16384,0.11617706616719563
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,16384,0.10844799677530925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2560,12288,0.3286933263142904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,128,256,0.008288000027338665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,12288,0.0880629301071167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,12288,0.0806933323542277
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2560,10240,0.2771519978841146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2560,65536,0.3067743937174479
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,10240,0.07474239667256674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,10240,0.07161280314127604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2560,8192,0.22187946637471517
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2560,12288,0.06214400132497152
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,8192,0.05983253320058187
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,8192,0.05544960101445516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2560,7168,0.20539093017578125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2560,7168,0.039434667428334555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,7168,0.05268266598383585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,7168,0.04866346518198649
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2560,6144,0.17086505889892578
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2560,10240,0.05297386646270752
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,6144,0.04559040069580078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,6144,0.04198293288548787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2560,5120,0.13825492858886718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,128,256,0.005974400043487549
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,5120,0.039051731427510575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,5120,0.03428693215052287
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2560,4096,0.10829119682312012
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2560,8192,0.043636266390482584
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,4096,0.03195520043373108
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,4096,0.02834239999453227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2560,3584,0.09759253660837809
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2560,6144,0.0345578670501709
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,3584,0.027743999163309736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,3584,0.025693867603937787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2560,3072,0.08651306629180908
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2560,3072,0.019723733266194664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,3072,0.023890133698781332
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,3072,0.023056000471115112
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2560,2560,0.0777237335840861
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2560,5120,0.02902719974517822
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,2560,0.02074026664098104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,2560,0.02024959921836853
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2560,2048,0.0639786680539449
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2560,2048,0.015710933009783427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,2048,0.017413334051767985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,2048,0.017501866817474364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2560,1536,0.05440959930419922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2560,4096,0.024081067244211832
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,1536,0.014365866780281067
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,1536,0.014792533715566
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2560,1024,0.041161600748697916
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2560,3584,0.02188373406728109
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,1024,0.011108266313870747
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,1024,0.012446932991345723
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2560,768,0.03613866567611694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,128,256,0.006695466736952464
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,768,0.009706667065620423
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,768,0.011592533191045125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2560,512,0.03191039959589641
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2560,2560,0.01731946667035421
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,512,0.008001066744327545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,512,0.01037440001964569
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2560,256,0.028205867608388262
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2560,1536,0.013470932841300964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2560,1024,0.011241599917411804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,256,0.006499200065930684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,256,0.009335466225941976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2560,128,0.026625066995620728
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,128,0.005197866757710775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,128,0.008797867099444072
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2560,64,0.026587732632954914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,64,0.004741333425045013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,64,0.009128533800443013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2560,32,0.026443733771642046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,32,0.005013333261013031
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,32,0.009136000275611877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2560,768,0.010158933202425639
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2560,512,0.009305600325266521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2048,65536,0.2751317342122396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,65536,0.28939946492513025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,65536,0.3634751955668131
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2048,65536,1.4568373362223306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,16384,0.09095253149668375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2048,16384,0.3858282725016276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,16384,0.06808959643046061
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2048,12288,0.05758399963378906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2048,12288,0.29794772466023767
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,12288,0.0691818634668986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,12288,0.050481065114339196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2048,10240,0.2549781322479248
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,10240,0.05763626495997111
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,10240,0.04445653359095256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2048,8192,0.20154773394266762
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2048,8192,0.03896426757176717
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,8192,0.046690134207407634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,8192,0.03712533315022786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2048,7168,0.19002025922139484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2048,7168,0.035035733381907144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,7168,0.041110400358835855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,7168,0.032001066207885745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2048,6144,0.1584448019663493
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2048,6144,0.030763733386993408
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,6144,0.0359007994333903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,6144,0.026276266574859618
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2048,5120,0.12744853496551514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2560,256,0.008158933122952778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,5120,0.030144000053405763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,5120,0.022381865978240968
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2048,4096,0.09877226352691651
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2560,128,0.0074080000321070354
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,4096,0.024550400177637734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,4096,0.019054933389027914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2048,3584,0.08966506322224935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2048,16384,0.07181546688079835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,3584,0.021861332654953002
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,3584,0.017813332875569663
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2048,3072,0.07975680033365885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2048,10240,0.048094932238260904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,3072,0.01862186590830485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,3072,0.01612160007158915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2048,2560,0.07156480153401693
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2048,5120,0.02646080056826274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,2560,0.016025599837303162
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,2560,0.014693333705266317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2048,2048,0.0599840005238851
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2048,2048,0.013894400000572205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,2048,0.01354986627896627
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,2048,0.012880000472068786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2048,1536,0.0512223998705546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2048,1536,0.012334932883580525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,1536,0.01112000048160553
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,1536,0.011363200346628825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2048,1024,0.038472533226013184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2048,1024,0.010315733154614766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,1024,0.008700799942016602
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,1024,0.009990400075912476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2048,768,0.034443732102711996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2048,4096,0.021038933595021566
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2048,3584,0.020103466510772706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,768,0.007551999886830647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,768,0.009296000003814697
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2048,512,0.030173865954081218
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,512,0.006313600142796834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,512,0.008620799581209818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2048,256,0.026655999819437663
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2048,256,0.007521066566308339
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,256,0.005202133456865946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,256,0.007981866598129272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2048,128,0.025277866919835405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2048,3072,0.01795413295427958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,128,0.00443200021982193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,128,0.007512533167997997
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2048,64,0.02480640014012655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,64,0.004066133250792822
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,64,0.007821866869926452
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,2048,32,0.024829866488774617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,32,0.004259199897448222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,32,0.007899733384450276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2048,2560,0.01595946649710337
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1536,65536,1.417959467569987
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,65536,0.2844789187113444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,65536,0.2779285430908203
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1536,16384,0.3799071947733561
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2048,768,0.009416533509890239
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2048,512,0.008586666981379191
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,16384,0.07149973710378012
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,16384,0.06568426688512166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1536,12288,0.29344107309977213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2048,128,0.00694400022427241
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,12288,0.05362879832585653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,12288,0.05058666865030924
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1536,10240,0.24853760401407876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,10240,0.046298666795094805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,10240,0.04358293215433757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1536,8192,0.19683413505554198
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1536,65536,0.2606250603993734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,8192,0.03666666746139526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,8192,0.03550399939219157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1536,7168,0.18669226964314778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1536,16384,0.0696394681930542
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,7168,0.033973332246144614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,7168,0.029346134265263873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1536,6144,0.15470399856567382
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1536,12288,0.05383679866790771
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,6144,0.028293333450953168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,6144,0.025118933121363325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1536,5120,0.1255232016245524
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1536,10240,0.046374400456746415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,5120,0.026745599508285523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,5120,0.02225066622098287
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1536,4096,0.09748373031616211
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1536,8192,0.03780586719512939
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,4096,0.01917333404223124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,4096,0.0189301331837972
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1536,3584,0.08859093189239502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1536,7168,0.0341813325881958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,3584,0.016577066977818807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,3584,0.017454934120178223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1536,3072,0.07892693678538004
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1536,6144,0.02905706763267517
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,3072,0.015657599767049155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,3072,0.01571626663208008
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1536,2560,0.07087999979654948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1536,2560,0.015565866231918335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,2560,0.013489066561063131
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,2560,0.01428053379058838
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1536,5120,0.024768000841140746
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1536,2048,0.05896960099538168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1536,2048,0.013738666971524557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,2048,0.01176533301671346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,2048,0.012710400422414145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1536,1536,0.04840000073115031
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1536,1536,0.012026666601498922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,1536,0.009275733431180318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,1536,0.01120960017045339
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1536,1024,0.03916800022125244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1536,1024,0.010026666522026061
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,1024,0.007298133273919423
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,1024,0.009780266880989074
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1536,768,0.03391679922739665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,768,0.006457599997520447
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,768,0.009168000022570292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1536,512,0.03010666569073995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1536,512,0.00800960014263789
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,512,0.0055178667108217875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1536,4096,0.02084266742070516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,512,0.00853760043780009
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1536,256,0.026450133323669432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1536,256,0.007075199981530507
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,256,0.00472320020198822
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,256,0.00790719985961914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1536,3584,0.018946133057276406
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1536,128,0.02512213389078776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,128,0.00402453343073527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,128,0.007493333518505096
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1536,64,0.024758400519688924
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,64,0.0037664001186688742
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,64,0.007703466713428498
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1536,32,0.024639999866485594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,32,0.003882666677236557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,32,0.007660800218582153
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1024,65536,1.4209685007731119
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,65536,0.19123306274414062
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,65536,0.26147306760152184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1024,16384,0.3801482518513998
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1536,3072,0.017220266660054526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,16384,0.053192532062530516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,16384,0.06474239826202392
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1024,12288,0.2929141362508138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1536,768,0.008989866574605305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,12288,0.04224960009256999
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,12288,0.04948906501134236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1024,10240,0.24846186637878417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1024,10240,0.039833601315816244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,10240,0.041365333398183185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1536,128,0.006548266609509785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,10240,0.04247039953867594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1024,8192,0.19667840003967285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,8192,0.03182186683019002
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,8192,0.03293333252271016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1024,7168,0.18509546915690106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1024,65536,0.2094186623891195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,7168,0.027867732445398967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1024,16384,0.056415998935699464
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,7168,0.028051199515660603
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1024,6144,0.15579412778218588
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1024,12288,0.04582933187484741
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,6144,0.0247978667418162
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,6144,0.02467306653658549
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1024,5120,0.12722559769948322
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,5120,0.020221867163976035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,5120,0.021666133403778078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1024,4096,0.09733333587646484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1024,8192,0.03063253362973531
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,4096,0.015661866466204325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,4096,0.018552533785502114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1024,3584,0.088700803120931
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1024,3584,0.016548267006874083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,3584,0.015198933084805808
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,3584,0.01696746746699015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1024,3072,0.07861653168996176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1024,7168,0.02751893401145935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,3072,0.013591466347376504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,3072,0.015559466679890952
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1024,2560,0.07023999691009522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1024,2560,0.013432533542315165
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,2560,0.011550933122634888
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,2560,0.014051199952761332
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1024,2048,0.05851840178171793
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1024,2048,0.011872000495592753
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,2048,0.010084266463915508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,2048,0.012460800011952718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1024,1536,0.048794666926066085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1024,6144,0.02390186587969462
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,1536,0.008070399860541026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,1536,0.010987733801205952
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1024,1024,0.037557331720987956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1024,5120,0.020490666230519615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,1024,0.006222933530807495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,1024,0.009611733754475911
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1024,768,0.03356159925460815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1024,4096,0.01792959968249003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,768,0.005676800012588501
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,768,0.009051733215649923
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1024,512,0.029770666360855104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1024,3072,0.015046399831771851
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,512,0.004884266853332519
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,512,0.008346666892369587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1024,256,0.02648959954579671
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1024,256,0.006444799900054932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,256,0.004198400179545084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,256,0.007784533500671387
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1024,128,0.02502506573994954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1024,128,0.0060032000144322716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,128,0.003719466676314672
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,128,0.007334400216738384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1024,64,0.02474773327509562
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,64,0.003402666747570038
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,64,0.007517866790294647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,1024,32,0.024472532669703166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,32,0.0034933333595593774
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1024,1536,0.010635733604431152
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,32,0.007493333518505096
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,768,65536,0.21235200564066567
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,65536,0.1464010715484619
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,65536,0.26031039555867513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,768,65536,1.4049610137939452
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,768,16384,0.3768106778462728
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,16384,0.04242240190505982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1024,1024,0.008690133690834045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,16384,0.06447679996490478
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,768,12288,0.2902496019999186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,12288,0.04127360184987386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,12288,0.05008960167566935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,768,10240,0.2465504010518392
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1024,768,0.0078015998005867004
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,10240,0.03561919927597046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,10240,0.04119786818822225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,768,8192,0.19624320665995282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1024,512,0.007245866457621257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,8192,0.022754132747650146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,8192,0.03152746756871541
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,768,7168,0.1851263999938965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,768,16384,0.058482134342193605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,7168,0.025534933805465697
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,768,12288,0.04466773271560669
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,7168,0.02790186603864034
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,768,6144,0.022827732563018798
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,768,6144,0.15384213129679364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,6144,0.02021333376566569
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,6144,0.02444266676902771
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,768,5120,0.12400106588999431
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,5120,0.019898666938145956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,5120,0.021564799547195434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,768,10240,0.037589331467946366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,768,4096,0.09639786879221598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,4096,0.014153599739074707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,4096,0.018532266219456993
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,768,3584,0.08806080023447672
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,768,8192,0.029579732815424604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,3584,0.012613333264986672
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,3584,0.01716586748758952
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,768,3072,0.07818559805552164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,768,7168,0.026555732885996504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,3072,0.010994133353233338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,3072,0.015550933281580605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,768,2560,0.0701141357421875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,768,2560,0.012273066242535909
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,2560,0.009488000472386678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,2560,0.014107732971509298
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,768,2048,0.05856639941533407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,768,2048,0.010921600461006164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,2048,0.008313600222269695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,2048,0.012548266847928365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,768,1536,0.04966400067011516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,768,5120,0.01932800014813741
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,1536,0.006753066678841908
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,1536,0.01104853351910909
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,768,1024,0.04023466507593791
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,768,1024,0.008116266628106435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,1024,0.005628799895445505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,1024,0.00963200032711029
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,768,4096,0.015793066223462424
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,768,768,0.033590400218963624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,768,768,0.0075082664688428235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,768,0.005197866757710775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,768,0.00904319981733958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,768,512,0.029769599437713623
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,768,3584,0.015096533298492431
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,512,0.004444799820582072
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,512,0.008410666386286418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,768,256,0.02640320062637329
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,768,256,0.006229333579540253
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,256,0.003884800026814143
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,256,0.007814399898052216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,768,128,0.025000532468159992
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,128,0.00348693331082662
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,128,0.0074442664782206226
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,768,64,0.024336000283559166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,64,0.0032501332461833954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,64,0.007590400179227193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,768,32,0.024205867449442545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,32,0.003403733422358831
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,32,0.007534933090209961
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,768,3072,0.013649066289265951
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,512,65536,1.4042261759440104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,65536,0.102237868309021
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,65536,0.25865279833475746
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,512,16384,0.3770357449849447
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,768,1536,0.009915733337402343
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,16384,0.03391253153483073
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,16384,0.06489706834157308
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,512,12288,0.2900416056315104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,768,512,0.006932266553243001
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,12288,0.02877013285954793
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,12288,0.048716799418131514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,768,128,0.005892266829808554
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,512,10240,0.033035733302434284
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,512,10240,0.24607787132263184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,10240,0.03030400077501933
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,10240,0.03882773319880168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,512,8192,0.19542187054951984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,8192,0.020756266514460244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,8192,0.030793599287668866
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,512,7168,0.18485867182413737
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,512,65536,0.17776320775349935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,7168,0.019719467560450236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,7168,0.02758079965909322
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,512,6144,0.15430720647176105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,512,6144,0.01958400011062622
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,6144,0.017554134130477905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,6144,0.024580266078313193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,512,5120,0.12671466668446857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,512,16384,0.048072532812754316
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,5120,0.014536533753077188
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,5120,0.021389865875244142
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,512,4096,0.09659626483917236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,512,12288,0.03802666664123535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,4096,0.011044266819953918
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,4096,0.018348799149195353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,512,3584,0.08802666664123535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,512,3584,0.01341546674569448
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,3584,0.010963199536005656
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,3584,0.01689599951108297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,512,3072,0.07804373105367025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,512,8192,0.024555732806523643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,3072,0.009492266178131103
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,3072,0.015427199999491372
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,512,2560,0.07011199792226155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,512,7168,0.022709333896636964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,2560,0.009081600109736125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,2560,0.013937066992123923
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,512,2048,0.05835946798324585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,512,2048,0.010242133339246114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,2048,0.007267199953397115
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,2048,0.01239466667175293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,512,1536,0.048369065920511885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,512,1536,0.009130666653315227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,1536,0.006260266900062561
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,1536,0.01090666651725769
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,512,1024,0.037825067838033036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,512,5120,0.01709866722424825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,1024,0.005014400184154511
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,512,4096,0.01479680041472117
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,1024,0.009556266665458679
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,512,768,0.033343998591105144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,512,768,0.007079466680685679
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,768,0.004579199850559235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,768,0.008949333429336548
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,512,512,0.02962133288383484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,512,0.0041514667371908825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,512,0.008480000495910644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,512,256,0.02606186668078105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,512,3072,0.012283733487129212
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,256,0.003704533229271571
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,256,0.00768746683994929
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,512,128,0.024731733401616416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,512,128,0.005622399846712748
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,128,0.0033759998778502146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,128,0.007381333410739899
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,512,64,0.02409706711769104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,64,0.003209600100914637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,64,0.007428266604741414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,512,32,0.023795199394226075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,32,0.003319466610749563
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,32,0.007390933235486348
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,256,65536,1.4040351867675782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,512,2560,0.011187199751536052
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,65536,0.07000106970469157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,65536,0.25898027420043945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,256,16384,0.37726399103800456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,512,1024,0.007597866654396057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,16384,0.02408533294995626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,16384,0.06368639866511026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,256,12288,0.28966293334960935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,512,512,0.006533333162466686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,12288,0.020746666193008422
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,12288,0.04606613318125407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,256,10240,0.24639466603597007
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,256,10240,0.027314132452011107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,10240,0.018875734011332194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,10240,0.036841599146525066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,256,8192,0.19582932790120441
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,512,256,0.006044800082842508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,8192,0.013949867089589438
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,8192,0.03067733248074849
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,256,7168,0.1847050666809082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,256,65536,0.1528223991394043
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,7168,0.01360319952170054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,7168,0.027194666862487792
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,256,6144,0.15404267311096193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,256,16384,0.04241173267364502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,6144,0.014556800325711569
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,6144,0.02425280014673869
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,256,5120,0.12484160264333088
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,256,12288,0.032374399900436404
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,5120,0.012250666817029316
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,5120,0.021418666839599608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,256,4096,0.09669973055521647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,256,4096,0.01325973371664683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,4096,0.010105599959691364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,4096,0.018311466773351034
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,256,3584,0.08800106843312581
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,256,3584,0.012326400478680928
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,3584,0.009614933530489604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,3584,0.01708586613337199
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,256,3072,0.07790719668070475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,256,8192,0.022032000621159873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,3072,0.008411733309427898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,3072,0.015405866503715514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,256,2560,0.07001067002614339
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,256,2560,0.010328533252080281
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,2560,0.007449600100517273
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,2560,0.013920000195503235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,256,2048,0.058481065432230625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,256,2048,0.009429333607355754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,2048,0.006841599941253662
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,2048,0.01232319970925649
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,256,1536,0.04879680077234904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,256,7168,0.019821866353352865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,1536,0.006031999985376993
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,1536,0.010870400071144103
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,256,1024,0.03798186779022217
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,256,6144,0.01731626590092977
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,1024,0.005102933446566264
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,1024,0.009541333715120951
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,256,5120,0.015243732929229736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,256,768,0.033123199144999185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,768,0.004586666822433472
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,768,0.008919466535250347
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,256,512,0.029415466388066608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,256,512,0.006275199850400289
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,512,0.0040554667512575785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,512,0.008250666658083598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,256,256,0.026002132892608644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,256,3072,0.0114464004834493
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,256,0.0035914666950702667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,256,0.00764160007238388
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,256,128,0.024918399254480996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,256,1536,0.008462933699289958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,128,0.0033098667860031127
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,128,0.007322666545708974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,256,64,0.024088533719380696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,256,1024,0.007321600119272869
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,64,0.0031776001056035364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,64,0.007386666536331177
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,256,32,0.024168533086776734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,32,0.0032448001205921174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,32,0.007494399944941203
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,128,65536,1.4049067179361978
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,65536,0.05948586861292521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,65536,0.2595242659250895
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,128,16384,0.37765439351399743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,256,768,0.006805333495140076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,16384,0.01686506668726603
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,16384,0.06311466693878173
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,128,12288,0.2894517262776693
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,256,256,0.0058335999647776285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,12288,0.014079999923706055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,12288,0.044487468401590985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,128,10240,0.24603519439697266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,256,128,0.005554133156935374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,10240,0.013446399569511413
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,10240,0.036748798688252766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,128,8192,0.19596160252888997
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,128,8192,0.021211733420689903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,8192,0.010828800002733866
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,8192,0.03015786608060201
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,128,7168,0.18542826970418294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,128,65536,0.14875200589497883
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,7168,0.009671466549237569
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,7168,0.027298132578531902
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,128,6144,0.15418346722920734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,128,16384,0.0413482666015625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,6144,0.01334826648235321
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,6144,0.024222934246063234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,128,5120,0.12404159704844157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,128,5120,0.015029333035151162
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,5120,0.008565333485603333
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,5120,0.021269333362579346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,128,4096,0.09660693009694418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,128,12288,0.0314848005771637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,4096,0.008749866485595703
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,4096,0.01819733381271362
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,128,3584,0.08818666934967041
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,128,10240,0.026165332396825152
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,3584,0.007737599809964498
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,3584,0.016834133863449098
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,128,3072,0.07787306308746338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,128,7168,0.01907520095507304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,3072,0.007718400160471599
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,3072,0.015473066767056783
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,128,2560,0.0699018637339274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,128,2560,0.00997226635615031
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,2560,0.006631466746330261
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,2560,0.013972266515096029
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,128,2048,0.05833280086517334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,128,6144,0.01700053413709005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,2048,0.0059338668982187905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,2048,0.012322133779525757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,128,1536,0.04815786679585775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,128,4096,0.012900267044703165
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,1536,0.005253333350022634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,1536,0.01092800001303355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,128,1024,0.0384554664293925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,128,3072,0.011083733042081196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,128,3584,0.012038399775822956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,1024,0.004424533247947693
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,1024,0.009551999966303508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,128,768,0.03314346671104431
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,768,0.004136533290147781
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,768,0.008871466914812724
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,128,512,0.02932586669921875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,128,2048,0.009180800120035807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,512,0.003772799919048945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,512,0.008257066706816356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,128,256,0.026055467128753663
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,128,1536,0.008285866677761078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,256,0.0035071998834609987
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,256,0.0076906666159629825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,128,128,0.02480640014012655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,128,1024,0.007112533350785573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,128,0.0031829332311948144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,128,0.007332266867160797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,128,64,0.023983999093373617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,64,0.003047466774781545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,64,0.007353599866231282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,128,32,0.023770666122436522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,32,0.0031690667072931922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,32,0.007314133147398631
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,64,65536,1.4044148763020834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,65536,0.05905386606852213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,65536,0.2582101345062256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,64,16384,0.3774357477823893
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,16384,0.012216533223787945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,16384,0.06338346799214681
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,64,12288,0.29001172383626306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,12288,0.010358400146166484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,12288,0.04282559951146443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,64,10240,0.2458944002787272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,10240,0.008872532844543457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,10240,0.03631893396377563
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,64,8192,0.19570560455322267
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,128,768,0.006605866551399231
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,128,512,0.006198399762312571
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,8192,0.007976533472537994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,8192,0.03025173346201579
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,7168,0.007748266557852428
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,64,7168,0.18508159319559733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,7168,0.027221333980560303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,6144,0.009241599837938945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,64,6144,0.15368746121724447
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,6144,0.02422399918238322
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,5120,0.008599467078844706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,64,5120,0.12474986712137859
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,5120,0.021346133947372437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,4096,0.007913599908351897
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,64,4096,0.0962719996770223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,4096,0.018283732732137046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,3584,0.00759680022795995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,64,3584,0.08747093677520752
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,3584,0.016835200786590575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,3072,0.007143466671307881
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,64,3072,0.07770453294118246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,3072,0.01548479994138082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,2560,0.0064522668719291685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,64,2560,0.06963093280792236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,2560,0.013885866602261862
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,64,2048,0.058261334896087646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,2048,0.005761066575845083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,2048,0.012389333049456278
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,64,1536,0.04820053180058797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,1536,0.005048533280690512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,1536,0.010867200295130412
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,64,1024,0.03728319803873698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,1024,0.004369066655635833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,1024,0.009530666470527648
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,64,768,0.03318079908688863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,768,0.004021333406368891
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,768,0.009000533819198608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,64,512,0.029294933875401812
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,512,0.003685333331425985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,512,0.00817386656999588
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,64,256,0.026152533292770386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,256,0.0032992000381151833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,256,0.007760000228881836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,64,128,0.02446613311767578
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,128,0.003160533308982849
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,128,0.007267199953397115
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,128,256,0.005723733206590017
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,64,64,0.024180267254511514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,64,0.00296426663796107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,64,0.007435733576615651
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,64,32,0.023965867360432942
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,32,0.0030218665798505146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,32,0.0073738664388656614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,65536,0.05695679982503256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,65536,0.2582922617594401
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,16384,0.011277866363525391
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,16384,0.06306879917780558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,12288,0.009830400347709656
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,32,16384,0.3772255897521973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,12288,0.04250986576080322
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,32,12288,0.28974294662475586
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,10240,0.009144533673922222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,10240,0.03631786505381267
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,32,10240,0.2455029328664144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,8192,0.007935999830563863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,32,65536,1.404529062906901
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,8192,0.030037333567937214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,32,8192,0.19544320106506347
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,7168,0.027101866404215497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,7168,0.007707733412583668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,32,7168,0.18456320762634276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,6144,0.007074133555094402
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,6144,0.024251733223597208
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,32,6144,0.15390292803446454
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,5120,0.007397333284219106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,32,5120,0.12336533069610596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,5120,0.021491199731826782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,4096,0.006988800068696339
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,32,4096,0.09724480311075846
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,3584,0.007380266487598419
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,4096,0.01827733318010966
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,32,3584,0.08830400307973227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,3072,0.007043200234572093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,3584,0.016889599959055583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,32,3072,0.0774826685587565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,3072,0.01544319987297058
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,32,2560,0.0698474645614624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,2560,0.006344533463319142
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,2560,0.013883733749389648
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,32,2048,0.05823359886805216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,2048,0.005603200197219849
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,2048,0.012268799543380737
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,32,1536,0.04816426833470662
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,1536,0.004977066814899444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,1536,0.010924800237019857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,32,1024,0.037028264999389646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,1024,0.0042698666453361515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,1024,0.009499733646710713
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,32,768,0.033292800188064575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,768,0.003966933240493139
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,768,0.00897920032342275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,32,512,0.029317333300908403
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,512,0.0036138666172822317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,512,0.008267733454704284
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,32,256,0.026096000274022417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,256,0.003291733314593633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,256,0.007613866527875264
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,32,128,0.024683733781178795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,128,0.003047466774781545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,128,0.007260799904664357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,32,64,0.024125866095225015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,64,0.0028789333999156954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,64,0.007264000177383423
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1024,32,32,0.023924267292022704
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,32,0.007260799904664357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,32,0.0029098667204380036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,65536,16384,1.489311981201172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,65536,16384,2.317858123779297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,65536,16384,1.252456537882487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,65536,12288,1.1455829620361329
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,65536,12288,0.944481086730957
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,65536,12288,1.793075180053711
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,65536,10240,0.7673877080281575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,65536,10240,0.9597951889038085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,65536,10240,0.7929525375366211
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,65536,10240,1.5122933705647787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,65536,8192,0.7404746373494466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,65536,8192,0.6302538553873698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,65536,16384,1.2182154337565103
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,65536,8192,0.6190773646036785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,65536,8192,1.1007413228352865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,65536,7168,0.6639167785644531
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,65536,7168,1.0254037221272787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,65536,12288,0.8689407984415689
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,65536,7168,0.5379445393880208
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,65536,6144,0.5764352162679036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,65536,6144,0.8931648254394531
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,65536,6144,0.46618560155232747
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,65536,5120,0.46989866892496746
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,65536,5120,0.3703370730082194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,65536,5120,0.388918399810791
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,65536,5120,0.728268814086914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,65536,4096,0.32098452250162757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,65536,4096,0.38533973693847656
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,65536,4096,0.31398293177286785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,65536,4096,0.5843168258666992
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,65536,3584,0.33558292388916017
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,65536,3584,0.48548161188761396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,65536,3584,0.27445545196533205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,65536,3072,0.29229866663614906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,65536,3072,0.22826879819234214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,65536,7168,0.5117621421813965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,65536,3072,0.4192842801411946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,65536,3072,0.22819306055704752
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,65536,2560,0.24636905988057456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,65536,2560,0.19641706148783367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,65536,2560,0.18986666997273763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,65536,2560,0.3542730649312337
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,65536,2048,0.20025386810302734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,65536,2048,0.1606239954630534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,65536,6144,0.44118827184041337
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,65536,2048,0.2888458569844564
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,65536,2048,0.16752746899922688
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,65536,1536,0.1578495979309082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,65536,1536,0.1263541301091512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,65536,1536,0.1210975964864095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,65536,1536,0.2191904067993164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,65536,1024,0.12322666645050048
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,65536,1024,0.09212586879730225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,65536,1024,0.1527114709218343
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,65536,768,0.07739840348561605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,65536,768,0.10806933244069417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,65536,768,0.12008960247039795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,65536,768,0.08112106323242188
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,65536,512,0.09677546819051107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,65536,512,0.06401066780090332
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,65536,512,0.08658133347829183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,65536,512,0.07251839637756348
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,65536,256,0.05501546859741211
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,65536,256,0.0841205358505249
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,65536,256,0.057852800687154135
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,65536,256,0.06048746506373087
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,65536,128,0.0795413335164388
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,65536,128,0.049628798166910806
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,65536,128,0.049958399931589764
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,65536,128,0.057239464918772374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,65536,64,0.080348801612854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,65536,64,0.04501333236694336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,65536,64,0.058183467388153075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,65536,32,0.0796895980834961
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,65536,32,0.0454090674718221
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,65536,32,0.05866026480992635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,65536,3584,0.2736522674560547
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,16384,65536,1.1612010955810548
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,16384,65536,2.1807711283365885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,65536,2.3793983459472656
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,65536,1.273895517985026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,16384,16384,0.5568192164103191
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,16384,0.536782964070638
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,16384,0.29290879567464195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,16384,12288,0.4306474685668945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,16384,12288,0.22403839429219566
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,12288,0.40023358662923175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,12288,0.23072427113850913
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,65536,1024,0.09291520118713378
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,16384,10240,0.3542197227478027
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,10240,0.3346336046854655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,10240,0.19397974014282227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,16384,8192,0.26748266220092776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,16384,8192,0.15283946990966796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,8192,0.269486935933431
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,8192,0.1464618682861328
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,16384,7168,0.2548170725504557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,16384,7168,0.13569067319234213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,7168,0.23655254046122232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,7168,0.12998826503753663
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,16384,6144,0.21842560768127442
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,6144,0.20404267311096191
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,6144,0.11465706825256347
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,16384,5120,0.1815989335378011
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,16384,5120,0.10083413124084473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,5120,0.17115839322408039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,5120,0.09699412981669107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,16384,4096,0.14933120409647624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,16384,4096,0.08242773214975993
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,16384,16384,0.30462506612141926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,4096,0.13838613828023275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,4096,0.07872107028961181
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,16384,3584,0.13311253388722738
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,3584,0.12208000024159749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,3584,0.06762239933013917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,16384,3072,0.11844480037689209
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,16384,3072,0.06393173138300577
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,3072,0.1061088005701701
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,3072,0.060071468353271484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,16384,2560,0.10108693440755208
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,16384,2560,0.05548160076141358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,16384,10240,0.18951360384623211
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,2560,0.09045866330464682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,2560,0.05528639952341715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,16384,2048,0.08594880104064942
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,2048,0.07370239893595378
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,2048,0.043825066089630126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,16384,1536,0.07242560386657715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,16384,1536,0.037979733943939206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,1536,0.05722026824951172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,1536,0.03563733498255412
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,16384,1024,0.05811946789423624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,16384,1024,0.029590400060017903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,1024,0.0410805344581604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,1024,0.02863466739654541
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,16384,768,0.051994665463765466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,16384,768,0.0254314661026001
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,768,0.033421866099039715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,768,0.025307732820510864
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,16384,512,0.04690560102462769
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,16384,512,0.020003199577331543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,512,0.026052266359329224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,512,0.021221333742141725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,16384,256,0.040072532494862874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,16384,256,0.014938666423161825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,256,0.017875200510025023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,256,0.017249067624409996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,16384,128,0.036710401376088456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,16384,128,0.013432533542315165
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,16384,6144,0.11848426659901937
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,128,0.013969066739082336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,128,0.015384533007939658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,16384,64,0.03522773186365764
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,64,0.012151466806729634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,64,0.016706132888793947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,16384,32,0.034560000896453856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,32,0.013022933403650919
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,32,0.01743040084838867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,12288,65536,0.9108746846516927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,128,128,0.005996799965699514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,16384,3584,0.07472000122070313
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,12288,65536,1.9250709533691406
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,65536,1.0472949345906577
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,12288,16384,0.4916693369547526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,65536,1.7285451253255206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,16384,0.23241705894470216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,16384,0.38934081395467124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,12288,0.29372587203979494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,12288,12288,0.39009173711140954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,12288,0.17897706031799315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,12288,10240,0.14122559229532877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,12288,10240,0.31950505574544275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,16384,2048,0.04577813148498535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,10240,0.15453866322835286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,10240,0.2456607977549235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,12288,8192,0.11699946721394856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,12288,8192,0.23992533683776857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,8192,0.1981536070505778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,8192,0.12305173079172771
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,12288,7168,0.10181653499603271
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,12288,7168,0.2279029369354248
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,7168,0.17709439595540363
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,7168,0.10498027006785075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,12288,6144,0.09008533159891764
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,6144,0.1497909386952718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,12288,6144,0.19479680061340332
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,6144,0.09443199634552002
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,12288,5120,0.16045546531677246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,12288,5120,0.07530879974365234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,5120,0.1281599998474121
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,5120,0.0801578680674235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,12288,4096,0.13279786904652913
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,12288,4096,0.061792000134785974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,4096,0.10359040101369221
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,4096,0.0652405341466268
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,12288,3584,0.11971200307210286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,3584,0.09165546894073487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,3584,0.05646933317184448
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,12288,3072,0.10600639979044597
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,3072,0.07919253508249918
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,3072,0.05030080080032349
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,12288,2560,0.09023253122965494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,12288,2560,0.04312320152918498
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,2560,0.06693867047627768
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,128,128,0.005505066613356272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,2560,0.0440661350886027
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,12288,2048,0.07721280256907145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,12288,2048,0.036261332035064694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,2048,0.05422293345133463
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,2048,0.036968533198038736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,12288,1536,0.06518400112787882
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,1536,0.042438399791717527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,1536,0.02969493269920349
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,12288,1024,0.051991466681162515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,12288,16384,0.22360960642496744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,1024,0.03086293339729309
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,1024,0.02332693338394165
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,12288,12288,0.17464426358540852
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,12288,768,0.04589866797129313
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,768,0.024156800905863442
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,768,0.020269866784413657
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,12288,512,0.04102186759312947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,12288,512,0.016586666305859886
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,512,0.01797013282775879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,512,0.01754666765530904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,12288,256,0.03402986526489258
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,12288,256,0.012820266683896384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,256,0.01267519990603129
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,256,0.014653866489728292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,12288,128,0.03267733256022136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,12288,128,0.011173333724339802
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,128,0.009939199686050415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,128,0.013590400417645773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,12288,64,0.03264213403065999
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,64,0.008406399687131246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,64,0.014371200402577718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,12288,32,0.03221546610196431
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,32,0.008784000078837078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,32,0.014890666802724203
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,12288,3584,0.055700266361236574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,12288,3072,0.049428268273671465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,10240,65536,0.7639488220214844
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,10240,65536,1.7139466603597004
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,128,128,0.0067669332027435304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,10240,16384,0.19883413314819337
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,65536,0.8064405441284179
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,10240,16384,0.445362122853597
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,65536,1.4661418914794921
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,16384,0.21264853477478027
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,16384,0.3414250691731771
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,12288,0.14583679835001628
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,10240,12288,0.34375359217325846
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,12288,0.2855488141377767
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,128,128,0.005625600119431814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,10240,10240,0.12691840330759685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,10240,0.21247679392496743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,10240,10240,0.28587306340535484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,10240,0.1250421365102132
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,10240,8192,0.21644585927327475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,8192,0.09817386468251546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,8192,0.17291520436604818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,10240,7168,0.2087519963582357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,10240,7168,0.09077333609263102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,12288,1536,0.030296534299850464
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,7168,0.15191359519958497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,7168,0.08936533133188883
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,10240,6144,0.08016106287638346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,10240,6144,0.1742784023284912
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,6144,0.1307978630065918
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,6144,0.07612160046895346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,12288,1024,0.022523732980092366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,5120,0.11016426881154377
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,10240,5120,0.14477119445800782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,5120,0.06437120040257772
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,10240,4096,0.11923306783040363
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,4096,0.08942613601684571
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,4096,0.053718400001525876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,10240,3584,0.10871360301971436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,3584,0.0781279961268107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,3584,0.04746666749318441
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,10240,3072,0.09777173201243082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,12288,768,0.019627734025319418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,3072,0.06778879960378012
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,3072,0.04277226527531942
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,10240,2560,0.08343573411305746
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,10240,2560,0.03905066649119059
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,2560,0.057564798990885416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,2560,0.03676799933115642
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,10240,2048,0.07208746274312337
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,10240,2048,0.03262293338775635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,2048,0.0472490668296814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,2048,0.030253867308298748
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,10240,1536,0.060994132359822595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,10240,1536,0.026130133867263795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,1536,0.03671786785125732
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,1536,0.02467413345972697
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,10240,1024,0.04816746711730957
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,10240,1024,0.019899733861287437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,1024,0.02577706575393677
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,1024,0.018936532735824584
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,10240,768,0.0430890679359436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,10240,768,0.017387733856836955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,768,0.021005866924921672
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,768,0.016999467213948568
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,10240,512,0.03686079978942871
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,10240,512,0.014935466647148132
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,512,0.016268799702326454
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,512,0.015010133385658264
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,10240,256,0.03273386756579082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,10240,256,0.011686399579048157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,256,0.0115146666765213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,256,0.01304213305314382
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,10240,128,0.031086933612823487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,10240,128,0.010941867033640544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,128,0.01200320025285085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,128,0.012075733145078022
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,64,0.009701333443323771
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,10240,64,0.031009066104888915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,64,0.012682666381200155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,10240,32,0.03025813301404317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,32,0.010284800330797832
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,32,0.012964266538619994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,8192,65536,0.6235701243082683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,10240,12288,0.15138880411783856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,8192,65536,1.550175984700521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,10240,8192,0.10237013498942057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,65536,0.6374368031819662
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,65536,1.10960324605306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,8192,16384,0.15362559954325358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,16384,0.2758218765258789
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,16384,0.15780372619628907
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,8192,16384,0.40209067662556963
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,10240,5120,0.06735786596934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,8192,12288,0.3094208081563314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,12288,0.21739840507507324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,12288,0.11840533415476481
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,8192,10240,0.0988106648127238
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,8192,10240,0.2576927979787191
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,10240,4096,0.05592106580734253
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,10240,0.09703466892242432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,10240,0.17782185872395834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,8192,8192,0.1946069399515788
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,8192,0.07597866853078207
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,8192,0.1387104034423828
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,8192,7168,0.07123733361562093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,10240,3584,0.05033493439356486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,8192,7168,0.1942421277364095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,7168,0.12123093605041504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,7168,0.06897386709849039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,8192,6144,0.06316693226496378
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,8192,6144,0.15528853734334308
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,6144,0.10775146484375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,6144,0.06051413218180338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,10240,3072,0.044299733638763425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,8192,5120,0.13096319834391276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,5120,0.08681920369466146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,8192,4096,0.10892159938812256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,5120,0.05169386863708496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,4096,0.07148479620615641
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,4096,0.04376426537831624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,8192,3584,0.0994815985361735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,3584,0.06320639848709106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,3584,0.037544532616933184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,8192,3072,0.08919680118560791
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,3072,0.05410240093866984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,3072,0.034306132793426515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,8192,2560,0.07599146366119384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,8192,2560,0.031515733400980635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,2560,0.04591360092163086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,2560,0.02911786635716756
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,8192,2048,0.06581759850184123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,8192,2048,0.025755733251571655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,2048,0.037436799208323164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,2048,0.023145600159962972
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,8192,1536,0.056032001972198486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,8192,1536,0.021435733636220297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,1536,0.028914133707682293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,1536,0.019099734226862588
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,8192,1024,0.04241919914881388
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,8192,1024,0.01704746683438619
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,1024,0.020541866620381675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,1024,0.016167466839154564
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,8192,768,0.037973332405090335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,8192,768,0.015154133240381876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,768,0.01680319905281067
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,768,0.013863466183344521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,8192,512,0.03388479948043823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,512,0.013099732995033263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,512,0.01263146698474884
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,8192,256,0.030686932802200317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,8192,256,0.010787199934323628
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,256,0.009485866626103718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,256,0.011179733276367187
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,8192,128,0.029260800282160444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,8192,128,0.009805867075920105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,128,0.007663999994595845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,128,0.010440533359845478
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,8192,64,0.029528532425562543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,64,0.006684799989064534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,64,0.010972799857457478
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,8192,32,0.028882133960723876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,32,0.006850133339564006
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,32,0.011336533228556316
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,8192,12288,0.11904213428497315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,8192,8192,0.0800490697224935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,65536,0.9473941167195639
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,8192,5120,0.05398720105489095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,7168,65536,1.5051050821940104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,8192,4096,0.04437226851781209
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,8192,3584,0.041341865062713624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,65536,0.6069248199462891
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,7168,16384,0.39541333516438804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,8192,3072,0.03568106492360433
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,16384,0.14629012743632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,7168,16384,0.147435728708903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,16384,0.23793279329935707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,12288,0.11271039644877116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,12288,0.1799221356709798
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,7168,12288,0.31058454513549805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,7168,10240,0.25324692726135256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,10240,0.15414826075236004
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,10240,0.095741868019104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,7168,8192,0.07700800100962321
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,7168,8192,0.19340267181396484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,8192,0.12406293551127116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,8192,0.07522559960683187
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,7168,7168,0.18648427327473957
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,7168,0.10835733413696289
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,7168,0.06680853366851806
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,7168,6144,0.1532704035441081
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,6144,0.09298773606618246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,6144,0.05824746688206991
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,8192,512,0.013063466548919678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,7168,5120,0.0513589342435201
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,7168,5120,0.1284063975016276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,5120,0.07656853199005127
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,5120,0.05024853150049845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,7168,4096,0.04291413227717082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,7168,4096,0.10664746761322022
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,4096,0.06246933142344156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,4096,0.04136319955190022
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,7168,3584,0.09666666984558106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,7168,3584,0.0389514684677124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,3584,0.054915201663970944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,3584,0.03751786549886067
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,7168,3072,0.03459200064341227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,7168,3072,0.08798720041910807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,3072,0.048153599103291825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,3072,0.03292160034179688
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,7168,2560,0.0761578639348348
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,2560,0.03996906677881877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,2560,0.028100266059239702
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,7168,2048,0.06572266817092895
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,7168,2048,0.025654399394989015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,2048,0.032525867223739624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,2048,0.022507733106613158
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,7168,1536,0.05552639961242676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,7168,1536,0.021191465854644775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,1536,0.025197867552439374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,1536,0.01911999980608622
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,7168,1024,0.042455466588338216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,7168,1024,0.01684160033861796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,1024,0.017989333470662436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,1024,0.01597653329372406
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,7168,768,0.03813440004984538
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,7168,768,0.01520639955997467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,768,0.01467626690864563
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,768,0.014260266224543253
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,7168,512,0.03442666530609131
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,7168,512,0.013064533472061157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,512,0.01128000020980835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,512,0.012618666887283326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,7168,256,0.03043839931488037
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,7168,256,0.010709333419799804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,256,0.008806399504343669
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,256,0.01111466685930888
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,7168,128,0.029014400641123456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,7168,128,0.009617066383361817
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,128,0.007807999849319458
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,128,0.010295466581980387
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,7168,64,0.02911679943402608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,64,0.006676266590754191
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,64,0.010915199915568035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,7168,32,0.028621866305669146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,32,0.0068256000677744556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,32,0.011143466830253601
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,6144,65536,1.4993653615315756
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,7168,65536,0.5760757446289062
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,6144,65536,0.5975285212198893
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,65536,0.7796021143595377
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,6144,16384,0.15304640134175618
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,65536,0.6220767974853516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,6144,16384,0.41414826711018876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,7168,12288,0.11306133270263671
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,16384,0.19585919380187988
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,16384,0.14447466532389325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,7168,7168,0.06800639629364014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,12288,0.14762239456176757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,6144,12288,0.308132266998291
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,7168,10240,0.0946773370107015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,12288,0.11045440038045247
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,6144,10240,0.24899306297302246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,10240,0.09503786563873291
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,10240,0.12534399827321369
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,6144,8192,0.1915082613627116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,8192,0.09983893235524496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,7168,6144,0.06105386813481649
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,8192,0.07380053202311197
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,7168,0.09041706720987955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,6144,7168,0.18535253206888835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,7168,0.06771093209584554
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,6144,6144,0.15195306142171222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,6144,0.07710399627685546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,6144,0.05842666625976563
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,6144,5120,0.12871572971343995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,5120,0.06483840147654216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,5120,0.05025706688563029
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,6144,4096,0.10645973682403564
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,7168,2560,0.030373332897822063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,4096,0.05241066614786784
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,4096,0.041069865226745605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,6144,3584,0.09654506842295328
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,6144,8192,0.07791679700215658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,6144,12288,0.11641173362731934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,3584,0.045607467492421463
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,6144,10240,0.09821013609568277
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,3584,0.0363103985786438
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,6144,3072,0.08631146748860677
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,6144,3072,0.035035733381907144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,3072,0.04040960073471069
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,3072,0.030771199862162275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,6144,2560,0.07534506320953369
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,6144,2560,0.03031040032704671
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,2560,0.03370346625645955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,6144,6144,0.061349332332611084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,6144,2048,0.0650816003481547
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,2560,0.02681066592534383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,2048,0.028275199731190997
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,2048,0.022168533007303873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,6144,7168,0.07010666529337564
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,6144,1536,0.054858668645222985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,6144,1536,0.020895999670028687
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,1536,0.02144320011138916
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,6144,5120,0.05328640143076578
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,1536,0.018915200233459474
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,6144,1024,0.04245866537094116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,1024,0.01580586632092794
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,1024,0.01585599978764852
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,6144,4096,0.044072532653808595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,6144,768,0.03784533341725667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,6144,768,0.015100799997647605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,6144,512,0.034135464827219644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,768,0.013172266880671182
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,768,0.013963733116785684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,512,0.010607999563217164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,6144,256,0.029710932572682695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,512,0.012537599603335062
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,256,0.007878399888674418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,256,0.010855467120806376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,6144,3584,0.03940373261769613
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,6144,128,0.02879040042559306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,6144,128,0.009529599547386169
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,128,0.006487466891606649
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,128,0.010212266445159912
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,6144,64,0.028780800104141236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,64,0.0058218667904535925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,64,0.010732799768447876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,32,0.010614400108655293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,6144,32,0.02798186739285787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,32,0.0060693333546320595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,65536,0.45653120676676434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,65536,0.6741568247477214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,5120,16384,0.35313920974731444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,5120,65536,1.321826171875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,16384,0.17195413907368978
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,16384,0.10509440104166667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,5120,12288,0.2716789245605469
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,5120,12288,0.08272319634755453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,12288,0.13018346627553304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,12288,0.08051306406656901
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,5120,10240,0.07011199792226155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,5120,10240,0.2222218672434489
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,10240,0.10829439957936604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,10240,0.06727573076883951
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,5120,8192,0.16758720080057782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,6144,2048,0.02532373269399007
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,8192,0.054670933882395426
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,8192,0.0881440003712972
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,5120,7168,0.16473813056945802
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,6144,1024,0.01721386710802714
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,7168,0.075709867477417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,7168,0.04882133404413859
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,5120,6144,0.04528533220291138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,5120,6144,0.1372586727142334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,6144,0.06587093273798625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,6144,0.04273173411687215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,5120,5120,0.1153429349263509
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,5120,0.05560106833775839
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,5120,0.03729493220647176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,6144,512,0.01284160017967224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,5120,4096,0.09610346953074136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,6144,256,0.010358400146166484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,4096,0.04540479977925618
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,4096,0.030101333061854047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,5120,3584,0.08809066613515218
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,3584,0.040133333206176756
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,3584,0.026581333080927534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,5120,3072,0.07956053415934244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,5120,3072,0.026250666379928587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,3072,0.03498560190200806
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,3072,0.023033599058787026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,5120,2560,0.06921172936757405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,5120,65536,0.4086079915364583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,2560,0.029892265796661377
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,2560,0.020500266551971437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,5120,16384,0.10827840169270833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,5120,2048,0.060710398356119785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,2048,0.023921066522598268
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,2048,0.017413334051767985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,5120,1536,0.04942400058110555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,5120,1536,0.016570666432380678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,1536,0.018920532862345376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,1536,0.014753066500027976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,5120,1024,0.04015680154164632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,5120,1024,0.013578666249910989
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,1024,0.014245333274205527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,1024,0.012418133020401002
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,5120,768,0.03535466591517131
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,5120,768,0.012318933010101318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,768,0.012025599678357441
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,768,0.011293866237004598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,5120,512,0.03187626600265503
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,5120,512,0.011001599828402202
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,5120,8192,0.057190398375193276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,512,0.009699199597040813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,512,0.010213333368301391
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,5120,256,0.028537599245707194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,5120,256,0.009480533003807069
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,256,0.007426133255163829
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,256,0.009291733304659527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,5120,128,0.027383466561635334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,5120,128,0.008589866757392883
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,128,0.006132266422112783
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,128,0.008952533205350239
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,5120,64,0.027269333600997925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,64,0.0055285334587097164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,64,0.009202133615811665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,5120,32,0.026796799898147584
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,32,0.0055637334783871974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,32,0.009372799595197042
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,4096,65536,1.2968981424967447
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,65536,0.5346922556559245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,5120,7168,0.05145066579182943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,65536,0.43453226089477537
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,4096,16384,0.347380256652832
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,5120,5120,0.03890346686045329
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,4096,16384,0.08676479657491049
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,16384,0.13729599316914876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,16384,0.10121599833170573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,12288,0.10151360034942628
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,4096,12288,0.2697728157043457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,12288,0.07949120203653971
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,4096,10240,0.21791359583536782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,10240,0.08644479910532633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,5120,4096,0.03248213330904643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,10240,0.06531840165456136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,5120,3584,0.029734400908152263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,4096,8192,0.04583253463109334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,4096,8192,0.16637120246887208
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,8192,0.07124266624450684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,8192,0.0538154681523641
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,7168,0.06121493180592855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,4096,7168,0.16266454060872396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,7168,0.04795519908269246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,4096,6144,0.036260267098744706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,4096,6144,0.13356693585713703
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,6144,0.05292266607284546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,6144,0.04146453142166138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,5120,2560,0.0228000005086263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,4096,5120,0.11298453013102214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,5120,0.044232531388600664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,5120,0.03490560054779053
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,4096,4096,0.09423253536224366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,4096,4096,0.02590186595916748
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,4096,0.03602773348490397
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,5120,2048,0.019811199108759562
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,4096,0.028536534309387206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,4096,3584,0.08625813325246176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,4096,3584,0.024077866474787393
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,3584,0.025471999247868853
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,3584,0.03201066652933757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,4096,3072,0.07857813040415446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,3072,0.027425066630045576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,3072,0.022653865814208984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,4096,2560,0.06810452938079833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,4096,2560,0.019033600886662803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,2560,0.023337600628534953
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,2560,0.01998186707496643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,4096,2048,0.05915946563084921
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,2048,0.019131733973821004
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,2048,0.017335466543833413
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,4096,1536,0.04983786741892497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,4096,1536,0.014146133263905843
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,1536,0.01542080044746399
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,1536,0.014709333578745524
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,4096,1024,0.039129598935445147
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,4096,1024,0.011864533027013142
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,1024,0.011760000387827556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,1024,0.012389333049456278
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,4096,65536,0.3309642791748047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,4096,768,0.035396265983581546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,768,0.009904000163078307
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,768,0.011338667074839274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,4096,512,0.03198613325754802
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,4096,512,0.009842133522033692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,512,0.00808426688114802
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,512,0.010140800476074218
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,4096,256,0.028618667523066205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,4096,256,0.008515200018882752
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,256,0.0063498665889104204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,256,0.009242666761080424
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,4096,128,0.027438932657241823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,4096,128,0.007843199868996937
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,128,0.0053962667783101406
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,128,0.008716799815495809
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,4096,64,0.027268266677856444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,64,0.004717866579691568
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,4096,12288,0.06628906726837158
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,64,0.009041066964467366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,4096,32,0.026469333966573076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,32,0.004960000018278758
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,32,0.00920853316783905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,4096,10240,0.05534186760584513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,65536,0.4596927960713704
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,4096,7168,0.04229226509730021
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,65536,0.42509972254435224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3584,65536,1.2946954091389975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3584,16384,0.3465973218282064
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,16384,0.11940800348917643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,16384,0.10218026638031005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,4096,5120,0.031684267520904544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3584,12288,0.26790294647216795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,12288,0.09046080112457275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,12288,0.07666773001352946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3584,10240,0.0535103996594747
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3584,10240,0.21677227020263673
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,10240,0.07606613636016846
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,10240,0.06390186548233032
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3584,8192,0.16504747072855633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,8192,0.06120213270187378
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,4096,3072,0.021236266692479452
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,8192,0.05238720178604126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3584,7168,0.03960106770197551
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3584,7168,0.16085866292317708
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,7168,0.05401920080184937
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,7168,0.0466538667678833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3584,6144,0.13447999954223633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,6144,0.046722133954366044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,6144,0.04132800102233887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,4096,2048,0.016424533724784852
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3584,5120,0.11299306551615398
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3584,5120,0.03097066680590312
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,5120,0.03963093360265096
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,5120,0.03402560154596965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3584,4096,0.09461867014567057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,4096,0.03253759940465291
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,4096,0.027705599864323933
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3584,3584,0.0868287960688273
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3584,3584,0.023269333442052207
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,3584,0.02887786626815796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,3584,0.025304534037907916
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3584,3072,0.0788810650507609
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3584,3072,0.020824533700942994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,3072,0.024975999196370443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,3072,0.022743467489878336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3584,2560,0.06802773475646973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3584,2560,0.018528000513712565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,2560,0.021321600675582884
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,2560,0.02008426586786906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3584,2048,0.059018667538960776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3584,2048,0.01634239951769511
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,2048,0.017909334103266398
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,4096,768,0.010847999652226766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,2048,0.01735573410987854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3584,1536,0.05031786759694418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3584,65536,0.3162048021952311
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,1536,0.01442026694615682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,1536,0.014506666858990987
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3584,1024,0.03896213372548421
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3584,1024,0.011711999773979187
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,1024,0.011205333471298217
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,1024,0.012288000186284382
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3584,768,0.03469333251317342
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3584,768,0.010734933614730834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,768,0.009547733267148336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,768,0.011224533120791119
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3584,16384,0.08220160007476807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3584,512,0.03139413396517436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3584,512,0.009794132908185323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,512,0.008107733229796093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,512,0.01011840005715688
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3584,256,0.02858026623725891
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,256,0.006540800134340922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,256,0.009326933821042379
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3584,128,0.02684906721115112
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3584,12288,0.06488853295644124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3584,128,0.007785599927107494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,128,0.0056533331672350565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,128,0.00862506628036499
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3584,64,0.027237333854039508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,64,0.005207466582457224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,64,0.009106133381525676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3584,32,0.026602667570114136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,32,0.0054282665252685545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,32,0.0090037335952123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3584,8192,0.04473813374837239
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,65536,0.4109813372294108
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,65536,0.4117727915445964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3584,6144,0.03527466853459676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3072,16384,0.3525312105814616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,16384,0.10348373254140217
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3072,65536,1.2893770853678386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,16384,0.10445653597513835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,12288,0.07996479670206705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3072,12288,0.06492693424224853
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3072,12288,0.26669333775838217
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,12288,0.07891200383504232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3584,4096,0.025250132878621417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,10240,0.06724692980448405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3072,10240,0.21618240674336753
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,10240,0.0639082670211792
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3072,8192,0.04466240008672078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3072,8192,0.16518294016520182
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,8192,0.052254935105641685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,8192,0.051774934927622474
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3072,7168,0.1622378667195638
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,7168,0.04630399942398071
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,7168,0.047870934009552
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3072,6144,0.13342827161153156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,6144,0.04015573263168335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,6144,0.03882559935251872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3072,5120,0.11200533707936604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3072,5120,0.030349866549173994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,5120,0.03404479821523031
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,5120,0.032841600974400836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3072,4096,0.09432319800059
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3072,4096,0.025545599063237508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,4096,0.027923200527826948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,4096,0.027923200527826948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3072,3584,0.08647680282592773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3072,3584,0.0232586661974589
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,3584,0.024629332621892295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,3584,0.025573333104451496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3072,3072,0.07891200383504232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3072,3072,0.020933334032694498
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,3072,0.021848533550898233
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,3072,0.022776534159978233
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3072,2560,0.06771732966105143
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3072,2560,0.018633600076039633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,2560,0.01825066606203715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,2560,0.02013333241144816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3072,2048,0.058964268366495765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3584,1536,0.01393280029296875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3072,2048,0.016290133198102318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,2048,0.01727786660194397
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,2048,0.015401599804560342
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3072,1536,0.04952640136082967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,1536,0.012452266613642375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,1536,0.014604799946149192
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3072,1024,0.038762664794921874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3072,1024,0.012125866611798604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,1024,0.00988266666730245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,1024,0.01220373312632243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3072,768,0.03492266734441121
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3072,768,0.010909866293271382
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,768,0.008166400094827015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3584,256,0.00844266712665558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,768,0.011081600189208984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3072,512,0.03157653411229451
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3072,512,0.009711999694506328
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,512,0.006853333115577698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,512,0.010199466347694397
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3072,256,0.02821226716041565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3072,256,0.008378666639328004
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,256,0.005606399973233541
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,256,0.009180800120035807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3072,128,0.027078400055567425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3072,16384,0.08525226910909017
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3072,65536,0.32812798817952477
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,128,0.004841599861780802
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,128,0.00855573316415151
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3072,64,0.02688746651013692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,64,0.004350933432579041
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,64,0.009063466389973959
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3072,10240,0.05457813342412313
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,3072,32,0.0263264000415802
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,32,0.0045855998992919925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,32,0.008770133058230083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3072,7168,0.04058239857355754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,65536,0.26600640614827475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2560,16384,0.3046047846476237
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,65536,0.3507285435994466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2560,16384,0.06539200146993002
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3072,6144,0.0362335999806722
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,16384,0.08806186517079671
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,16384,0.06223893165588379
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,12288,0.06702293554941813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,12288,0.04896959861119588
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2560,12288,0.23744853337605795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2560,65536,1.1398016611735025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2560,10240,0.19078720410664876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,10240,0.056599466005961094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,10240,0.040440531571706136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2560,8192,0.03534293174743652
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2560,8192,0.14594772656758626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,8192,0.046187734603881835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,8192,0.033580799897511796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2560,7168,0.1450549284617106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,7168,0.04037333329518636
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,7168,0.029501867294311524
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2560,6144,0.027806933720906573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2560,6144,0.12081813017527263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,6144,0.03497920036315918
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,6144,0.024317866563796996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2560,5120,0.023669334252675374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2560,5120,0.10065173308054606
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,5120,0.02969493269920349
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,5120,0.021441066265106203
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2560,4096,0.0850549300511678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,4096,0.02394346594810486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,4096,0.018281600872675576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2560,3584,0.07841066519419351
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2560,3584,0.018219733238220216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,3584,0.02099840044975281
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,3584,0.016936532656351724
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2560,3072,0.07158613204956055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2560,3072,0.016581333676973977
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,3072,0.01812053322792053
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,3072,0.01548693378766378
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2560,2560,0.06138773361841837
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2560,2560,0.01511253317197164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,2560,0.01572480003039042
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,2560,0.014064000050226847
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2560,2048,0.053641601403554284
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3072,1536,0.013758933544158936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2560,2048,0.01333440045515696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,2048,0.013454932967821756
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,2048,0.012380799651145935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2560,1536,0.011572266618410747
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2560,1536,0.04650133450826009
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,1536,0.010974933703740437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,1536,0.011011200149854024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2560,1024,0.03626240094502767
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2560,1024,0.009975467125574749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,1024,0.008665600419044494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,1024,0.009664000074068705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2560,768,0.009204266468683879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2560,768,0.032891732454299924
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,768,0.007577600081761678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,768,0.009109333157539368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2560,512,0.029623466730117797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3072,128,0.0076885332663853955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2560,512,0.008498133222262064
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,512,0.006325333317120869
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,512,0.00848533312479655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2560,256,0.027130667368570966
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,256,0.00516480008761088
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,256,0.007870933413505555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2560,128,0.025764266649882
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2560,128,0.006859733164310456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,128,0.004558933277924856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,128,0.007402666906515758
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2560,64,0.02544533411661784
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,64,0.004347733159859975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,64,0.0077002664407094315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2560,32,0.025256532430648803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,32,0.004446933170159658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,32,0.007732266684373219
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2048,65536,0.25839254061381023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,65536,0.27309331893920896
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2048,65536,1.1180693308512368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,65536,0.24488213857014976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2048,16384,0.06813866297403971
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2048,16384,0.30396906534830725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,16384,0.09904853502909342
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,16384,0.06028159856796265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2048,12288,0.23413227399190267
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2048,12288,0.053445335229237875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2560,65536,0.2516618728637695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,12288,0.07331306934356689
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,12288,0.047430400053660074
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2048,10240,0.18874133427937825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2560,12288,0.051520001888275144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,10240,0.06189759969711304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2560,10240,0.04349013169606526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,10240,0.040770133336385087
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2048,8192,0.1454367955525716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,8192,0.045960533618927005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,8192,0.031872000296910605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2048,7168,0.033666133880615234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2048,7168,0.14584852854410807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2560,7168,0.03224959969520569
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,7168,0.04350293477376302
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,7168,0.027532800038655596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2048,6144,0.028548266490300494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2048,6144,0.12047253449757893
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,6144,0.03640213410059611
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,6144,0.024235733350118003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2048,5120,0.10122559865315754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,5120,0.030142933130264282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,5120,0.021441066265106203
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2048,4096,0.08486826419830322
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2560,4096,0.020053333044052123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,4096,0.01818880041440328
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,4096,0.020056533813476562
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2048,3584,0.0778218666712443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2048,3584,0.01926506757736206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,3584,0.017220266660054526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,3584,0.017017600933710735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2048,3072,0.01730666756629944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2048,3072,0.0714698632558187
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,3072,0.016420267025629678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,3072,0.015317333738009134
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2048,2560,0.061709864934285485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2048,2560,0.015546666582425437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,2560,0.014644267161687216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,2560,0.014035200079282125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2048,2048,0.053933866818745936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,2048,0.01241600016752879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,2048,0.012599466244379678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2048,1536,0.04424639940261841
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2048,1536,0.01225386659304301
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,1536,0.009870933492978413
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,1536,0.011153067151705425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2048,1024,0.03614613215128581
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2048,1024,0.010233599940935771
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,1024,0.007426133255163829
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,1024,0.009656533598899841
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2048,768,0.03299306631088257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2048,768,0.009125333031018574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,768,0.006450133522351582
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,768,0.00913706620534261
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2048,512,0.029706666866938274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2048,512,0.008187733093897502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,512,0.005605333546797434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,512,0.008518399794896443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2048,256,0.026748800277709962
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,256,0.004737066725889841
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,256,0.007831466694672901
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2048,128,0.02558079957962036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2560,256,0.007402666906515758
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,128,0.0040949332217375435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,128,0.007286400099595388
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2048,64,0.025135999917984007
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,64,0.003739733248949051
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,64,0.007523199915885926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,2048,32,0.024740266799926757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,32,0.0039007999002933502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,32,0.007561600208282471
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2048,8192,0.037096532185872395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1536,65536,1.1157396952311198
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1536,65536,0.23522133827209474
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,65536,0.2087519963582357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,65536,0.23448425928751626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1536,16384,0.30369173685709633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2048,5120,0.024388267596562704
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,16384,0.06078826586405436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,16384,0.05904959837595621
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1536,12288,0.04964480002721151
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1536,12288,0.23367573420206705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,12288,0.044997334480285645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,12288,0.04623466730117798
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2048,10240,0.045201067129770914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1536,10240,0.18853119214375813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,10240,0.03703573147455851
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,10240,0.0397546648979187
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2048,4096,0.021388800938924153
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2048,2048,0.013763200243314108
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1536,8192,0.14548373222351074
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1536,8192,0.03529706796010335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,8192,0.030294400453567506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,8192,0.029813333352406816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1536,7168,0.14481706619262696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,7168,0.027191466093063353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,7168,0.02683626612027486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1536,6144,0.11859839757283527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1536,6144,0.027130667368570966
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,6144,0.022285866737365722
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,6144,0.023400533199310302
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1536,5120,0.10006720225016277
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,5120,0.019269333283106486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,5120,0.02111039956410726
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1536,4096,0.08418880303700765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1536,4096,0.02020906607309977
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2048,128,0.00650133341550827
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,4096,0.01572480003039042
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,4096,0.018054399887720743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1536,3584,0.07788693110148112
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1536,3584,0.01845226685206095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,3584,0.014338133732477823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,3584,0.016730666160583496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1536,3072,0.07136853535970053
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,3072,0.01271573305130005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,3072,0.015243732929229736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2048,256,0.007084799806276958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1536,2560,0.0611402670542399
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1536,2560,0.015044266978899637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,2560,0.011366400122642516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,2560,0.013821867108345032
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1536,2048,0.05322879950205485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,2048,0.00976213316122691
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,2048,0.01220906674861908
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1536,1536,0.04440853198369344
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1536,1536,0.011782399813334147
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,1536,0.008628267049789428
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,1536,0.01088213324546814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1536,1024,0.03589653174082438
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,1024,0.007829333345095318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,1024,0.00958720048268636
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1536,768,0.03274666666984558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1536,768,0.009227733810742695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,768,0.0067071999112765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,768,0.009035733342170716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1536,512,0.029470932483673096
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1536,16384,0.06429546674092611
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,512,0.005474133292833964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,512,0.008364799618721008
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1536,256,0.026715733607610065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1536,10240,0.043237332503000894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1536,5120,0.023517866929372154
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,256,0.0043818667531013485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,256,0.007713066538174947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1536,128,0.025568000475565594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,128,0.0038560000558694207
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1536,7168,0.0323253333568573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,128,0.007256533205509186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1536,64,0.024987733364105223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,64,0.003692800054947535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,64,0.007412266731262207
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1536,32,0.02479040026664734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,32,0.0037674665451049804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,32,0.00746559997399648
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1536,3072,0.016759467124938966
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,65536,0.14441919326782227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1024,65536,1.1095743815104167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,65536,0.23193705876668297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1024,16384,0.054390398661295566
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,16384,0.042131201426188154
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1024,16384,0.30209280649820963
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,16384,0.059418666362762454
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1024,12288,0.2319925308227539
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,12288,0.053123199939727785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1536,2048,0.013480533162752786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,12288,0.04532373348871867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1024,10240,0.18770666122436525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,10240,0.04196586608886719
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,10240,0.03587626616160075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1024,8192,0.14527360598246258
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1536,1024,0.010456533233324686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1024,8192,0.02784000039100647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,8192,0.035225598017374675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,8192,0.02917120059331258
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1024,7168,0.024613332748413087
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1024,7168,0.14335467020670573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,7168,0.030083199342091877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,7168,0.026961066325505573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1024,6144,0.11895893414815266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,6144,0.026191999514897663
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,6144,0.023411200443903605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1536,512,0.00795306662718455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1024,5120,0.09940480391184489
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,5120,0.02227413256963094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,5120,0.020830933252970377
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1024,4096,0.08372906843821207
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1536,128,0.006346666812896728
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,4096,0.014444800217946372
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1536,256,0.006764799853165944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,4096,0.017697066068649292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1024,3584,0.07761173248291016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,3584,0.012382933497428894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,3584,0.01658986707528432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1024,3072,0.07114880084991455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1024,65536,0.19769387245178222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,3072,0.011432533462842304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1024,12288,0.04139413436253865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,3072,0.01520639955997467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1024,2560,0.06122666597366333
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1024,10240,0.034560000896453856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,2560,0.009431466460227966
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,2560,0.013707733154296875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1024,2048,0.05246719916661581
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1024,6144,0.02217386762301127
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,2048,0.00867733359336853
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,2048,0.012089600165685017
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1024,5120,0.018879999717076622
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1024,1536,0.04349013169606526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,1536,0.007454933226108551
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,1536,0.010792533556620281
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1024,1024,0.035860268274943034
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1024,1024,0.008249600231647492
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,1024,0.005714133381843567
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,1024,0.009459199508031209
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1024,768,0.03256640036900838
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1024,4096,0.016570666432380678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1024,768,0.007740800082683563
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,768,0.0052149335543314615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,768,0.008851200342178345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1024,512,0.029611732562383013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,512,0.004598399996757508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,512,0.008259200056393941
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1024,256,0.026405332485834758
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,256,0.004005333284536997
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,256,0.00769706666469574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1024,128,0.025297067562739056
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1024,128,0.005916800101598104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,128,0.0036256000399589538
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,128,0.0072618668278058365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1024,64,0.024940800666809083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,64,0.0033919999996821085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,64,0.007361066838105519
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,1024,32,0.024651734034220378
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,32,0.003570133447647095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,32,0.007387733459472657
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,768,65536,1.1045876820882161
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1024,3584,0.015017599860827128
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,65536,0.11484266916910808
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,65536,0.23016959826151528
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,768,16384,0.30067199071248374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,768,16384,0.04759786526362102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,16384,0.03511893351872762
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,16384,0.058303999900817874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,768,12288,0.23138346672058105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,768,12288,0.036798934141794845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,12288,0.028180267413457232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1024,3072,0.013680000106493631
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,12288,0.04472106695175171
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,768,10240,0.1877237319946289
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,10240,0.023835732539494833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,10240,0.03535679976145427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,768,8192,0.144488525390625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1024,2560,0.012266666690508524
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,8192,0.019010132551193236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,8192,0.02924373348553975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,768,7168,0.1437023957570394
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1024,2048,0.011271466811498005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,7168,0.017373865842819212
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,7168,0.026270933945973712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,768,6144,0.11762133439381917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1024,512,0.007074133555094402
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,6144,0.015278933445612588
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,6144,0.023589332898457847
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,768,5120,0.10010986328125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,768,5120,0.017805866400400796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,5120,0.013124266266822815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,5120,0.0207914670308431
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1024,1536,0.009633066256841023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,768,4096,0.08386773268381754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,768,4096,0.015321600437164306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,4096,0.011101866761843365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,4096,0.017732266585032144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,768,3584,0.07743679682413737
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,3584,0.009875200192133586
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,3584,0.01658560037612915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,768,3072,0.07083520094553629
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1024,256,0.006236800054709116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,3072,0.008963200449943542
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,3072,0.015165866414705912
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,768,2560,0.06080213387807211
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,768,2560,0.011767466862996418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,2560,0.011563733220100403
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,2560,0.013683199882507324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,768,2048,0.052451201279958094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,768,2048,0.010597333312034607
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,2048,0.00981333355108897
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,2048,0.012087466319402058
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,768,1536,0.0449504017829895
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,768,1536,0.009352533022562663
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,1536,0.008181333541870117
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,1536,0.010796800255775452
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,768,1024,0.03555946747461955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,768,65536,0.1756981372833252
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,1024,0.006494933366775512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,1024,0.009417600433031718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,768,768,0.03212479948997497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,768,10240,0.030649600426355998
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,768,768,0.007624533275763195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,768,0.005560533205668131
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,768,0.008782933155695598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,768,512,0.029040000836054486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,768,512,0.006816000243028005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,512,0.004691199958324432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,512,0.008203733464082081
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,768,256,0.02632746696472168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,768,7168,0.02262293299039205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,768,8192,0.025370667378107708
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,256,0.003939199944337209
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,256,0.007635200023651123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,768,128,0.025154133637746174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,128,0.003623466690381368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,128,0.007289599875609081
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,768,64,0.02493013342221578
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,64,0.003268266717592875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,64,0.007416533430417378
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,768,32,0.02443199952443441
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,32,0.0033834666013717652
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,32,0.007388799885908763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,512,65536,1.104151407877604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,65536,0.08156159718831381
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,65536,0.23109653790791831
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,512,16384,0.3017429351806641
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,768,6144,0.02038080096244812
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,16384,0.02813439965248108
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,16384,0.05820906559626261
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,512,12288,0.23124586741129555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,512,12288,0.033624533812204996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,12288,0.021897600094477335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,12288,0.0420576016108195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,512,10240,0.18672320048014324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,768,3072,0.013050666451454163
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,10240,0.019961599508921304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,768,3584,0.014299733440081277
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,10240,0.035019731521606444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,512,8192,0.023411200443903605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,512,8192,0.1439850648244222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,8192,0.01611733337243398
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,8192,0.029063467184702558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,512,7168,0.14461439450581867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,7168,0.014647466937700906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,7168,0.026266666253407796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,512,6144,0.11885546843210856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,512,6144,0.01892906626065572
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,6144,0.01309333344300588
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,6144,0.023483733336130776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,512,5120,0.10022613207499187
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,512,5120,0.016587733229001363
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,5120,0.011680000027020772
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,5120,0.020777599016825358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,512,4096,0.08363200028737386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,768,256,0.006083199878533682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,512,4096,0.014495999614397685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,768,1024,0.008236800134181977
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,4096,0.00981760025024414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,4096,0.01774186690648397
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,512,3584,0.07729173501332601
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,3584,0.009205333391825358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,3584,0.016436266899108886
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,512,3072,0.0705183982849121
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,512,3072,0.012220799922943115
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,3072,0.008463999629020691
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,3072,0.015107199549674988
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,512,2560,0.06046613454818726
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,2560,0.0074986666440963745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,2560,0.01353600025177002
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,512,2048,0.052535466353098546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,512,2048,0.010035199920336406
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,2048,0.006514133512973785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,2048,0.012001066406567892
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,512,1536,0.045414400100708005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,768,128,0.0058005332946777345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,1536,0.0058037335673968
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,1536,0.010703999797503154
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,512,1024,0.035558398564656576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,512,65536,0.16188586552937825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,1024,0.00487360010544459
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,1024,0.009309867024421692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,512,768,0.03218026757240296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,512,16384,0.04397759834925334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,768,0.004554666578769684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,768,0.00876800020535787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,512,512,0.029088000456492107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,512,512,0.006558933357397716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,512,0.004130133241415024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,512,0.008089600006739299
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,512,256,0.026272000869115193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,512,10240,0.02808319926261902
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,256,0.003634133438269297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,256,0.007464533547560374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,512,128,0.02511253356933594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,512,3584,0.013241599996884665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,512,7168,0.020925867557525634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,128,0.0033632000287373864
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,128,0.007143466671307881
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,512,64,0.024498132864634196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,64,0.0032085334261258446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,64,0.007276799778143566
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,512,32,0.024301866690317787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,32,0.0032970666885375976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,32,0.007212799787521362
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,256,65536,1.103118896484375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,256,65536,0.13189120292663575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,512,2560,0.011219200491905213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,65536,0.058842666943868006
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,65536,0.23173653284708656
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,256,16384,0.3015872001647949
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,16384,0.016577066977818807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,16384,0.055569068590799964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,512,1536,0.00879039963086446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,256,12288,0.027360000212987262
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,12288,0.015925332903862
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,256,12288,0.23189226786295572
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,12288,0.040167466799418135
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,256,10240,0.023528534173965453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,256,10240,0.18731733957926433
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,10240,0.014122666915257773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,10240,0.03468480110168457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,256,8192,0.14424853324890136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,8192,0.01239466667175293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,512,768,0.0070933332045873005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,8192,0.028742400805155437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,256,7168,0.14325440724690755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,7168,0.011526399850845337
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,512,1024,0.0076682666937510175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,7168,0.025987199942270917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,256,6144,0.11868906815846761
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,6144,0.011794132987658183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,6144,0.023226666450500488
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,256,5120,0.09941973686218261
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,5120,0.010404266913731893
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,5120,0.020516266425450645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,256,4096,0.08326506614685059
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,512,256,0.005962666869163513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,4096,0.00888320008913676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,4096,0.01775253415107727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,256,3584,0.07714560031890869
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,256,3584,0.011358933647473653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,3584,0.00810346653064092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,3584,0.016446933150291443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,256,3072,0.07060373624165853
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,256,3072,0.010517332951227825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,3072,0.007472000022729237
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,3072,0.015110400319099427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,256,2560,0.06044160127639771
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,512,128,0.005690666536490122
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,2560,0.007065600156784058
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,2560,0.013525333007176718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,256,2048,0.0520362655321757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,256,2048,0.009173333644866943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,2048,0.006514133512973785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,2048,0.012116266290346782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,256,1536,0.04380586544672648
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,256,8192,0.01983039975166321
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,1536,0.005490133166313171
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,256,7168,0.01768746574719747
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,1536,0.01053013304869334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,256,1024,0.03516053358713786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,1024,0.004647466540336609
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,1024,0.00937600036462148
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,256,16384,0.0367082675298055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,256,768,0.03200533390045166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,768,0.004271999994913737
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,768,0.008716799815495809
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,256,512,0.02882560094197591
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,256,6144,0.016105600198109946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,512,0.003944533318281174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,256,5120,0.014132266243298849
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,512,0.008187733093897502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,256,256,0.026428800821304322
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,256,0.0035914666950702667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,256,0.007523199915885926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,256,128,0.025245867172876996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,256,128,0.00547733356555303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,128,0.0032778667906920114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,128,0.007131733496983846
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,256,64,0.024817067384719848
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,64,0.0031210665901501974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,64,0.0072170664866765336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,256,32,0.024296534061431885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,32,0.003159466634194056
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,32,0.007253333429495494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,128,65536,1.1041685740152993
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,128,65536,0.12962026596069337
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,65536,0.048509867986043294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,256,4096,0.012435199817021687
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,65536,0.23176639874776206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,128,16384,0.3021162668863932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,128,16384,0.03577920198440552
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,16384,0.01138879954814911
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,16384,0.05289066632588705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,128,12288,0.23157760302225747
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,12288,0.012165333827336628
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,12288,0.040037333965301514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,128,10240,0.18760639826456707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,256,2560,0.009741866588592529
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,10240,0.011150933305422465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,256,1536,0.007865599791208903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,10240,0.03435626824696859
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,128,8192,0.019375999768575035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,128,8192,0.1442954699198405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,8192,0.009852799773216247
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,8192,0.028717867533365887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,128,7168,0.14383145968119304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,7168,0.00950933297475179
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,7168,0.026035199562708538
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,128,6144,0.11759573618570965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,256,1024,0.007019733389218648
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,6144,0.008931199709574383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,6144,0.023031467199325563
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,128,5120,0.09887893199920654
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,256,512,0.006212266782919565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,5120,0.008842666943868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,5120,0.020551466941833497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,128,4096,0.08262399832407633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,256,768,0.006630399823188781
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,4096,0.008188800017038981
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,4096,0.017658666769663493
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,128,3584,0.07705279986063639
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,128,3584,0.011527466773986816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,3584,0.00792746643225352
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,3584,0.016338133811950685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,128,3072,0.07019840081532797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,256,256,0.0057205334305763245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,128,3072,0.010363733768463135
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,3072,0.007176533341407776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,3072,0.0149536003669103
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,128,2560,0.0602399984995524
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,2560,0.006540800134340922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,2560,0.013553067048390707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,128,2048,0.05209706624348959
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,128,12288,0.026580266157786053
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,2048,0.005794133245944977
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,2048,0.011913599570592244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,128,1536,0.0435477336247762
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,128,1536,0.007577600081761678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,1536,0.005073066552480062
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,1536,0.010565333565076192
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,128,1024,0.035234133402506515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,128,1024,0.006839466591676076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,1024,0.004401066899299621
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,1024,0.009321600198745728
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,128,768,0.031915734211603805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,128,10240,0.022932267189025878
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,768,0.004102399945259095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,768,0.008700799942016602
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,128,512,0.02877013285954793
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,128,5120,0.013774933417638144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,512,0.0036778666079044344
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,512,0.008075733482837678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,128,256,0.026102399826049803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,128,4096,0.012245333194732666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,256,0.0033557333052158357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,128,7168,0.01742080052693685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,256,0.007492266595363617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,128,6144,0.01598186691602071
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,128,128,0.024859732389450072
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,128,128,0.005392000079154968
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,128,0.0031776001056035364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,128,0.007106133302052816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,128,64,0.02472213308016459
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,64,0.0030261332790056865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,64,0.0072202667593956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,128,32,0.024133332570393882
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,32,0.003068800022204717
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,32,0.007162666817506154
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,65536,0.23215360641479493
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,65536,0.048469332853953044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,16384,0.013193600376447043
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,16384,0.05320746501286825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,64,16384,0.3006218592325846
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,12288,0.0107722669839859
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,64,65536,1.1050389607747397
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,64,12288,0.23223679860432944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,12288,0.03981333176294963
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,10240,0.009971200426419576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,10240,0.03433599869410197
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,64,10240,0.18714133898417157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,8192,0.009062400460243225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,64,8192,0.14435307184855145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,8192,0.02877119978268941
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,7168,0.008615466952323913
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,7168,0.02601066629091899
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,64,7168,0.1439370632171631
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,6144,0.007945600152015685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,64,6144,0.11808853149414063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,6144,0.02325973312060038
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,5120,0.007592533528804779
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,64,5120,0.09892373085021973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,5120,0.020447999238967896
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,64,4096,0.08334506352742513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,4096,0.007083733379840851
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,4096,0.01776960094769796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,3584,0.007549866537253062
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,64,3584,0.07703039646148682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,3584,0.016434133052825928
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,64,3072,0.06962239742279053
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,3072,0.007243733108043671
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,3072,0.01492800017197927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,64,2560,0.06012800137201944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,2560,0.006552533308664958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,2560,0.013556266824404398
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,64,2048,0.05160746574401855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,2048,0.00574186642964681
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,2048,0.011916800340016683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,128,2560,0.009594666957855224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,64,1536,0.04292693138122559
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,1536,0.005006933212280273
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,1536,0.010595200459162395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,128,2048,0.008797867099444072
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,64,1024,0.035345065593719485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,1024,0.004308266441027323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,1024,0.009278933207194012
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,64,768,0.03197120030721028
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,768,0.003987200061480204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,768,0.00867306689421336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,64,512,0.028802132606506346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,512,0.003734400123357773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,512,0.008037333190441132
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,64,256,0.026009599367777508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,256,0.0033471999069054925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,256,0.007459199925263722
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,64,128,0.0250218669573466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,128,0.0030400000512599947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,128,0.007055999835332234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,64,0.0070613334576288865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,64,64,0.024652800957361855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,64,0.002926933268706004
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,64,32,0.024155733982721965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,32,0.0029557332396507262
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,32,0.00717439999183019
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,65536,0.04527466694513957
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,65536,0.23190827369689943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,16384,0.010039466619491576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,32,16384,0.3020394643147787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,16384,0.05200533469518026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,12288,0.008954667051633199
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,12288,0.03982079823811849
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,32,12288,0.23202667236328126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,10240,0.008278400202592214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,10240,0.03439679940541585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,8192,0.0076896001895268755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,32,10240,0.18753706614176432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,32,65536,1.1045045216878255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,8192,0.028841600815455122
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,32,8192,0.14446293512980143
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,7168,0.007235200206438701
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,32,7168,0.1427178700764974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,7168,0.026104533672332765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,32,6144,0.11944426695505779
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,6144,0.00698773314555486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,6144,0.02320746580759684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,32,5120,0.09962879816691081
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,5120,0.007435733576615651
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,5120,0.020578134059906005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,4096,0.006941866874694824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,32,4096,0.08274133205413818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,4096,0.017517866690953572
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,32,3584,0.07698132991790771
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,3584,0.00739519993464152
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,3584,0.01634986698627472
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,32,3072,0.06962239742279053
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,3072,0.0070250665148099255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,3072,0.015010133385658264
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,32,2560,0.05989120006561279
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,32,2048,0.05124799807866415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,2048,0.011935999989509583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,2560,0.006378666559855144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,1536,0.004993066688378652
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,2560,0.013493333260218301
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,2048,0.005613866448402405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,32,1536,0.04272533257802327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,1536,0.01055999994277954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,32,1024,0.03520426750183105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,1024,0.004260266820589701
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,1024,0.009243733684221904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,32,768,0.03189546664555867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,768,0.004001066585381826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,768,0.008730666836102803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,32,512,0.028914133707682293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,512,0.0036159999668598174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,512,0.00802346666653951
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,32,256,0.026174932718276978
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,256,0.0032469332218170166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,128,768,0.006438399851322174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,256,0.0075093333919843035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,32,128,0.02506133317947388
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,128,0.0030389333764712016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,128,0.0071274667978286745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,32,64,0.024660267432530723
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,32,0.007147733370463054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,64,0.002930133293072383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,64,0.007097599903742473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,768,32,32,0.024277333418528238
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,32,0.0028778667251269023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,128,512,0.006071466704209646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,65536,16384,0.7059872309366862
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,65536,12288,0.731173324584961
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,65536,16384,0.9545525232950846
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,65536,12288,1.1009323120117187
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,65536,16384,1.5759508768717447
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,128,256,0.00553599993387858
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,65536,12288,0.5729344050089519
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,65536,10240,0.626748784383138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,65536,10240,0.5096533457438152
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,65536,10240,0.9224853515625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,65536,8192,0.46755412419637044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,65536,8192,0.3933781305948893
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,65536,8192,0.38186025619506836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,65536,8192,0.7241098403930664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,65536,7168,0.34649813969930016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,65536,7168,0.3233386675516764
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,65536,7168,0.4310890515645345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,65536,7168,0.6447669347127278
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,65536,6144,0.29946559270222983
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,65536,6144,0.3766773223876953
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,65536,6144,0.2737290700276693
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,65536,6144,0.5618687947591146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,65536,5120,0.2559157371520996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,65536,5120,0.31506134668986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,65536,4096,0.20843839645385742
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,65536,5120,0.24188159306844076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,65536,5120,0.45694828033447266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,65536,4096,0.25402773221333824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,65536,4096,0.37548373540242513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,65536,4096,0.1910144011179606
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,65536,3584,0.183570130666097
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,65536,3584,0.22398079236348473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,65536,3584,0.1752768039703369
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,65536,3584,0.3310271898905436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,65536,3072,0.19702933629353842
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,65536,3072,0.1512608051300049
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,65536,3072,0.2802506764729818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,65536,2560,0.16925652821858722
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,65536,2560,0.13641494115193684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,65536,2560,0.23997759819030762
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,65536,2560,0.12508373260498046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,65536,2048,0.11354026794433594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,65536,2048,0.14176534016927084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,65536,2048,0.1958207925160726
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,65536,2048,0.10622719923655193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,65536,1536,0.11698133150736492
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,65536,1536,0.15026133855183918
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,65536,1536,0.0856661319732666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,65536,1024,0.09498240152994791
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,65536,1024,0.10383893648783367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,65536,1024,0.06748053232828775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,65536,768,0.05772906541824341
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,65536,768,0.0815605322519938
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,65536,768,0.08312640190124512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,65536,768,0.056082133452097574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,65536,512,0.07520319620768229
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,65536,512,0.061749335130055746
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,65536,512,0.049781334400177
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,65536,256,0.06553920110066733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,65536,256,0.03845973412195842
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,65536,256,0.044188801447550455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,65536,256,0.042659199237823485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,65536,16384,0.767520014444987
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,65536,128,0.06214613517125448
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,65536,128,0.03631999890009562
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,65536,128,0.039869864781697587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,65536,12288,0.5955904006958008
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,65536,64,0.06294826666514078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,65536,64,0.035869868596394856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,65536,64,0.04121599992116292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,65536,32,0.06733866532643637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,65536,32,0.03206186691919963
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,65536,32,0.04361813465754191
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,65536,10240,0.48531198501586914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,16384,65536,1.4447755177815755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,65536,1.4342676798502603
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,16384,16384,0.3723050753275553
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,65536,0.7943381627400716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,16384,0.3662218729654948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,65536,3072,0.15860586166381835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,16384,0.2015765349070231
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,16384,12288,0.16457173029581706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,16384,12288,0.26398506164550783
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,12288,0.26998186111450195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,12288,0.14381333986918132
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,65536,1536,0.09078613122304281
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,65536,1024,0.06698666413625082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,16384,10240,0.2252554734547933
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,10240,0.2285055955251058
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,10240,0.12161386807759603
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,16384,8192,0.17375680605570476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,8192,0.18380799293518066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,8192,0.10181012948354085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,65536,512,0.05196373462677002
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,16384,7168,0.1727519989013672
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,16384,7168,0.09457813103993734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,7168,0.15875412623087565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,7168,0.08609600067138672
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,16384,6144,0.1543765385945638
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,16384,6144,0.08164693514506022
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,6144,0.13944427172342938
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,6144,0.07820586363474527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,16384,5120,0.07066240310668945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,16384,5120,0.13103893597920735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,5120,0.11585493087768554
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,65536,128,0.0399946649869283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,5120,0.06479253371556601
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,16384,4096,0.10931200186411541
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,4096,0.0929248015085856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,4096,0.053787732124328615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,16384,3584,0.05240960121154785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,16384,3584,0.09721813201904297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,3584,0.08306986490885417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,3584,0.04798080126444499
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,16384,3072,0.0456000010172526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,16384,3072,0.08909227053324381
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,3072,0.07121173540751138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,3072,0.04225493272145589
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,16384,2560,0.07864747047424317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,16384,2560,0.04213546514511109
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,2560,0.06045120159784952
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,2560,0.037435734272003175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,16384,2048,0.06460373401641846
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,16384,2048,0.03464213212331136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,2048,0.04967893362045288
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,2048,0.03212160070737203
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,16384,1536,0.05567786693572998
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,16384,65536,0.8021856307983398
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,16384,1536,0.028990934292475384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,1536,0.038856534163157146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,1536,0.028972800572713214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,16384,1024,0.046470399697621664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,16384,1024,0.021064533789952596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,1024,0.030671999851862593
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,1024,0.019980800151824952
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,16384,768,0.041523198286692306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,768,0.022402133544286093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,768,0.01687893271446228
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,16384,512,0.03736319939295451
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,16384,512,0.014778666694959006
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,512,0.014737066626548768
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,512,0.01704960068066915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,16384,256,0.03282240033149719
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,16384,256,0.011891200145085653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,256,0.01183893382549286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,256,0.012745599945386252
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,16384,128,0.030876799424489336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,16384,128,0.011149866382280986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,128,0.009382399916648864
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,128,0.011813333630561829
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,64,0.008065066734949748
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,16384,64,0.03054400086402893
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,64,0.012597333391507468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,16384,32,0.030661332607269286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,32,0.008400000135103862
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,32,0.013550933202107748
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,16384,16384,0.2017279942830404
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,12288,65536,0.5998058954874674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,65536,0.5814773559570312
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,65536,1.184424591064453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,12288,16384,0.3042218526204427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,12288,65536,1.2000202178955077
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,12288,16384,0.15661120414733887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,16384,0.15048960049947102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,16384,0.3214922587076823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,16384,10240,0.12839573224385578
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,12288,0.21373653411865234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,12288,12288,0.23765120506286622
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,12288,0.12078719933827717
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,12288,10240,0.1929642677307129
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,10240,0.09610773722330729
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,10240,0.17586879730224608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,12288,8192,0.0812661329905192
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,12288,8192,0.15123626391092937
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,16384,8192,0.10622933705647786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,8192,0.1412021319071452
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,8192,0.07852586905161539
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,12288,7168,0.15186239878336588
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,7168,0.12427840232849122
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,7168,0.06997973124186198
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,12288,6144,0.13439146677652997
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,6144,0.10784426530202229
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,6144,0.059273600578308105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,12288,5120,0.11775039831797282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,5120,0.09072213172912598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,5120,0.05064213275909424
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,12288,4096,0.09682880242665609
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,12288,4096,0.04439359903335571
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,16384,4096,0.05815466642379761
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,4096,0.0730773369471232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,4096,0.04184853235880534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,12288,3584,0.08899306456247966
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,3584,0.06447466611862182
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,3584,0.038252798716227214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,12288,3072,0.08024960358937581
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,12288,3072,0.03570346832275391
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,3072,0.05563946564992269
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,3072,0.0343285322189331
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,12288,2560,0.07172266642252603
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,12288,2560,0.03199466665585836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,2560,0.04728960196177165
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,16384,768,0.017796266078948974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,2560,0.030458666880925495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,12288,2048,0.06021120150883993
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,2048,0.03878399928410848
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,2048,0.02456106742223104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,12288,1536,0.04982613325119019
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,12288,1536,0.02159893314043681
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,1536,0.030078933636347456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,1536,0.01914773384730021
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,12288,1024,0.040990932782491045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,12288,1024,0.01676693360010783
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,1024,0.0211136003335317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,1024,0.015528532862663268
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,12288,12288,0.11749546527862549
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,12288,768,0.03636480172475179
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,12288,10240,0.10031039714813232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,768,0.016706132888793947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,768,0.013986133535703025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,12288,512,0.035607465108235675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,12288,512,0.01279253363609314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,512,0.012988799810409546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,12288,7168,0.07274560133616129
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,512,0.012315733234087627
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,12288,256,0.029934932788213093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,256,0.009266133109728496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,256,0.010870400071144103
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,12288,128,0.02855573296546936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,12288,128,0.009669333696365356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,12288,6144,0.06238293250401815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,128,0.007467733323574066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,128,0.010082133611043294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,12288,5120,0.05321066776911417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,64,0.0065098668138186145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,12288,64,0.029102933406829835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,64,0.01071679989496867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,32,0.006886399785677592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,12288,32,0.028505599498748778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,32,0.011389866471290588
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,10240,65536,0.5192170778910319
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,12288,3584,0.04068373441696167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,65536,0.564684804280599
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,65536,0.9332842508951822
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,10240,16384,0.13488213221232098
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,10240,16384,0.30222078959147136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,16384,0.22613439559936524
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,16384,0.16991359392801922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,10240,65536,1.1754325866699218
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,10240,12288,0.10403947035471599
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,10240,12288,0.22545280456542968
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,12288,0.1191541353861491
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,12288,0.1714431921641032
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,10240,0.10046613216400146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,10240,0.1449685255686442
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,10240,10240,0.19183252652486166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,10240,8192,0.14882879257202147
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,8192,0.1174400011698405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,8192,0.07639146645863851
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,10240,7168,0.14988373120625814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,7168,0.10241493384043376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,7168,0.06509973208109537
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,10240,6144,0.13096319834391276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,6144,0.0885205348332723
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,6144,0.05817493200302124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,10240,5120,0.11410346825917561
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,12288,2048,0.027405865987141925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,10240,5120,0.04776639938354492
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,5120,0.07626346747080484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,5120,0.04952746629714966
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,10240,4096,0.09444373448689779
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,10240,4096,0.03988373279571533
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,4096,0.061427199840545656
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,4096,0.041324798266092935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,10240,3584,0.08653972943623861
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,10240,3584,0.036638931433359785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,3584,0.054722134272257486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,3584,0.03756800095240275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,10240,3072,0.0785802682240804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,10240,3072,0.03203199903170268
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,3072,0.047492265701293945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,3072,0.033352533976236984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,10240,2560,0.07026346524556479
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,10240,2560,0.027229867378870648
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,2560,0.04056426684061686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,2560,0.02876586715380351
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,10240,2048,0.058532265822092686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,10240,2048,0.023670399188995363
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,2048,0.03341653347015381
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,2048,0.022447999318440756
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,10240,1536,0.04941973288853963
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,10240,1536,0.019193599621454872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,1536,0.027790933847427368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,1536,0.018613332509994508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,10240,1024,0.039961600303649904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,10240,1024,0.015050666530927024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,1024,0.01856213410695394
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,1024,0.015530666708946228
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,10240,768,0.036252800623575845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,10240,768,0.01384213368097941
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,768,0.015165866414705912
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,768,0.013899733622868856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,10240,512,0.032942932844161985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,12288,768,0.014778666694959006
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,512,0.011589333415031433
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,512,0.012330666184425354
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,10240,256,0.02979733347892761
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,12288,256,0.010679466525713603
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,256,0.008642133076985676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,256,0.010578133662541707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,10240,128,0.028612265984217327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,128,0.00714026689529419
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,128,0.009955199559529622
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,10240,64,0.02911253372828166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,64,0.006316799918810527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,64,0.01074026624361674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,10240,32,0.02844480077425639
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,32,0.006607999900976817
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,32,0.010914132992426554
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,10240,10240,0.0884991963704427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,10240,8192,0.0707594633102417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,10240,7168,0.06415146589279175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,8192,65536,0.412385082244873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,8192,65536,1.0265045166015625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,10240,6144,0.055683199564615884
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,65536,0.6996629079182942
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,65536,0.4022656122843425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,8192,16384,0.11398719946543376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,8192,16384,0.2637258688608805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,16384,0.17766399383544923
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,16384,0.11841066678365071
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,8192,12288,0.08359359900156657
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,8192,12288,0.19467946688334148
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,12288,0.07919039726257324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,12288,0.13712746302286785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,8192,10240,0.07083840370178222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,8192,10240,0.1790239969889323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,10240,0.11558187007904053
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,10240,0.06632959842681885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,8192,8192,0.05783040126164755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,8192,8192,0.12781333128611247
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,8192,0.09195840358734131
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,8192,0.05458346605300903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,8192,7168,0.051994665463765466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,8192,7168,0.13092373212178549
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,7168,0.08178880214691162
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,7168,0.050152532259623205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,8192,6144,0.04575039943059285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,8192,6144,0.11667413711547851
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,6144,0.06932693322499593
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,6144,0.04279786745707194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,8192,5120,0.10448640187581379
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,5120,0.03662079970041911
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,8192,5120,0.03951893250147502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,5120,0.05941546758015951
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,8192,4096,0.08532373110453287
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,4096,0.047378134727478025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,4096,0.03136746684710185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,8192,3584,0.0784447987874349
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,8192,3584,0.0307861328125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,3584,0.04403626521428426
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,3584,0.027247999111811323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,8192,3072,0.0717034657796224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,8192,3072,0.025996800263722735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,3072,0.036685868104298906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,3072,0.0233407994111379
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,8192,2560,0.06370026667912801
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,8192,2560,0.02224000096321106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,2560,0.031378134091695147
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,8192,2048,0.053364264965057376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,2560,0.019992534319559732
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,8192,2048,0.019045333067576088
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,2048,0.02558293342590332
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,8192,1536,0.04503893454869588
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,2048,0.01731200019518534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,10240,512,0.01181226670742035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,1536,0.0197269340356191
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,1536,0.014434132973353067
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,8192,1024,0.03677013317743937
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,1024,0.01463573376337687
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,1024,0.012124799688657125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,8192,768,0.03362666765848796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,8192,768,0.012238933642705282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,768,0.012292266885439555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,768,0.011158399780591329
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,8192,512,0.030289065837860108
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,8192,512,0.010843732953071594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,512,0.009920000036557516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,512,0.00988159974416097
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,10240,256,0.009956266482671101
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,256,0.007392000158627827
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,256,0.009035733342170716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,10240,128,0.009013332923253377
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,8192,256,0.02823253273963928
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,8192,128,0.0269375999768575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,128,0.006382933259010315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,128,0.008517332871754964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,8192,64,0.02661120096842448
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,64,0.005612800021966299
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,64,0.009050666292508443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,8192,32,0.026705066363016765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,32,0.005719466507434845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,32,0.009406933188438415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,7168,65536,0.410753059387207
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,7168,65536,0.9782410939534506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,65536,0.6583818435668946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,65536,0.39251626332600914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,7168,16384,0.27636054356892903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,7168,16384,0.10673600037892658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,16384,0.17051092783610028
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,16384,0.0994271993637085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,7168,12288,0.08178239663441976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,7168,12288,0.19499732653299967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,12288,0.1261514663696289
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,12288,0.07866986592610678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,7168,10240,0.1565066655476888
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,10240,0.1058773358662923
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,10240,0.06473280191421509
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,7168,8192,0.1259584029515584
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,8192,0.08657493591308593
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,8192,0.05227200190226237
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,8192,4096,0.033190399408340454
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,7168,7168,0.13004586696624756
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,7168,7168,0.050855465730031336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,7168,0.07621119817097982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,7168,0.04743146498998006
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,7168,6144,0.11412800153096517
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,6144,0.06594026486078898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,6144,0.04091093142827352
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,7168,5120,0.0997109333674113
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,8192,1536,0.016473600268363954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,8192,1024,0.013684266805648803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,5120,0.055538133780161536
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,5120,0.03609493176142375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,7168,4096,0.08341866334279378
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,7168,4096,0.03359359900156657
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,4096,0.04331306616465251
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,4096,0.03030719955762227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,7168,3584,0.03073173364003499
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,7168,3584,0.07732693354288736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,3584,0.026258132855097455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,3584,0.03879893223444621
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,7168,3072,0.07009066740671793
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,8192,256,0.009448533256848652
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,3072,0.03334720134735107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,3072,0.02247040073076884
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,7168,2560,0.06209493478139242
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,7168,2560,0.022873600323994957
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,8192,128,0.008710400263468424
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,2560,0.027888000011444092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,2560,0.019428267081578573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,7168,2048,0.05158186753590902
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,2048,0.022370133797327676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,2048,0.01694186727205912
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,7168,1536,0.04296533266703288
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,1536,0.018266665935516357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,1536,0.01441493332386017
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,7168,1024,0.03661226828893026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,7168,10240,0.06932480335235595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,1024,0.012433066964149475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,1024,0.012119467059771221
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,7168,768,0.03349226713180542
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,7168,768,0.012808533509572348
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,768,0.010201600193977357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,768,0.011171199878056844
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,7168,8192,0.057070934772491456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,7168,512,0.03102506597836812
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,7168,512,0.011475200454394024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,512,0.00804373323917389
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,512,0.010010666648546855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,7168,256,0.028296534220377607
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,256,0.006281599899133046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,256,0.009000533819198608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,7168,6144,0.044622933864593504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,7168,128,0.027189334233601887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,7168,128,0.00902400016784668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,128,0.0053727999329566956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,128,0.008505599697430928
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,7168,64,0.027246934175491334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,64,0.005067733426888784
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,64,0.0090037335952123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,7168,32,0.02672959963480632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,32,0.005312000215053558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,32,0.009474133451779682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,7168,5120,0.03909333149592082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,65536,0.5639167785644531
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,6144,65536,0.9774826685587564
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,65536,0.4177205403645833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,6144,16384,0.2568266709645589
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,16384,0.1395136038462321
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,7168,3072,0.026050132513046265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,16384,0.10085439682006836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,6144,12288,0.06378133296966552
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,6144,12288,0.18637973467508953
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,7168,2048,0.01952106753985087
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,12288,0.1055519978205363
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,12288,0.0762506643931071
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,6144,10240,0.05414186716079712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,6144,10240,0.15423466364542643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,10240,0.09179413318634033
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,10240,0.06467839876810709
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,7168,1536,0.01690773367881775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,6144,8192,0.045830400784810384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,8192,0.07177813053131103
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,7168,1024,0.014298666516939798
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,6144,8192,0.12454933325449627
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,8192,0.052515200773874914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,6144,7168,0.040566400686899824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,6144,7168,0.12927786509195965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,7168,0.06376640001932779
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,7168,0.04667093356450398
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,6144,6144,0.11338666280110676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,6144,6144,0.035571201642354326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,6144,0.0548746665318807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,6144,0.040829865137736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,6144,5120,0.030789333581924438
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,6144,5120,0.09971626599629721
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,5120,0.04620373249053955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,5120,0.034991999467213944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,6144,4096,0.08269866307576498
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,4096,0.03754026492436727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,4096,0.027559467156728107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,6144,3584,0.07638506889343262
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,6144,3584,0.023162666956583658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,3584,0.03377279837926229
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,3584,0.025092266003290814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,6144,3072,0.06859947045644124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,3072,0.02872320016225179
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,3072,0.02248959938685099
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,6144,2560,0.061578667163848876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,6144,2560,0.01781760056813558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,2560,0.02455893357594808
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,2560,0.019334399700164796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,6144,2048,0.015424000223477683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,6144,2048,0.05161600112915039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,2048,0.020011732975641884
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,2048,0.01707520087560018
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,6144,1536,0.013673599561055502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,6144,1536,0.0420693318049113
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,1536,0.015516799688339234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,1536,0.014016000429789224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,6144,1024,0.036644268035888675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,6144,1024,0.011346133550008138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,1024,0.01167039970556895
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,1024,0.012056533495585125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,6144,768,0.0334826668103536
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,768,0.009909333785374959
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,768,0.011058132847150166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,6144,512,0.031042132774988813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,6144,512,0.009622400005658466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,512,0.008089600006739299
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,512,0.009946667154630025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,6144,256,0.02805440028508504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,6144,256,0.008452266454696655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,7168,256,0.00990826686223348
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,256,0.006341333190600078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,256,0.009037866195042928
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,6144,128,0.026959999402364092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,128,0.005339733262856802
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,128,0.008475733796755473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,6144,64,0.02725333372751872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,64,0.004823466638724009
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,64,0.00899733304977417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,6144,32,0.0265610675017039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,32,0.005048533280690512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,32,0.009337600072224934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,5120,65536,0.2738410631815592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,6144,65536,0.31189438501993816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,65536,0.464189879099528
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,5120,65536,0.9736127853393555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,65536,0.3851978619893392
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,5120,16384,0.07784533500671387
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,5120,16384,0.2565653324127197
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,16384,0.11753599643707276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,16384,0.10032426516215007
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,5120,12288,0.1877962589263916
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,12288,0.08995520273844401
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,12288,0.07650132973988852
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,5120,10240,0.154530127843221
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,10240,0.07568533420562744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,10240,0.06350933313369751
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,6144,16384,0.0840768019358317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,5120,8192,0.12393706639607746
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,8192,0.06089813311894735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,8192,0.05062506596247355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,5120,7168,0.12859093348185222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,5120,7168,0.03688640197118123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,7168,0.05366933345794678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,7168,0.04660373528798421
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,5120,6144,0.11285866896311443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,5120,6144,0.032313599189122515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,6144,0.046476801236470536
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,6144,0.040292267004648843
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,5120,5120,0.09894506931304932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,5120,5120,0.02728853424390157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,5120,0.039315199851989745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,5120,0.03392533461252849
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,6144,4096,0.02576853235562642
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,5120,4096,0.08171306451161703
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,4096,0.03213546673456828
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,4096,0.027292799949645997
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,5120,3584,0.075436798731486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,6144,3072,0.020668800671895346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,5120,3584,0.02102186679840088
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,3584,0.029448533058166505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,3584,0.024501333634058632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,5120,3072,0.06800106366475424
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,5120,3072,0.018323200941085815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,3072,0.02585066755612691
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,3072,0.021941334009170532
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,5120,2560,0.060331734021504727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,6144,768,0.010538666447003683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,2560,0.02147093415260315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,2560,0.019375999768575035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,5120,2048,0.05039360125859579
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,6144,128,0.007657599945863088
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,2048,0.01744106610616048
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,2048,0.01684373418490092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,5120,1536,0.04175146818161011
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,5120,1536,0.012422399719556172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,1536,0.014284800489743552
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,1536,0.014333867033322654
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,5120,1024,0.010619733730951946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,5120,1024,0.03619306484858195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,1024,0.011211733023325603
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,1024,0.012171733379364013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,5120,768,0.033070933818817136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,768,0.009620267152786254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,768,0.01112320025761922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,5120,512,0.03054933349291484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,5120,512,0.00899733304977417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,512,0.00806826651096344
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,512,0.009890133142471313
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,5120,256,0.027962666749954224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,5120,12288,0.05709760189056397
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,5120,256,0.007941333452860515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,256,0.006495999793211619
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,256,0.00901759962240855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,5120,128,0.02680213252703349
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,5120,128,0.007336533566315968
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,128,0.005605333546797434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,128,0.008507733543713886
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,5120,64,0.02669973373413086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,64,0.004695466657479604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,5120,10240,0.04921386639277141
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,64,0.008660266796747845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,5120,32,0.026436267296473186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,32,0.004906666775544485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,32,0.00902400016784668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,5120,8192,0.04069653352101644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,65536,0.3648746808369955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,65536,0.22615253130594887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,4096,65536,0.8266730626424155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,4096,16384,0.22350613276163736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,16384,0.05959359804789225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,16384,0.09453759988149008
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,4096,12288,0.04960426489512126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,4096,12288,0.15791786511739095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,12288,0.07118079662322999
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,12288,0.04626453320185343
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,4096,10240,0.04241173267364502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,4096,10240,0.13067200183868408
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,10240,0.05874346494674683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,10240,0.03954879840215047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,4096,8192,0.03496640125910441
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,4096,8192,0.10590826670328776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,8192,0.04788159926732381
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,8192,0.03323733409245809
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,4096,7168,0.11210239728291829
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,7168,0.042115199565887454
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,7168,0.028596266110738115
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,5120,4096,0.0228000005086263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,4096,6144,0.09947840372721353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,6144,0.036517333984375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,6144,0.02482453385988871
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,4096,5120,0.0880074659983317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,4096,5120,0.023283199469248454
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,5120,0.030908799171447753
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,5120,0.02118933399518331
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,4096,4096,0.07288320064544677
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,4096,4096,0.019447465737660728
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,4096,0.025337600708007814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,4096,0.01777600049972534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,4096,3584,0.0679146687189738
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,4096,3584,0.01776640017827352
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,3584,0.02239146629969279
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,3584,0.01632213294506073
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,4096,3072,0.0616159995396932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,4096,3072,0.016116266449292503
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,3072,0.019088000059127808
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,3072,0.015081600348154704
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,5120,2048,0.014046933253606161
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,4096,2560,0.054662398497263585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,2560,0.01602453291416168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,2560,0.013653332988421122
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,4096,2048,0.04663679997126262
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,4096,2048,0.01269546647866567
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,2048,0.013688533504803976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,2048,0.012319999933242797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,4096,1536,0.03950399955113729
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,4096,1536,0.011348266402880352
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,1536,0.011173333724339802
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,5120,2560,0.016215466459592185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,1536,0.010789333780606588
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,4096,1024,0.033506135145823165
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,5120,768,0.009778133034706116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,1024,0.008665600419044494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,4096,65536,0.23620160420735678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,1024,0.00943893293539683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,4096,768,0.03177066644032796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,768,0.0077344000339508055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,768,0.008919466535250347
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,4096,16384,0.06355520089467367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,4096,512,0.02927359938621521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,4096,512,0.008363733688990276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,512,0.006279466549555461
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,512,0.00836906631787618
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,4096,256,0.026798933744430542
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,256,0.005224533379077911
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,256,0.007683200140794118
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,4096,128,0.025544534126917522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,4096,128,0.006839466591676076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,128,0.004524800181388855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,128,0.007166933516661327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,64,0.004091733445723852
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,4096,64,0.025257599353790284
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,64,0.007461333274841308
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,4096,7168,0.03140693306922913
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,32,0.004305066665013632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,4096,32,0.024925865729649863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,32,0.007620266576608021
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,65536,0.316488520304362
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,4096,6144,0.02736426591873169
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,65536,0.22186452547709146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3584,65536,0.8058058420817057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3584,16384,0.06594453255335489
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,16384,0.08327893416086832
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3584,16384,0.213155206044515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,16384,0.05676266749699911
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,12288,0.0616159995396932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3584,12288,0.1547935962677002
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,12288,0.04445759852727254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3584,10240,0.12797760168711345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,10240,0.052230401833852136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,10240,0.03833386500676473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3584,8192,0.10348800023396809
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3584,8192,0.035689600308736164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,8192,0.042071465651194254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,8192,0.031754666566848756
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3584,7168,0.11055573622385662
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3584,7168,0.032407466570536295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,7168,0.03721706469853719
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,7168,0.027868799368540448
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3584,6144,0.09787093003590902
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,4096,2560,0.014432000120480857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3584,6144,0.029219200213750202
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,6144,0.03269866704940796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,6144,0.023781333367029825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3584,5120,0.08646079699198404
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3584,5120,0.02482453385988871
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,5120,0.029833600918451948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,5120,0.020906666914621987
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3584,4096,0.07199040253957113
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3584,4096,0.02076479991277059
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,4096,0.021875200668970744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,4096,0.017568000157674155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3584,3584,0.0667306661605835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3584,3584,0.019385600090026857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,3584,0.019382399320602418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,3584,0.0164682666460673
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3584,3072,0.060670932133992515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3584,3072,0.01741973360379537
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,3072,0.017990400393803917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,3072,0.014789332946141561
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,4096,1024,0.009773866335550944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3584,2560,0.05432960192362467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3584,2560,0.015262933572133383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,2560,0.015466666221618653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,2560,0.013329066832860312
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3584,2048,0.04526079893112182
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,2048,0.013291733463605246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,2048,0.01209493378798167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3584,1536,0.038429868221282956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3584,1536,0.011892267068227132
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,1536,0.010244266192118327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,1536,0.010737066467603047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3584,1024,0.03343679904937744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3584,1024,0.010437333583831787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,1024,0.007981866598129272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,1024,0.009411199887593587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3584,768,0.03095146616299947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,768,0.0070271998643875126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3584,768,0.009361066420873006
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,768,0.008797867099444072
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3584,512,0.028823467095692952
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,512,0.005886933207511902
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,512,0.008151466647783916
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3584,256,0.026525866985321046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3584,256,0.007533866663773854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,256,0.004819199939568837
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,256,0.007539199789365132
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3584,128,0.025638399521509807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,4096,768,0.009040000041325887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,4096,256,0.007368533313274384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,128,0.004357333481311798
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,128,0.007079466680685679
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3584,64,0.025091199080149333
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,64,0.003905066599448522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,64,0.007310933371384938
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3584,32,0.024628265698750814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,32,0.004112000018358231
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,32,0.007477333148320516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3584,65536,0.252237860361735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3072,65536,0.22365226745605468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3072,65536,0.8031295776367188
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,65536,0.21727466583251953
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,65536,0.2862058639526367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3584,10240,0.04329066673914592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3584,12288,0.050475732485453284
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3072,16384,0.06139839887619018
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,16384,0.07165226936340333
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,16384,0.05521066586176554
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3072,12288,0.04801173210144043
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3072,16384,0.21268800099690757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3072,12288,0.15314240455627443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,12288,0.05473706722259521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,12288,0.04383680025736491
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3072,10240,0.12732160091400146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,10240,0.03743786811828613
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,10240,0.04693333307902019
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3072,8192,0.034687999884287515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3072,8192,0.10293653011322021
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,8192,0.03683946530024211
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,8192,0.030771199862162275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3072,7168,0.10995413462320965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,7168,0.032762666543324784
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,7168,0.026238934199015303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3072,6144,0.09689173698425294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3072,6144,0.02675093412399292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,6144,0.029487999280293782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,6144,0.022940800587336222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3072,5120,0.08470079898834229
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,5120,0.028221867481867474
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,5120,0.020281600952148437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3072,4096,0.07110506693522135
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3072,4096,0.01872106591860453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,4096,0.018761599063873292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,4096,0.01738986571629842
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3072,3584,0.06606719891230264
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3072,3584,0.01755733291308085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,3584,0.017498666048049928
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,3584,0.015989333391189575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3072,3072,0.06006079912185669
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3072,3072,0.015754666924476624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,3072,0.017245866854985557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,3072,0.014553599556287131
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3072,2560,0.05373973449071249
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3072,2560,0.014064000050226847
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,2560,0.014111999670664468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,2560,0.01325653294722239
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3072,2048,0.044385067621866864
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3072,2048,0.01237440009911855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,2048,0.011009066303571065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,2048,0.011961600184440613
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3072,1536,0.038551465670267744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3072,1536,0.011125333110491435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,1536,0.009709866841634114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,1536,0.010636799534161885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3072,1024,0.03378986517588298
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3072,1024,0.009546666344006857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,1024,0.007314133147398631
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,1024,0.009293867150942485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3072,768,0.031198932727177935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3072,768,0.008770133058230083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,768,0.006523733337720235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,768,0.008769067128499348
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3072,512,0.028870399792989093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3072,512,0.007796266674995422
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,512,0.00554666668176651
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,512,0.008122666676839193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3584,2048,0.013448533415794373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3072,256,0.02650559941927592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3072,256,0.006898133456707001
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,256,0.004614399870236715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,256,0.0075573335091273
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3072,128,0.02547840078671773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,128,0.0041002665956815084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,128,0.0070592001080513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3072,64,0.025054933627446492
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,64,0.003788800040880839
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,64,0.0073642666141192125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,3072,32,0.024664533138275147
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,32,0.003882666677236557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,32,0.007451733450094859
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2560,65536,0.20805120468139648
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3584,512,0.008518399794896443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,65536,0.23951360384623208
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,65536,0.20854934056599936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2560,65536,0.7982655843098958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2560,16384,0.05778666734695435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2560,16384,0.2104853312174479
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,16384,0.06116480032602946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,16384,0.05420053402582804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2560,12288,0.15071360270182294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2560,12288,0.0449834664662679
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,12288,0.05514026482899984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,12288,0.04261973301569621
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2560,10240,0.12634027004241943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3584,128,0.007092266778151194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,10240,0.05003626743952433
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,10240,0.0369045337041219
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2560,8192,0.10157866477966308
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2560,8192,0.03111039996147156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,8192,0.0324565331141154
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,8192,0.02885119915008545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2560,7168,0.10862720012664795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,7168,0.0343285322189331
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,7168,0.02638186613718669
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2560,6144,0.09606186548868814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2560,6144,0.024638932943344117
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,6144,0.030903466542561847
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,6144,0.02379306753476461
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2560,5120,0.08514560063680013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2560,5120,0.021027199427286782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,5120,0.02630293369293213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,5120,0.02040426731109619
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2560,4096,0.07049173514048258
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2560,4096,0.017492266496022542
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,4096,0.018756266434987387
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,4096,0.01727466583251953
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2560,3584,0.06595306793848674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2560,3584,0.01614400049050649
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,3584,0.01702293356259664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,3584,0.015945600469907124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2560,3072,0.05999146699905396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,3072,0.016394666830698647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,3072,0.014879999558130899
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3072,10240,0.04130986531575521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2560,2560,0.0536629319190979
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,2560,0.013486933708190919
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3072,7168,0.030821333328882854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,2560,0.013392000397046407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2560,2048,0.044719998041788736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2560,2048,0.011462400356928509
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,2048,0.011335466305414836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,2048,0.011982933680216471
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2560,1536,0.03823893467585246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2560,1536,0.010427733262379963
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,1536,0.009218133489290873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,1536,0.010692266623179118
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2560,1024,0.03339093526204427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2560,1024,0.00900266667207082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3072,5120,0.022486400604248048
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,1024,0.006885333359241486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,1024,0.009348266323407491
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2560,768,0.030777599414189654
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2560,768,0.008298666775226593
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,768,0.006247466802597046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,768,0.008732799688975017
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2560,512,0.028420267502466838
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2560,512,0.007528533538182576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,512,0.00806933343410492
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,512,0.0052159999807675685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2560,256,0.02648959954579671
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,256,0.004430933296680451
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,256,0.0076000000039736434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2560,128,0.025217066208521526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2560,128,0.006301866471767425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,128,0.003956266740957896
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,128,0.0070592001080513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2560,64,0.024856533606847128
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,64,0.003621333340803782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,64,0.007153066496054332
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2560,32,0.024850134054819742
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,32,0.0037471999724706015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,32,0.007477333148320516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2048,65536,0.1791594664255778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,65536,0.19016532897949218
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2048,65536,0.7969290415445964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3072,128,0.006417066852251689
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,65536,0.20743145942687988
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2048,16384,0.04633280038833618
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,16384,0.053880532582600914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2048,16384,0.21043732961018882
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,16384,0.05438719987869263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2048,12288,0.0370901346206665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,12288,0.04893759886423747
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2048,12288,0.1499392032623291
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,12288,0.04230719804763794
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2048,10240,0.03351253271102905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2048,10240,0.12548906803131105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,10240,0.043703468640645345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,10240,0.03551146586736043
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2048,8192,0.026180267333984375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2048,8192,0.10130133628845214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,8192,0.03175786733627319
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,8192,0.02769920031229655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2048,7168,0.02365760008494059
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2048,7168,0.10746453603108723
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,7168,0.029612799485524494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,7168,0.02542613347371419
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2048,6144,0.0949951966603597
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,6144,0.0272490660349528
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,6144,0.022699733575185142
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2048,5120,0.08440639972686767
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2048,5120,0.01839573383331299
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,5120,0.023396267493565878
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2560,10240,0.03841600020726522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,5120,0.020130133628845213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2048,4096,0.07033920288085938
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2048,4096,0.01573013365268707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,4096,0.01685653328895569
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,4096,0.01727573275566101
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2048,3584,0.01492693324883779
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2048,3584,0.06539200146993002
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,3584,0.015727999806404113
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,3584,0.015714133779207863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2560,7168,0.027083732684453327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2048,3072,0.05909973382949829
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,3072,0.014264532923698425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,3072,0.014536533753077188
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2048,2560,0.05272853374481201
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2048,2560,0.012307199835777282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,2560,0.012378666798273723
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,2560,0.013208533326784769
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2048,2048,0.04394133488337199
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2048,2048,0.0105813334385554
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,2048,0.010452266534169514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,2048,0.011850666999816895
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2048,1536,0.038126933574676516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2048,1536,0.00967680017153422
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,1536,0.008541867136955261
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,1536,0.010520533720652262
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2048,1024,0.0331989328066508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2048,1024,0.008427733182907104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,1024,0.006386133531729381
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,1024,0.00922986666361491
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2048,768,0.03051946759223938
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2048,768,0.00766293356815974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,768,0.005757866799831391
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,768,0.008664533495903015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2048,512,0.02807253400484721
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2048,512,0.007099733253320058
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,512,0.004890666902065277
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,512,0.008051200211048127
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2048,256,0.02605866591135661
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2048,256,0.006364800035953522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,256,0.004167466859022776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,256,0.0074879998962084455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2560,3072,0.014450132846832275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2048,128,0.025146667162577314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2048,128,0.005960533519585928
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,128,0.0036981334288915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,128,0.006951466699441274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2048,64,0.024889600276947022
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,64,0.0034048000971476236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,64,0.007085866729418437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,2048,32,0.024014933904012045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,32,0.003551999976237615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,32,0.007150933146476746
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2560,2560,0.013114666938781739
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,65536,0.14762667020161946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,65536,0.20438292821248374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1536,16384,0.21011199951171874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1536,16384,0.05048533280690511
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,16384,0.04286186695098877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,16384,0.05394239823023478
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1536,65536,0.7897855758666992
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1536,12288,0.1495914618174235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,12288,0.040828800201416014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,12288,0.04073280096054077
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1536,10240,0.12333119710286458
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,10240,0.037596801916758224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,10240,0.03391146659851074
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1536,8192,0.10035093625386556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2560,256,0.0066101332505544026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,8192,0.02775999903678894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,8192,0.027339732646942137
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1536,7168,0.10689493020375569
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,7168,0.025639466444651288
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,7168,0.02501759926478068
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1536,6144,0.09447147051493326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2048,6144,0.020889600118001304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,6144,0.022719999154408775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,6144,0.02225386699040731
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1536,5120,0.08410560290018718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2048,3072,0.013490133484204612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,5120,0.01903466582298279
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,5120,0.019693867365519205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1536,4096,0.07002773284912109
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1536,65536,0.17928214073181153
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,4096,0.01515733301639557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,4096,0.017016534010569254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1536,3584,0.06512426535288493
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1536,3584,0.013330133756001792
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1536,7168,0.021604265769322714
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,3584,0.013641599814097086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,3584,0.015659733613332113
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1536,8192,0.02491626739501953
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1536,3072,0.05858773390452067
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,3072,0.012661332885424295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,3072,0.01437440017859141
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1536,2560,0.05284800132115682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1536,2560,0.011004799604415893
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,2560,0.010760533809661865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,2560,0.01307093302408854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1536,2048,0.044809599717458085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1536,6144,0.01885653336842855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1536,12288,0.038387199242909745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1536,2048,0.01011306643486023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,2048,0.008658132950464885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,2048,0.01171946624914805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1536,1536,0.03803733189900716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1536,10240,0.03176640073458354
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1536,1536,0.008945066730181377
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,1536,0.0074976002176602675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,1536,0.01046506663163503
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1536,1024,0.03347626527150472
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1536,1024,0.0076906666159629825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,1024,0.00591893345117569
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,1024,0.009164800246556599
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1536,768,0.03063253362973531
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,768,0.005212800204753875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,768,0.008631466825803121
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1536,512,0.028548266490300494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,512,0.004514133433500925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,512,0.008252800007661184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1536,256,0.02612373431523641
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1536,256,0.0064074665307998655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,256,0.003947733342647553
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,256,0.007474133372306823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1536,128,0.02505386670430501
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,128,0.0035690667728583017
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,128,0.00698773314555486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1536,64,0.02468693256378174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,64,0.003260799994071325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,64,0.007092266778151194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1536,32,0.024422399202982583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,32,0.0033610666791598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,32,0.007142400244871776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1024,65536,0.13605333964029948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,65536,0.10180586973826092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1024,65536,0.7883189519246419
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1536,5120,0.0164682666460673
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,65536,0.20444159507751464
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1024,16384,0.20895679791768393
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,16384,0.032289065917332965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,16384,0.05353493293126425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1024,12288,0.15000212987263997
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,12288,0.030628265937169392
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,12288,0.03863146702448527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1024,10240,0.12350186506907146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1024,10240,0.024649600187937416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,10240,0.028483200073242187
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,10240,0.032536532481511435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1536,4096,0.013895466923713684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1024,8192,0.09937280019124349
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,8192,0.020662399133046468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,8192,0.02713279922803243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1024,7168,0.10602453549702961
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1024,7168,0.017784533898035686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,7168,0.017845332622528076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,7168,0.02463679909706116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1024,6144,0.09404266675313314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1024,6144,0.016664533813794454
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,6144,0.017190400759379068
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,6144,0.02207466761271159
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1024,5120,0.08329066435496012
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1536,3072,0.012403200070063274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,5120,0.014127999544143677
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,5120,0.01945706605911255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1024,4096,0.06947093009948731
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1536,512,0.006811733543872833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,4096,0.010502400000890096
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,4096,0.016871466239293417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1536,768,0.007227733234564463
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1024,3584,0.06420586506525675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,3584,0.009658666451772054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,3584,0.015527466932932535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1536,128,0.005949866771697998
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1024,3072,0.058074665069580075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1024,3072,0.011032533645629884
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,3072,0.009142399827639262
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,3072,0.014320000012715658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1024,2560,0.052140800158182773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1024,2560,0.009913600484530131
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,2560,0.007655466596285502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,2560,0.013064533472061157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1024,2048,0.04391359885533651
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1024,2048,0.008781866232554118
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,2048,0.006779733300209046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,2048,0.011730133493741354
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1024,1536,0.03784000078837077
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,1536,0.005985066791375478
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,1536,0.010338133573532105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1024,1024,0.03322346607844035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1024,1024,0.007114666700363159
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,1024,0.005026133358478546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,1024,0.009097599983215332
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1024,768,0.03062186638514201
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1024,16384,0.03636480172475179
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,768,0.004574933151404063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1024,12288,0.02855573296546936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,768,0.00851093331972758
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1024,512,0.027853866418202717
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1024,512,0.006237866481145223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,512,0.004065066576004028
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,512,0.007906133433183034
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1024,256,0.02595840096473694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1024,8192,0.01912533243497213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,256,0.003647999962170919
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,256,0.007274666428565979
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1024,128,0.024485333760579427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1024,5120,0.014578133821487427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,128,0.003432533393303553
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,128,0.006971733272075653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1024,64,0.024734934171040855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,64,0.003253333270549774
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,64,0.007083733379840851
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,1024,32,0.02426346739133199
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,32,0.0034293333689371743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,32,0.007034666836261749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,768,65536,0.12197013696034749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,65536,0.08061227003733316
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,65536,0.20427519480387368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,768,65536,0.7882240295410157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,768,16384,0.2094165325164795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,16384,0.02717653314272563
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,16384,0.05083839893341065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1024,4096,0.01279253363609314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,768,12288,0.026366933186848955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,768,12288,0.1505194664001465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,12288,0.024953599770863852
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,12288,0.037921067078908285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,768,10240,0.1242730696996053
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,10240,0.02216213345527649
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,10240,0.03252266645431519
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,768,8192,0.10012693405151367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1024,3584,0.012036266922950744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,8192,0.016059733430544534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,8192,0.027318400144577027
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,768,7168,0.10666666825612385
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,768,7168,0.01732053359349569
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1024,1536,0.008105599880218506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,7168,0.014730667074521383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,7168,0.024729599555333458
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,768,6144,0.09443199634552002
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,768,6144,0.01575040022532145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,6144,0.013102933764457703
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,6144,0.022259199619293214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,768,5120,0.08351573149363199
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,768,5120,0.013894400000572205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1024,768,0.006674133241176605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,5120,0.014346667130788169
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,5120,0.019662932554880778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,768,4096,0.06915626525878907
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,768,4096,0.012223999698956807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,4096,0.010375466942787171
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,4096,0.01684266726175944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,768,3584,0.06422079801559448
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,768,3584,0.011615999539693196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,3584,0.009306666254997254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,3584,0.015570132931073507
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,768,3072,0.058474667867024745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,3072,0.008526933193206788
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,3072,0.014325333635012307
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,768,2560,0.05084799925486246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,768,2560,0.00962666670481364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,2560,0.007937066753705342
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,2560,0.01302826702594757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,768,2048,0.0431818683942159
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,2048,0.007077333331108093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,2048,0.011633066336313884
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,768,1536,0.03771413167317708
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1024,256,0.005860266586144766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,1536,0.006226133306821187
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,1536,0.010356266299883525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,768,1024,0.033156265815099076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1024,128,0.00558186670144399
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,1024,0.0053269331653912864
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,1024,0.009087999661763508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,768,768,0.030408533414204915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,768,16384,0.03516266743342082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,768,0.004886400202910105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,768,0.008550399541854858
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,768,512,0.028099199136098225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,768,8192,0.018900267283121743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,512,0.0042367999752362575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,512,0.00790293316046397
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,768,10240,0.02247999906539917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,768,256,0.025967999299367266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,768,256,0.005804799993832906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,256,0.003621333340803782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,256,0.007302399973074596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,768,128,0.024825600783030193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,768,128,0.005576533575852713
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,128,0.0032874666154384612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,128,0.006960000097751618
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,768,64,0.02443199952443441
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,64,0.0031093334158261614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,64,0.007054933408896129
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,768,32,0.024201599756876628
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,32,0.003201066702604294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,32,0.007034666836261749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,512,65536,0.7877770741780599
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,512,65536,0.11059306462605793
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,65536,0.059799468517303465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,65536,0.20336960156758627
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,512,16384,0.20902934074401855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,16384,0.020467199881871543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,768,3072,0.010633599758148194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,16384,0.051261866092681886
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,512,12288,0.02401813268661499
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,512,12288,0.14950186411539715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,12288,0.019934932390848793
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,12288,0.03773546616236369
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,512,10240,0.12422293027242023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,768,2048,0.008654933174451191
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,10240,0.017603200674057008
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,10240,0.033037867148717245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,512,8192,0.017544533809026083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,512,8192,0.10079572995503742
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,8192,0.014103466272354126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,8192,0.027321600914001466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,512,7168,0.10634986559549968
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,512,7168,0.016064000129699708
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,7168,0.013313066959381104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,7168,0.024658133586247764
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,512,6144,0.014626133441925048
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,512,6144,0.09438933531443278
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,768,1536,0.007853866616884867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,6144,0.013556266824404398
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,6144,0.02220266660054525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,512,5120,0.012938666343688964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,512,5120,0.08343040148417155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,5120,0.011345066626866658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,5120,0.019576533635457357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,512,4096,0.011526399850845337
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,768,1024,0.00699839989344279
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,512,4096,0.06875306765238444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,4096,0.009820800026257832
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,4096,0.01697173317273458
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,512,3584,0.06408640146255493
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,3584,0.008920533458391826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,3584,0.015585066874821982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,512,3072,0.05765653451283773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,3072,0.008145066599051159
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,3072,0.014203733205795288
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,512,2560,0.05250666538874308
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,512,2560,0.009078400333722432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,2560,0.0073642666141192125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,2560,0.012930132945378623
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,512,2048,0.04277226527531942
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,768,768,0.006630399823188781
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,2048,0.006631466746330261
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,2048,0.011570133765538533
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,512,1536,0.03782399892807007
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,512,1536,0.0074080000321070354
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,1536,0.005850666761398315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,1536,0.01030399998029073
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,512,1024,0.03311040004094441
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,1024,0.004931200047334035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,1024,0.009123200178146362
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,512,768,0.030265599489212036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,512,768,0.006398933132489522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,768,0.004499199986457825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,768,0.00860053300857544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,512,512,0.028078933556874592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,512,512,0.0060917332768440245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,512,0.004108799993991852
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,512,0.00790719985961914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,512,256,0.025894399483998614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,512,256,0.005662933488686879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,256,0.0034805332620938623
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,256,0.007371733089288075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,512,128,0.02451733350753784
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,768,512,0.006252799928188324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,128,0.0031786667803923286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,128,0.006929066777229309
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,512,64,0.024445867538452147
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,64,0.0030229332546393077
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,64,0.007048533360163371
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,512,32,0.024001065889994302
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,32,0.003107200066248576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,32,0.0070250665148099255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,256,65536,0.7885653177897136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,512,16384,0.03128959933916728
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,65536,0.052130134900410977
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,65536,0.2054368019104004
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,256,16384,0.20966293017069498
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,512,10240,0.020488532384236653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,16384,0.013778133193651834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,16384,0.04809173345565796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,256,12288,0.15045973459879558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,512,3584,0.010860799749692281
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,12288,0.011464533209800721
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,512,3072,0.010098133484522502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,12288,0.037673600514729816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,256,10240,0.12400853633880615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,10240,0.011367467045783997
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,10240,0.03254826664924622
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,512,2048,0.008208000163237254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,256,8192,0.10064213275909424
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,256,8192,0.01655359963575999
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,8192,0.009547733267148336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,512,1024,0.006772266825040181
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,8192,0.02733760078748067
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,256,7168,0.01530346671740214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,256,7168,0.10644799868265789
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,7168,0.009758933385213216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,7168,0.024625066916147867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,256,6144,0.09417280356089273
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,256,6144,0.013809067010879517
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,512,128,0.005352533360322317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,6144,0.011229866743087768
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,6144,0.02205973267555237
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,256,5120,0.08324906826019288
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,5120,0.00844266712665558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,5120,0.01959679921468099
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,256,4096,0.06900479793548583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,4096,0.008577066659927367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,4096,0.01686186591784159
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,256,3584,0.06390186548233032
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,3584,0.007726933558781941
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,3584,0.015583999951680503
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,256,3072,0.057518935203552245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,256,65536,0.10972053209940594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,3072,0.007393066585063934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,3072,0.01421440045038859
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,256,2560,0.0507967988650004
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,256,16384,0.0294869323571523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,2560,0.006631466746330261
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,2560,0.012810666362444559
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,256,2048,0.042320001125335696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,256,12288,0.022410666942596434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,2048,0.005860266586144766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,2048,0.011689600348472596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,256,1536,0.03750400145848592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,256,1536,0.007186133166154225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,1536,0.0051584000388781226
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,1536,0.010273067156473796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,256,1024,0.03300906618436177
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,256,1024,0.006516266862551372
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,1024,0.00444160004456838
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,1024,0.009034666419029235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,256,768,0.03041173418362935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,256,10240,0.019542400042215982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,768,0.004215466479460398
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,768,0.008502399921417237
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,256,512,0.027584000428517656
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,256,4096,0.011027200023333232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,512,0.0037770666182041167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,512,0.007910400132338206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,256,256,0.02603200078010559
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,256,256,0.0054837331175804135
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,256,0.003437866767247518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,256,0.007346133391062419
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,256,128,0.02481279969215393
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,256,128,0.005309866865475973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,256,5120,0.012358400225639343
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,128,0.003139200061559677
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,128,0.006961066524187725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,256,64,0.02453226645787557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,64,0.0030165334542592366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,64,0.007039999961853028
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,256,32,0.024142932891845704
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,32,0.0030464000999927522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,32,0.007004799942175548
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,256,3584,0.010462933778762817
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,65536,0.042682667573293046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,128,65536,0.7893258412679036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,65536,0.20563200314839683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,128,16384,0.20923306147257487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,16384,0.010975999633471172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,16384,0.04848853349685669
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,256,3072,0.009621333082516987
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,128,12288,0.15015145937601726
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,128,12288,0.023142399390538533
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,12288,0.00976746678352356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,12288,0.03755413293838501
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,128,10240,0.12451413472493489
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,10240,0.00895146628220876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,10240,0.03269546627998352
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,128,8192,0.1005237340927124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,256,2048,0.007921066880226136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,8192,0.007998933394749958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,8192,0.027034666140874224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,256,2560,0.008644266923268636
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,128,7168,0.10628586610158283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,7168,0.007736533383528392
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,7168,0.024545067548751832
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,128,6144,0.09362026850382486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,256,768,0.006308266520500183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,6144,0.009300266702969868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,6144,0.022077866395314536
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,128,5120,0.08294933636983236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,256,512,0.006099199752012888
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,5120,0.008664533495903015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,5120,0.019614932934443156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,128,4096,0.06845226287841796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,128,65536,0.11011626720428466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,4096,0.007867733140786488
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,128,16384,0.029553065697352093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,4096,0.016819200913111367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,128,3584,0.06385813156763712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,128,8192,0.01718719998995463
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,128,3584,0.010546132922172546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,3584,0.007670400043328603
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,3584,0.015518933534622192
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,128,10240,0.020243199666341145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,128,3072,0.009827199578285217
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,128,3072,0.05732266505559286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,3072,0.007213866710662842
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,3072,0.014147200187047324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,128,2560,0.05145813226699829
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,2560,0.006470400094985962
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,2560,0.012974933783213297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,128,2048,0.04215893348058065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,128,2048,0.007970133423805236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,2048,0.005743999779224396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,2048,0.011627733707427979
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,128,1536,0.03753066857655843
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,128,1536,0.007331199944019318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,1536,0.005038933455944061
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,1536,0.010251733660697936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,128,1024,0.03293333252271016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,128,1024,0.006660266717274983
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,1024,0.004331733286380768
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,1024,0.009055999914805095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,128,768,0.030190932750701904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,128,768,0.006298666695753734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,768,0.004004266609748204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,768,0.008453333377838134
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,128,512,0.02770773371060689
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,128,512,0.005939200023810069
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,512,0.0037077332536379496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,512,0.007905066510041555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,128,256,0.025612799326578777
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,128,256,0.00551146666208903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,256,0.003389866650104523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,256,0.007303466896216075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,128,128,0.024715733528137208
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,128,0.0031776001056035364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,128,0.006870399912198384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,128,64,0.02444266676902771
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,64,0.0030464000999927522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,64,0.007012266914049785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,128,32,0.023938133319218954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,32,0.0033600000043710076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,32,0.007017600039641063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,65536,0.03618026574452718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,128,6144,0.014225066701571146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,65536,0.2058186690012614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,16384,0.009864532947540283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,64,16384,0.20848426818847657
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,64,65536,0.788420295715332
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,16384,0.04823360045750936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,12288,0.008667733271916707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,64,12288,0.15001707077026366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,12288,0.03787413438161214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,10240,0.008102400104204814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,64,10240,0.124236798286438
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,10240,0.032526934146881105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,128,7168,0.0156960000594457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,8192,0.007537066439787547
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,64,8192,0.10029866695404052
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,8192,0.027177600065867107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,7168,0.007347199817498524
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,64,7168,0.10632320245107015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,64,6144,0.09385173320770264
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,7168,0.024516266584396363
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,6144,0.00703359991312027
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,6144,0.021941334009170532
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,5120,0.007425066828727722
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,64,5120,0.0833237330118815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,5120,0.019539199272791543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,64,4096,0.06899519761403403
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,4096,0.006996266543865204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,64,3584,0.06396906773249308
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,4096,0.016755199432373045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,3584,0.007474133372306823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,3584,0.015383467078208923
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,3072,0.007010133564472198
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,64,3072,0.05617813269297282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,3072,0.014198399583498635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,64,2560,0.05068480173746744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,2560,0.006324266890684764
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,128,5120,0.012517333030700684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,2560,0.012929067015647888
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,64,2048,0.04223039944966634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,2048,0.005594666798909505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,2048,0.011560533444086711
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,64,1536,0.03744213183720906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,1536,0.004981333514054617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,1536,0.010242133339246114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,64,1024,0.03274880051612854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,1024,0.004271999994913737
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,1024,0.008993066350618998
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,64,768,0.030163200696309407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,768,0.0039594667653242745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,768,0.008361599842707316
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,64,512,0.027752532561620073
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,512,0.003601066768169403
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,512,0.007778133451938629
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,64,256,0.025675733884175617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,256,0.0032810665667057036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,256,0.007281066477298736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,64,128,0.024842667579650878
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,128,0.0030773334205150605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,128,0.0067775999506314594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,64,64,0.02421120007832845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,64,0.002919466545184453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,64,0.006899199883143107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,64,32,0.02380160093307495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,32,0.002915200094381968
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,32,0.00690773328145345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,65536,0.20516586303710938
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,65536,0.033019733428955075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,32,16384,0.2088010629018148
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,16384,0.008308266599973042
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,16384,0.04784533182779948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,12288,0.007536000013351441
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,32,12288,0.1495840072631836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,32,65536,0.7885749181111653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,12288,0.03744639952977498
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,10240,0.007400533556938172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,32,10240,0.12401599884033203
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,10240,0.03258453408877055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,8192,0.007293866574764251
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,32,8192,0.09974079926808675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,8192,0.027053866783777875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,7168,0.007145600020885467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,32,7168,0.10637013117472331
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,7168,0.02449493408203125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,32,6144,0.09371626377105713
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,6144,0.006916266679763794
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,6144,0.02201813260714213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,5120,0.007233066856861115
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,32,5120,0.08312959671020508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,5120,0.019516799847284952
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,4096,0.006821333368619282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,32,4096,0.06905386447906495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,4096,0.016796799500783284
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,32,3584,0.06394773324330648
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,3584,0.007309866448243459
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,3584,0.01548906664053599
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,32,3072,0.056790399551391604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,3072,0.007021866738796234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,3072,0.014152533809343972
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,32,2560,0.0496234655380249
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,2560,0.00626453310251236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,2560,0.012916266918182373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,32,2048,0.04223466714223226
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,2048,0.005633066594600678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,2048,0.011572266618410747
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,32,1536,0.037564798196156816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,1536,0.004936533172925314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,1536,0.010237866640090942
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,32,1024,0.032959999640782674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,1024,0.004278400043646494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,32,768,0.030510934193929036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,1024,0.009052800138791402
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,768,0.003944533318281174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,768,0.008499200145403545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,128,4096,0.011183999975522359
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,32,512,0.027574400107065838
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,512,0.003603200117746989
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,512,0.007758933305740357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,32,256,0.02564053336779277
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,256,0.0033183999359607695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,256,0.007146666447321574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,32,128,0.024651734034220378
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,128,0.0030346666773160298
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,64,0.006909866631031036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,128,0.006899199883143107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,32,64,0.02410879929860433
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,64,0.002914133419593175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,512,32,32,0.02373653252919515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,32,0.002887466549873352
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,32,0.006911999980608622
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,65536,16384,0.5978848139444988
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,65536,12288,0.43944212595621746
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,65536,16384,1.131646982828776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,65536,12288,0.9207967758178711
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,65536,12288,0.4547871907552083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,65536,16384,1.2465994517008463
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,65536,12288,0.8415840148925782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,65536,10240,0.3757354736328125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,128,2560,0.00867733359336853
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,65536,10240,0.4239498774210612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,65536,10240,0.7724543889363606
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,65536,10240,0.7011317571004232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,65536,8192,0.6184736251831054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,65536,8192,0.30270506540934244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,65536,8192,0.5852309544881185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,65536,7168,0.2664896011352539
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,65536,7168,0.5564938863118489
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,65536,7168,0.4928138732910156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,128,128,0.01313706636428833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,65536,6144,0.23234666188557945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,65536,6144,0.4837183952331543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,65536,6144,0.22446293830871583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,65536,6144,0.4327871958414714
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,65536,5120,0.19683094024658204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,65536,5120,0.4027008056640625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,65536,5120,0.20321812629699706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,65536,5120,0.37208213806152346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,65536,4096,0.16056960423787434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,65536,4096,0.154530127843221
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,65536,4096,0.3258101463317871
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,65536,4096,0.293173344930013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,65536,3584,0.14619626998901367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,65536,3584,0.14277226130167645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,65536,3584,0.255623467763265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,65536,3584,0.2865215937296549
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,65536,3072,0.24932799339294434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,65536,3072,0.22126506169637045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,65536,3072,0.12135573228200276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,65536,2560,0.11083199977874755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,65536,2560,0.20883307456970215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,65536,2560,0.10439679622650147
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,65536,2560,0.18842453956604005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,65536,2048,0.09135680198669434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,65536,2048,0.17144427299499512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,65536,2048,0.15359253883361818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,65536,2048,0.08575680255889892
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,65536,1536,0.1354848066965739
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,65536,1536,0.0735530694325765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,65536,1536,0.11917973359425862
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,65536,1536,0.07225600083669027
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,65536,1024,0.055283200740814206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,65536,1024,0.10261226495107015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,65536,1024,0.08567999998728434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,65536,1024,0.055214933554331457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,65536,768,0.04833386739095052
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,65536,768,0.07868586381276449
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,65536,768,0.04632960160573323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,65536,768,0.06782720088958741
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,65536,512,0.06898986498514811
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,65536,512,0.052160000801086424
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,65536,512,0.04046186606089274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,65536,256,0.03186560074488322
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,65536,256,0.06228906710942587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,65536,256,0.038038400808970134
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,65536,256,0.03520533243815104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,65536,128,0.06009920040766398
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,65536,128,0.031318400303522745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,65536,128,0.03245546619097392
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,65536,64,0.060395733515421546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,65536,16384,0.5873685201009115
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,65536,64,0.025700267155965167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,65536,64,0.03375466664632161
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,65536,32,0.02778773307800293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,65536,32,0.07106133302052817
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,65536,32,0.0359658678372701
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,65536,1.199280039469401
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,65536,8192,0.30323947270711266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,65536,7168,0.26517333984375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,16384,65536,1.5981834411621094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,16384,16384,0.39239679972330727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,65536,0.6382389068603516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,16384,0.28285439809163415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,16384,0.15453866322835286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,16384,12288,0.29679253896077473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,12288,0.2173386732737223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,12288,0.12037653128306072
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,16384,10240,0.2455733299255371
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,16384,10240,0.1015455961227417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,10240,0.10300373236338298
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,10240,0.1818442662556966
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,16384,8192,0.0822282632191976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,16384,8192,0.1976672013600667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,8192,0.14621334075927733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,65536,3072,0.12428906758626301
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,8192,0.08102933565775552
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,16384,7168,0.07381546497344971
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,7168,0.12651093006134034
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,16384,7168,0.19504000345865885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,7168,0.07252906958262126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,16384,6144,0.1705578645070394
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,6144,0.11049493153889973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,6144,0.06300586859385172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,16384,5120,0.055233065287272134
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,16384,5120,0.14410667419433593
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,5120,0.09192106723785401
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,5120,0.05445119937260946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,16384,4096,0.04504426717758179
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,16384,4096,0.11959466934204102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,4096,0.07521813710530599
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,4096,0.04473280111948649
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,16384,3584,0.04100480079650879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,16384,3584,0.10785919825236004
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,3584,0.06562453508377075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,3584,0.04062186479568482
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,16384,3072,0.0963594675064087
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,3072,0.056807466348012293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,3072,0.036354132493336994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,65536,512,0.04233280022939046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,16384,2560,0.08103679815928141
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,2560,0.04824533462524414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,65536,128,0.02927680015563965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,2560,0.032536532481511435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,16384,2048,0.07026027043660482
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,16384,65536,0.6079242706298829
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,2048,0.03996479908625285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,2048,0.02852800091107686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,16384,1536,0.05970986684163412
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,16384,16384,0.15332694053649903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,1536,0.030902399619420366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,1536,0.0230730672677358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,16384,1024,0.04811946551005046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,16384,1024,0.0159850666920344
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,1024,0.02215999960899353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,16384,12288,0.11771093209584553
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,1024,0.0164192001024882
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,16384,768,0.04132266839345296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,768,0.01746666630109151
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,768,0.014003200332323709
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,16384,512,0.037588266531626384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,16384,512,0.01183786690235138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,512,0.013538133104642233
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,512,0.012314666310946147
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,16384,256,0.03132373293240865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,16384,256,0.010025599598884582
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,256,0.009633066256841023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,256,0.0107914666334788
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,16384,128,0.030152533451716108
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,16384,128,0.008933333555857341
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,128,0.00765119989713033
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,128,0.010079999764760334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,16384,64,0.029612799485524494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,64,0.006734933455785115
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,64,0.010661333799362183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,16384,32,0.030357333024342854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,32,0.006970666845639547
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,32,0.012007466952006022
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,16384,6144,0.06476373275121053
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,12288,65536,0.5481119791666667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,12288,65536,1.3134613037109375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,65536,0.5764415740966797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,65536,0.831173324584961
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,16384,3072,0.036696533362070724
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,12288,16384,0.14376959800720215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,16384,0.15044266382853191
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,16384,0.23239466349283852
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,12288,16384,0.32269013722737633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,12288,12288,0.1081397294998169
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,12288,12288,0.24147307078043617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,12288,0.16071893374125162
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,16384,2560,0.032646399736404416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,12288,0.11780587037404378
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,12288,10240,0.09200426737467447
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,12288,10240,0.20330346425374352
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,10240,0.13194986979166667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,10240,0.0970367987950643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,12288,8192,0.0735530694325765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,12288,8192,0.1657845338185628
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,8192,0.1062666654586792
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,8192,0.08007253011067708
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,16384,2048,0.027850667635599773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,12288,7168,0.16497599283854167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,7168,0.09190719922383626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,7168,0.06826026439666748
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,12288,6144,0.05720106760660807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,12288,6144,0.14766079584757488
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,6144,0.08190399805704753
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,6144,0.060363733768463136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,12288,5120,0.12239360014597575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,5120,0.05184853474299113
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,5120,0.06786026954650878
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,12288,4096,0.10252799987792968
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,4096,0.05538133382797241
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,4096,0.044949332873026535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,12288,3584,0.09299733638763427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,3584,0.04949013392130534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,3584,0.0389141321182251
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,12288,3072,0.03278719981511434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,12288,3072,0.08407573699951172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,3072,0.04182720184326172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,3072,0.03458026647567749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,12288,2560,0.07166079680124918
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,2560,0.03606506586074829
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,2560,0.029705599943796797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,12288,2048,0.0622218648592631
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,16384,1536,0.021911466121673585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,2048,0.02977386713027954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,2048,0.02353066603342692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,12288,1536,0.053033598264058436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,16384,768,0.013892267147699991
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,1536,0.023933867613474526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,1536,0.01885226567586263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,12288,1024,0.04202666680018107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,12288,1024,0.015717333555221556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,1024,0.016463999946912132
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,1024,0.015438933173815408
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,12288,768,0.037463466326395675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,12288,7168,0.06501013437906901
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,768,0.013622400164604188
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,768,0.013852799932161966
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,12288,512,0.034281599521636966
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,12288,512,0.0116565336783727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,512,0.010595200459162395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,12288,4096,0.04048853317896525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,12288,5120,0.049158398310343424
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,512,0.012332800030708312
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,12288,256,0.02999573349952698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,256,0.00783253312110901
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,12288,3584,0.036866132418314615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,256,0.010779733459154766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,12288,128,0.02841493288675944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,12288,128,0.008725333213806152
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,128,0.006528000036875407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,128,0.00993173321088155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,12288,64,0.028116265932718914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,64,0.005763199925422668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,64,0.010378666718800863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,12288,32,0.02813120086987813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,32,0.006118399898211161
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,32,0.011307733257611592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,12288,2560,0.02927359938621521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,12288,2048,0.023991467555363973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,65536,0.7168629328409831
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,10240,65536,0.8952042897542318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,65536,0.41109546025594074
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,12288,1536,0.019672532876332603
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,10240,16384,0.21760212580362953
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,16384,0.10296533107757569
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,10240,16384,0.30992533365885416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,16384,0.18060266176859538
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,10240,65536,1.2234293619791665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,10240,12288,0.16830933888753255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,10240,12288,0.22086079915364584
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,12288,0.13661120732625326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,12288,0.08580906391143799
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,10240,10240,0.19038186073303223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,10240,0.11609919865926106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,10240,0.06864853700002035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,8192,0.09224106470743815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,12288,768,0.01384213368097941
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,10240,8192,0.15459945996602376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,8192,0.056113068262736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,10240,7168,0.1047818660736084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,10240,7168,0.15358826319376628
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,7168,0.05035093228022257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,7168,0.0802453358968099
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,10240,6144,0.09004159768422446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,10240,6144,0.1391775925954183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,6144,0.044609065850575766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,6144,0.06977492968241374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,10240,5120,0.07492907047271728
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,5120,0.05980586608250936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,10240,5120,0.11569279829661053
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,5120,0.03828373352686564
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,10240,4096,0.05864853461583456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,10240,4096,0.09577919642130533
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,4096,0.05114453236262003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,4096,0.03210879961649577
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,10240,3584,0.088155730565389
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,3584,0.042745598157246906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,3584,0.02947733402252197
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,10240,3072,0.07981332937876383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,10240,3072,0.04512213468551636
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,3072,0.03770026763280233
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,3072,0.026038400332132977
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,10240,2560,0.06845653057098389
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,2560,0.03215893308321635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,2560,0.021938133239746093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,10240,2048,0.05976106723149618
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,10240,2048,0.028973867495854694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,2048,0.026946133375167845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,2048,0.017199999094009398
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,10240,1536,0.04973866542180379
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,10240,1536,0.020755199591318767
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,1536,0.021346133947372437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,1536,0.01446613371372223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,10240,1024,0.041859201590220135
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,1024,0.014443733294804893
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,1024,0.012329600254694621
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,10240,768,0.0369322657585144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,10240,768,0.013994666934013366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,768,0.01241600016752879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,768,0.011154133081436157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,10240,512,0.033524266878763836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,512,0.009751466910044353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,512,0.010066133737564088
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,10240,256,0.030148265759150188
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,10240,256,0.009452799956003826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,256,0.00737066666285197
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,256,0.009041066964467366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,10240,128,0.028833067417144774
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,10240,128,0.008389332890510559
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,128,0.0063178668419520065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,128,0.008619733651479085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,10240,64,0.02881493369738261
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,64,0.005619200070699056
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,64,0.00886079967021942
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,12288,256,0.009544533491134644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,10240,32,0.028758400678634645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,32,0.005916800101598104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,32,0.00985599954922994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,8192,65536,0.3151391983032227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,65536,0.5643423716227214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,8192,65536,1.0349077224731444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,65536,0.386025587717692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,8192,16384,0.25480106671651204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,10240,10240,0.14267733891805012
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,10240,8192,0.11744426886240641
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,16384,0.146725336710612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,16384,0.10328959623972575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,8192,12288,0.0644159992535909
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,8192,12288,0.18728747367858886
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,12288,0.10916799704233807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,12288,0.07726613680521646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,10240,0.09329919815063477
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,8192,10240,0.15853759447733562
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,10240,0.06725760300954184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,8192,8192,0.1290005366007487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,10240,3584,0.05236053466796875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,8192,0.07303466796875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,8192,0.055012265841166176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,8192,7168,0.040839465459187825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,8192,7168,0.1339893341064453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,7168,0.06370240052541097
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,7168,0.04771840174992879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,10240,2560,0.03801066478093465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,8192,6144,0.11886080106099446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,8192,6144,0.036287999153137206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,6144,0.05547306537628174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,6144,0.042269865671793624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,10240,1024,0.01627626617749532
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,8192,5120,0.09989120165506998
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,5120,0.04625920057296753
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,5120,0.037115732828776046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,8192,4096,0.08541653156280518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,4096,0.03766080141067505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,4096,0.030529065926869707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,8192,3584,0.07824107011159262
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,8192,3584,0.021734400590260824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,3584,0.03358826637268066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,3584,0.026348799467086792
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,8192,3072,0.018951465686162315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,8192,3072,0.07164159615834555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,3072,0.02965013384819031
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,10240,512,0.011797333757082622
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,3072,0.022462934255599976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,8192,2560,0.016806399822235106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,8192,2560,0.06128533283869425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,2560,0.02528853416442871
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,2560,0.019230933984120686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,8192,2048,0.054017066955566406
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,2048,0.0204586664835612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,2048,0.0170197327931722
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,8192,1536,0.0460970679918925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,8192,1536,0.012842667102813721
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,1536,0.0157642662525177
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,1536,0.014521599809328715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,8192,1024,0.037759999434153244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,8192,1024,0.010814932982126872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,1024,0.01181226670742035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,1024,0.012315733234087627
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,8192,768,0.034135464827219644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,8192,768,0.009699199597040813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,768,0.010057600339253743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,768,0.011077333490053813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,8192,512,0.030748800436655684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,8192,512,0.008731733759244282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,512,0.008195200065771738
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,512,0.010106666882832845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,8192,256,0.027828266223271687
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,8192,256,0.007657599945863088
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,256,0.006532266736030579
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,256,0.008969600001970928
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,8192,128,0.026804266373316447
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,8192,128,0.007050666709740956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,128,0.005492266515890757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,128,0.00840106705824534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,8192,64,0.026652799050013228
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,64,0.004896000027656555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,64,0.00870293378829956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,8192,32,0.026679466168085735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,32,0.005027199784914652
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,32,0.009304533402125042
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,7168,65536,0.5347071965535481
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,7168,65536,0.9779200236002603
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,65536,0.4892704010009766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,65536,0.37826560338338217
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,7168,16384,0.2588895956675211
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,8192,16384,0.08435093561808268
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,16384,0.125272536277771
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,7168,16384,0.1292202631632487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,16384,0.10129706859588623
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,7168,12288,0.09724266529083252
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,7168,12288,0.1806538740793864
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,12288,0.09705706437428793
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,12288,0.07691840330759683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,7168,10240,0.15075947443644205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,10240,0.08036479949951172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,10240,0.06754026412963868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,7168,8192,0.12390613555908203
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,7168,8192,0.06081493298212687
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,8192,0.06519039869308471
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,8192,0.053963732719421384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,8192,10240,0.0547274669011434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,7168,7168,0.12933013439178467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,7168,0.05731840133666992
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,8192,8192,0.04617706537246704
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,7168,0.048803198337554934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,7168,6144,0.11557439963022868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,6144,0.05007359981536865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,6144,0.04279893239339193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,7168,5120,0.09693333307902018
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,8192,5120,0.031548800071080525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,8192,4096,0.025389866034189863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,5120,0.036082132657368975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,5120,0.041900801658630374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,7168,4096,0.08157973289489746
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,4096,0.03431146542231242
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,4096,0.028407466411590577
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,7168,3584,0.07565333048502604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,7168,3584,0.027247999111811323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,3584,0.030832000573476154
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,3584,0.02518186569213867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,8192,2048,0.014963199694951376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,7168,3072,0.023788799842198692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,7168,3072,0.06896853446960449
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,3072,0.027154133717219038
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,3072,0.022217599550882976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,7168,2560,0.05977813402811686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,2560,0.023564799626668295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,2560,0.019099734226862588
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,7168,2048,0.052375467618306484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,7168,2048,0.01806186636288961
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,2048,0.018936532735824584
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,2048,0.01652479966481527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,7168,1536,0.04384746551513672
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,7168,1536,0.015460266669591268
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,1536,0.01499626636505127
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,1536,0.014155733585357665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,7168,1024,0.03741333484649658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,7168,1024,0.013147733608881631
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,1024,0.01153706709543864
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,1024,0.011859200398127238
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,7168,768,0.03363519906997681
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,768,0.009917866190274556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,768,0.01116480032602946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,7168,512,0.03111039996147156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,7168,512,0.010019200046857198
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,512,0.008311466872692108
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,512,0.010075733065605164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,7168,256,0.02840533256530762
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,7168,256,0.008539733290672303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,256,0.006853333115577698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,256,0.00902186632156372
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,7168,128,0.027037866910298664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,7168,128,0.0076885332663853955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,128,0.005880533158779145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,128,0.008412800232569377
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,7168,64,0.026613332827885944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,64,0.0054175997773806255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,64,0.00881173312664032
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,7168,32,0.02653119961420695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,32,0.00558186670144399
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,32,0.009136000275611877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,7168,10240,0.08108800252278646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,6144,65536,0.9258613586425781
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,65536,0.41773440043131516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,7168,7168,0.054898134867350255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,65536,0.36807680130004883
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,6144,16384,0.2285205364227295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,7168,6144,0.0483242670694987
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,6144,16384,0.07577280203501383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,16384,0.10912000338236491
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,16384,0.09839573701222738
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,6144,12288,0.16826133728027343
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,12288,0.08278079827626547
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,12288,0.07757759888966878
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,6144,10240,0.04941866795221965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,6144,10240,0.1444810708363851
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,7168,5120,0.03947306474049886
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,10240,0.06917866865793863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,10240,0.06311893463134766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,6144,8192,0.040278398990631105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,6144,8192,0.11894079844156902
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,8192,0.05575573444366455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,8192,0.052162134647369386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,6144,7168,0.03627946774164836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,6144,7168,0.12478400071461995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,7168,4096,0.03153600096702576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,7168,0.048809599876403806
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,7168,0.04676266511281331
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,6144,6144,0.032400000095367434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,6144,6144,0.11114453474680583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,6144,0.042225066820780435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,6144,0.04113813241322835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,6144,5120,0.028490666548411054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,6144,5120,0.09314346313476562
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,5120,0.03685653209686279
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,5120,0.03346133232116699
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,6144,4096,0.079585067431132
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,4096,0.029167999823888142
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,7168,2560,0.02096959948539734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,6144,3584,0.07417919635772705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,4096,0.026814933617909747
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,6144,3584,0.020396800835927327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,3584,0.026054400205612182
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,3584,0.0241376002629598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,6144,3072,0.06747946739196778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,3072,0.02318613330523173
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,3072,0.021897600094477335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,6144,2560,0.057966931660970056
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,6144,2560,0.016562133034070333
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,2560,0.02073600093523661
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,2560,0.01936960021654765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,7168,768,0.011283199985822041
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,6144,2048,0.049981868267059325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,6144,2048,0.014538666605949402
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,2048,0.017960532506306966
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,2048,0.0165173331896464
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,6144,1536,0.0425983985265096
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,6144,1536,0.012541866302490235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,1536,0.014839466412862143
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,1536,0.014354133605957031
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,6144,1024,0.037342933813730876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,6144,1024,0.010678399602572124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,1024,0.010710400342941285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,1024,0.011860266327857971
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,6144,768,0.03308373292287191
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,6144,768,0.00965226689974467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,768,0.00855573316415151
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,768,0.01070186694463094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,6144,512,0.030610134204228718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,512,0.007099733253320058
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,512,0.010081066687901815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,6144,256,0.027790933847427368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,6144,256,0.00755626658598582
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,256,0.00561599979797999
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,256,0.008932266632715862
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,6144,128,0.02659733295440674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,6144,128,0.006982400019963582
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,128,0.004827733337879181
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,128,0.008295466502507527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,6144,64,0.0262880007425944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,64,0.0043381333351135256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,64,0.008578133583068848
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,6144,32,0.026502400636672974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,32,0.004484266539414724
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,32,0.008905599514643352
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,5120,65536,0.48278506596883136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,65536,0.22351466814676918
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,5120,65536,0.9152277628580728
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,65536,0.3671797434488932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,5120,16384,0.2232202688852946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,16384,0.09274453322092692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,16384,0.0584885319073995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,5120,12288,0.16567999521891277
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,5120,12288,0.08775466283162435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,6144,65536,0.2905749320983887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,12288,0.0698848009109497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,12288,0.04593706528345744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,5120,10240,0.07488533655802408
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,5120,10240,0.14557545979817707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,10240,0.06058880090713501
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,10240,0.040022401014963786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,5120,8192,0.11910933653513592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,8192,0.04765760103861491
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,8192,0.034030934174855546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,5120,7168,0.05190399885177612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,5120,7168,0.1209994633992513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,7168,0.042786133289337155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,7168,0.03035413424173991
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,5120,6144,0.043511466185251875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,5120,6144,0.10745706558227539
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,6144,12288,0.05796586672465006
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,6144,0.03679680029551188
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,6144,0.026494934161504106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,5120,5120,0.037752532958984376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,5120,0.031905066967010495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,5120,5120,0.09167359670003256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,5120,0.02101226647694906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,5120,4096,0.07843093077341715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,4096,0.026051199436187743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,4096,0.017262933651606242
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,5120,3584,0.07291839917500814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,3584,0.02296533385912577
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,3584,0.01616853376229604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,5120,3072,0.06672853628794352
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,5120,3072,0.021754666169484457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,3072,0.01964799960454305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,3072,0.014854400356610616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,5120,2560,0.057411201794942225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,6144,4096,0.023277866840362548
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,5120,2560,0.018732800086339315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,2560,0.01680213411649068
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,2560,0.013871999581654868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,5120,2048,0.04861866633097331
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,5120,2048,0.015286399920781454
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,6144,3072,0.01844586730003357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,2048,0.013582932949066161
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,2048,0.012036266922950744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,5120,1536,0.041978665192921955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,1536,0.011274666587511698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,1536,0.010718933741251628
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,5120,1024,0.03618880112965901
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,5120,1024,0.010743467013041179
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,1024,0.008874666690826417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,1024,0.009489066402117411
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,5120,768,0.03271146615346272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,5120,768,0.009654399752616883
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,768,0.007730133334795634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,768,0.008846933643023174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,5120,512,0.03032960096995036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,5120,512,0.008477866649627686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,512,0.0062837332487106325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,512,0.00820479989051819
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,6144,512,0.008643200000127155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,5120,256,0.02799786726633708
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,5120,256,0.007438933352629344
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,256,0.005302399893601735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,256,0.007523199915885926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,5120,128,0.026817067464192705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,128,0.004470400015513102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,128,0.007149866720040639
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,5120,64,0.02632853388786316
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,64,0.004402133325735727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,64,0.007430399954319
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,5120,32,0.026383999983469648
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,32,0.00444160004456838
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,32,0.007672533392906189
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,4096,65536,0.7607776006062825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,5120,16384,0.1143829345703125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,65536,0.28390932083129883
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,65536,0.20945493380228677
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,4096,16384,0.18517333666483562
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,16384,0.08809706370035807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,16384,0.05408746798833212
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,4096,12288,0.13678080240885418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,5120,8192,0.05938666661580404
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,4096,12288,0.04598613182703654
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,12288,0.07454613049825033
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,12288,0.04305386543273926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,4096,10240,0.039537068208058676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,4096,10240,0.11602986653645833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,10240,0.06610560019810995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,10240,0.0374570647875468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,4096,8192,0.09577173391977946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,8192,0.04923413197199504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,8192,0.031040000915527343
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,4096,7168,0.10512320200602214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,5120,4096,0.0291157325108846
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,5120,3584,0.023619200785954794
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,4096,7168,0.02908693353335063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,7168,0.04583146572113037
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,7168,0.027614933252334595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,4096,6144,0.0945642630259196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,6144,0.03856000105539958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,6144,0.02369173367818197
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,4096,5120,0.08102293014526367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,4096,5120,0.021449599663416544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,5120,0.03490133285522461
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,5120,0.020103466510772706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,4096,4096,0.0699285348256429
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,4096,4096,0.0175872008005778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,4096,0.01725333333015442
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,4096,0.021208532651265464
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,4096,3584,0.06486506859461466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,3584,0.018273067474365235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,3584,0.015966932972272238
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,4096,3072,0.05925973256429037
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,4096,3072,0.01432960033416748
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,3072,0.015691733360290526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,5120,1536,0.012560000022252401
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,3072,0.0146506667137146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,4096,2560,0.05138453245162964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,4096,2560,0.013006933530171714
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,2560,0.014025599757830302
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,2560,0.013352533181508383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,4096,2048,0.0435263991355896
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,4096,2048,0.011427199840545655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,2048,0.01102293332417806
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,2048,0.01199893355369568
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,4096,1536,0.03848640124003093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,4096,1536,0.010129066308339436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,1536,0.009529599547386169
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,1536,0.010644267002741497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,4096,1024,0.033508265018463136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,4096,1024,0.008400000135103862
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,1024,0.00740479975938797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,1024,0.009271466732025146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,4096,768,0.030800000826517744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,4096,768,0.007632000247637431
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,768,0.006602666775385539
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,768,0.008694400389989216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,4096,512,0.028486400842666626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,4096,512,0.0068341334660847975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,512,0.0055189331372578945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,512,0.008089600006739299
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,4096,256,0.02630400061607361
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,4096,256,0.006192000210285186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,5120,128,0.006906666855017345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,256,0.004733866453170777
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,256,0.007468800246715546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,4096,128,0.025243733326594037
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,128,0.004131199916203817
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,128,0.007047466437021891
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,4096,64,0.024858667453130086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,64,0.0037600000699361167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,64,0.007249066730340321
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,4096,32,0.02478826642036438
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,32,0.003964799890915553
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,32,0.0074527998765309645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,4096,65536,0.22946346600850426
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3584,65536,0.738146146138509
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,65536,0.25668907165527344
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,65536,0.2025770664215088
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,4096,16384,0.05900373458862305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3584,16384,0.057093334197998044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3584,16384,0.17619199752807618
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,16384,0.053002667427062986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,16384,0.07076160113016763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3584,12288,0.13099946975708007
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,12288,0.05379413366317749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,12288,0.041867733001708984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3584,10240,0.11082560221354168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,10240,0.04447360038757324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,4096,8192,0.032476800680160525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,10240,0.03662826617558797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3584,8192,0.03143786589304606
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3584,8192,0.09294506708780924
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,8192,0.03619733254114787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,8192,0.03018346627553304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3584,7168,0.1015232006708781
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,7168,0.033139199018478394
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,7168,0.02605866591135661
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3584,6144,0.09218986829121908
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,4096,6144,0.024780799945195518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3584,6144,0.024281599124272666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,6144,0.022541866699854533
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,6144,0.02895146608352661
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3584,5120,0.07854932943979899
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,5120,0.025536000728607178
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,5120,0.019882667064666747
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,4096,3584,0.01588159998257955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3584,4096,0.06828373273213705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3584,4096,0.0176362673441569
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,4096,0.019828265905380248
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,4096,0.01700906753540039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3584,3584,0.06376853386561075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,3584,0.016151466965675355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,3584,0.015793066223462424
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3584,3072,0.057767466704050696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3584,3072,0.014227199554443359
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,3072,0.015154133240381876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,3072,0.014601600170135499
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3584,2560,0.04972906510035197
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3584,2560,0.012914133071899415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,2560,0.012692266702651977
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,2560,0.013211733102798462
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3584,2048,0.04263039827346802
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3584,2048,0.011345066626866658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,2048,0.0130730668703715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,2048,0.011783466736475626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3584,1536,0.03791786829630534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3584,1536,0.009991466999053955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,1536,0.009422933061917622
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,1536,0.010571733117103577
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3584,1024,0.0334826668103536
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3584,1024,0.008534399668375652
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,4096,128,0.005747200051943461
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,1024,0.00886826713879903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,1024,0.009301333626111349
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3584,768,0.030798933903376263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,768,0.007310933371384938
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,768,0.008638933300971985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3584,512,0.028538666168848675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3584,512,0.006853333115577698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,512,0.006022400160630544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,512,0.008030933141708375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3584,256,0.026362667481104534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3584,256,0.006140799820423126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,256,0.004862933357556661
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,256,0.00747519979874293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3584,128,0.025355732440948485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3584,128,0.005710933109124502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,128,0.004276266694068909
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,128,0.006996266543865204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3584,64,0.02486720085144043
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,64,0.003947733342647553
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,64,0.007214933137098948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3584,32,0.024795732895533242
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,32,0.004109866668780645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3584,65536,0.20621652603149415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,32,0.00747626672188441
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3072,65536,0.19555306434631348
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,65536,0.2150965372721354
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3072,65536,0.6992256164550781
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,65536,0.19550506273905438
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3584,12288,0.044446933269500735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3072,16384,0.1704810619354248
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3072,16384,0.05313706795374552
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,16384,0.06021546522776285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,16384,0.0512223998705546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3072,12288,0.12356159687042237
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3072,12288,0.041382400194803874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3584,10240,0.03831466833750407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,12288,0.045932801564534505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,12288,0.04100480079650879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3072,10240,0.10705066521962484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,10240,0.03900373379389445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,10240,0.03536746501922607
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3072,8192,0.02947733402252197
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3072,8192,0.08947626749674478
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,8192,0.03259306748708089
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3584,7168,0.027703466018040974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,8192,0.028267733256022137
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3072,7168,0.10011839866638184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,7168,0.02895466685295105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,7168,0.026195200284322102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3072,6144,0.09027307033538819
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3072,6144,0.022959999243418374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,6144,0.02467413345972697
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,6144,0.022634667158126832
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3072,5120,0.07741653124491374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,5120,0.022810665766398112
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,5120,0.019563732544581096
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3584,5120,0.020838399728139244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3072,4096,0.06671040058135987
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3072,4096,0.01714026729265849
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,4096,0.016032000382741295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,4096,0.01691733400026957
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3584,3584,0.015863466262817382
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3072,3584,0.061672532558441163
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,3584,0.015943466623624166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,3584,0.015503999590873719
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3072,3072,0.05634453296661377
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3072,3072,0.01385599970817566
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,3072,0.013986133535703025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,3072,0.014262400070826211
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3072,2560,0.049394134680430093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3072,2560,0.012542933225631714
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,2560,0.012139733632405598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,2560,0.012970667084058127
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3072,2048,0.04352213144302368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,2048,0.0108842670917511
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,2048,0.011737599968910217
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3072,1536,0.037272532780965165
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3072,1536,0.009732266267140705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,1536,0.009141332904497783
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,1536,0.010550399621327717
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3072,1024,0.03324480056762695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3072,1024,0.008899199962615966
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,1024,0.006805333495140076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,1024,0.009194667140642803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3072,768,0.030637866258621214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,768,0.006265600025653839
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,768,0.008748799562454224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3072,512,0.028485333919525145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3072,512,0.006798933446407318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,512,0.005268266797065735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,512,0.008099199831485748
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3072,256,0.026257065931955974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3072,256,0.0060362666845321655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,256,0.004408533374468485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,256,0.007546666761239369
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3072,128,0.025360000133514405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3584,768,0.0075552001595497135
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3072,128,0.005666133264700572
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,128,0.003868799904982249
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,128,0.007011199990908305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3072,64,0.024612265825271606
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,64,0.0037237333754698435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,64,0.0071370666225751235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,3072,32,0.02478613257408142
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,32,0.003899733225504557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,32,0.00732479989528656
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2560,65536,0.1582688013712565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,65536,0.18864639600118
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2560,65536,0.6726666768391927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,65536,0.19568533897399903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2560,16384,0.1639370600382487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2560,16384,0.04310400088628133
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,16384,0.05528746843338013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,16384,0.051863467693328856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2560,12288,0.03368959824244182
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2560,12288,0.12249173323313396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,12288,0.03974506855010986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,12288,0.040805331865946454
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2560,10240,0.10219093163808186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2560,10240,0.029234133164087933
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,10240,0.034270934263865155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,10240,0.03522666692733765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2560,8192,0.08652693430582682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2560,8192,0.0240831991036733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,8192,0.02717546621958415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,8192,0.027884799242019653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2560,7168,0.09684373537699381
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2560,7168,0.021057067314783733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,7168,0.02477546731630961
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,7168,0.024989867210388185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2560,6144,0.08717013200124105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,6144,0.021265067656834922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,6144,0.022191999355951945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2560,5120,0.0748416026433309
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3072,10240,0.03584853410720825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3072,7168,0.026250666379928587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,5120,0.01880533297856649
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,5120,0.019523199399312338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2560,4096,0.06525439818700154
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2560,4096,0.014177067081133523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3072,5120,0.019831466674804687
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,4096,0.015130666891733804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,4096,0.016732800006866454
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2560,3584,0.06040853261947632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2560,3584,0.01292693316936493
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,3584,0.013673599561055502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,3584,0.015494400262832641
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2560,3072,0.0543509324391683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2560,3072,0.01159999966621399
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3072,3584,0.015309866269429526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,3072,0.012154666582743327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,3072,0.014242133498191834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2560,2560,0.04853866497675578
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,2560,0.010846933722496033
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,2560,0.013032533725102744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2560,2048,0.04121919870376587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2560,2048,0.009399466713269551
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,2048,0.009249066313107807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,2048,0.011782399813334147
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2560,1536,0.03699306646982829
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2560,1536,0.008408533533414204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,1536,0.007758933305740357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3072,2048,0.011311999956766764
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,1536,0.010481066505114238
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2560,1024,0.03281066616376241
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2560,1024,0.007211733361085255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,1024,0.006146133442719777
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,1024,0.009157333771387737
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3072,768,0.007609599828720092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2560,768,0.030184533198674517
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2560,768,0.006656000018119812
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,768,0.005406933526198069
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,512,0.004737066725889841
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,768,0.008648533622423809
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,512,0.007950933277606964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2560,512,0.02816213369369507
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2560,256,0.026096000274022417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2560,256,0.005634133517742157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,256,0.004071466624736786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,256,0.007426133255163829
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2560,128,0.02524906595547994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2560,128,0.005235200126965841
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,128,0.003659733384847641
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,128,0.007109333574771881
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2560,64,0.02456106742223104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,64,0.0036202666660149893
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,64,0.0071168000499407455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2560,32,0.02441493272781372
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,32,0.0036533333361148832
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,32,0.007191466788450877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,65536,0.15115733146667482
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,65536,0.19207253456115722
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2048,16384,0.15731733640034992
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2048,65536,0.6627989451090495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,16384,0.05013653437296549
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,16384,0.05192426840464274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2048,12288,0.11810026963551838
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,12288,0.04723946650822957
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,12288,0.03982400099436442
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2048,10240,0.1002570629119873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2048,10240,0.028615466753641766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,10240,0.04138026634852092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,10240,0.0327946662902832
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2048,8192,0.08506027062733969
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2048,8192,0.0231221338113149
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,8192,0.031140265862147014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,8192,0.02709760069847107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2048,7168,0.09518933296203613
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2048,7168,0.019662932554880778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,7168,0.028221867481867474
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,7168,0.024217599630355836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2048,6144,0.0859328031539917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2048,6144,0.016771199305852254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,6144,0.023891200621922813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,6144,0.021732266743977866
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2560,6144,0.01858666737874349
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2048,5120,0.07362879912058512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,5120,0.020115200678507486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,5120,0.01928640007972717
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2560,5120,0.01639893352985382
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2048,4096,0.06342080036799112
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2048,4096,0.012573867042859396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,4096,0.014602667093276978
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,4096,0.01672853430112203
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2048,3584,0.057980799674987794
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,3584,0.013347199559211731
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,3584,0.015521066387494406
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2048,3072,0.05342719952265421
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2048,3072,0.010743467013041179
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,3072,0.012513066331545511
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,3072,0.014243200421333313
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2048,2560,0.04538346529006958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2048,2560,0.009719467163085938
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,2560,0.010637866457303365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2560,2560,0.010660266876220703
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,2560,0.012964266538619994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2048,2048,0.041017599900563556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2560,512,0.006190933287143707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,2048,0.00897920032342275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,2048,0.011557333668073018
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2048,1536,0.007758933305740357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2048,1536,0.03700266679128011
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,1536,0.007720533510049183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,1536,0.01037013332049052
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2048,1024,0.032866134246190386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,1024,0.005975466469923655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,1024,0.009096533060073853
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2048,768,0.03019839922587077
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2048,768,0.006279466549555461
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,768,0.005383466680844625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,768,0.008547199765841167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2048,512,0.028036266565322876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2048,512,0.005845333139101664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,512,0.0046069333950678505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2048,65536,0.1690293312072754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,512,0.007933866480986278
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2048,256,0.026114133993784587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2048,256,0.005407999952634176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,256,0.003934933245182038
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,256,0.0073642666141192125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2048,128,0.024895999828974405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2048,128,0.005147733290990194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,128,0.0035743998984495797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2048,16384,0.04627626736958822
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,128,0.0069567998250325514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2048,64,0.02452053427696228
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,64,0.003386666625738144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2048,12288,0.034782934188842776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,64,0.007086933155854543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,2048,32,0.023937066396077476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,32,0.0035968000690142312
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,32,0.007099733253320058
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1536,65536,0.14468159675598144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,65536,0.12950719992319742
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,65536,0.19251093864440919
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1536,16384,0.03847786585489909
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1536,16384,0.15872000058492025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1536,65536,0.6356682459513346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,16384,0.03711680173873901
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,16384,0.05111253261566162
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1536,12288,0.11747519969940186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1536,12288,0.030096000432968138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,12288,0.03041173418362935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,12288,0.03779093424479167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1536,10240,0.10028479894002278
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1536,10240,0.025586134195327757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,10240,0.0262880007425944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,10240,0.031959466139475506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1536,8192,0.020873600244522096
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1536,8192,0.08500906626383463
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,8192,0.020863999923070274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,8192,0.026713599761327107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1536,7168,0.0957098642985026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,7168,0.018424532810846963
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,7168,0.024036266406377158
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1536,6144,0.08649706840515137
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1536,6144,0.016472533345222473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,6144,0.01602133313814799
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,6144,0.02193173368771871
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1536,5120,0.0144405335187912
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1536,5120,0.07381760279337565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,5120,0.01921066641807556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,5120,0.014984533190727234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2048,5120,0.014773333072662353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1536,4096,0.06382613182067871
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,4096,0.012796800335248312
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,4096,0.016718933979670204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1536,3584,0.05876053174336752
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1536,3584,0.011851732929547627
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,3584,0.010936533411343891
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,3584,0.015421866377194723
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2048,3584,0.011762133240699768
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1536,3072,0.05263359944025675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,3072,0.009578667084376017
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,3072,0.014121599992116294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1536,2560,0.046741334597269694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1536,2560,0.009567999839782714
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,2560,0.010013866424560546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,2560,0.012935466567675271
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1536,2048,0.04071466525395711
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,2048,0.008903466661771138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,2048,0.011496532956759136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1536,1536,0.03654613494873047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,1536,0.007180800040562947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,1536,0.010397866368293762
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1536,1024,0.03274986743927002
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1536,1024,0.00695253312587738
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,1024,0.005763199925422668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,1024,0.009136000275611877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1536,768,0.030113067229588824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1536,768,0.006420266628265381
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,768,0.005106133222579956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,768,0.008534399668375652
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1536,512,0.027923200527826948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2048,2048,0.008583466211954754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2048,1024,0.006844800213972728
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,512,0.004461866617202759
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,512,0.007912533481915791
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1536,256,0.025760000944137572
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,256,0.003885866701602936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,256,0.007314133147398631
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1536,128,0.0248906672000885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1536,128,0.005144533514976501
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,128,0.0036085332433382668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,128,0.006985599795977275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1536,64,0.024434133370717367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,64,0.0033098667860031127
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,64,0.007075199981530507
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1536,32,0.024399999777475992
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,32,0.0033290666838486993
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,32,0.007044266661008198
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1024,65536,0.635869852701823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1024,65536,0.12087679704030355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,65536,0.09778986771901449
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,65536,0.1928864002227783
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1024,16384,0.15797972679138184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1536,7168,0.01876693367958069
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,16384,0.029502934217453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,16384,0.04943466583887736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1024,12288,0.11736000378926595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1536,4096,0.012597333391507468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1536,3072,0.010616532961527507
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,12288,0.02807360092798869
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,12288,0.03716586828231812
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1536,2048,0.008585600058237712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1024,10240,0.1008682648340861
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1024,10240,0.020372267564137778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,10240,0.0247871994972229
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,10240,0.03213653365770976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1024,8192,0.017041067282358803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1024,8192,0.08527146975199382
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,8192,0.01667520006497701
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,8192,0.026718932390213012
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1024,7168,0.015588266650835672
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1024,7168,0.09547626972198486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1536,1536,0.007710933188597361
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,7168,0.01493333379427592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,7168,0.024206932385762533
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1024,6144,0.08612799644470215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1024,6144,0.01373546620210012
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,6144,0.013318399588267008
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,6144,0.02175253431002299
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1024,5120,0.012319999933242797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1024,5120,0.07368426322937012
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,5120,0.011665067076683045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,5120,0.01935360034306844
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1024,4096,0.06340479850769043
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1024,4096,0.010846933722496033
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,4096,0.010199466347694397
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,4096,0.016666666666666666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1024,3584,0.0577621340751648
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1024,3584,0.010108799735705058
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,3584,0.009131733576456707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,3584,0.01541866660118103
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1024,3072,0.05318933327992757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1024,3072,0.009092266360918682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,3072,0.00844266712665558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,3072,0.014180266857147216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1536,512,0.005874133110046387
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1024,2560,0.0451904018719991
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,2560,0.008215466638406117
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,2560,0.01290986637274424
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1024,2048,0.04029119809468587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,2048,0.0069461335738499955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,2048,0.01158186693986257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1024,1536,0.03622186581293742
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,1536,0.006095999975999197
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,1536,0.010285866260528565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1024,1024,0.03216639955838521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1024,1024,0.006190933287143707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,1024,0.0050016000866889955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,1024,0.00904960036277771
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1024,768,0.029890133937199907
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,768,0.004523733258247375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,768,0.008525866270065307
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1024,512,0.02770026723543803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1024,512,0.005525333185990652
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1536,256,0.005379199981689453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,512,0.004105599969625473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,512,0.007855999966462452
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1024,256,0.025865600506464644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,256,0.0037109332780043284
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,256,0.007302399973074596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1024,128,0.024943999449412026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1024,128,0.004907733201980591
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,128,0.003389866650104523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,128,0.006994133194287618
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1024,64,0.024371200799942018
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,64,0.0032117334504922234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,64,0.007034666836261749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,1024,32,0.023794132471084594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,32,0.003311999887228012
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,32,0.0070165331164995836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,768,65536,0.636634635925293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1024,16384,0.03067306677500407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,768,65536,0.10807253519694011
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,65536,0.06789440313975016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,65536,0.19204373359680177
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,768,16384,0.15750080744425457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1024,12288,0.023468800385793052
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,768,16384,0.02994026740392049
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,16384,0.025630933046340943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,16384,0.048077865441640215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,768,12288,0.023066665728886923
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,768,12288,0.11785279909769694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,12288,0.020753065745035805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,12288,0.03736213445663452
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,768,10240,0.10054720242818196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,10240,0.019514666001001994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,10240,0.03163093328475952
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,768,8192,0.08477760155995687
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,8192,0.01618666648864746
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,8192,0.026743467648824053
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,768,7168,0.09557226498921713
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1024,2560,0.008408533533414204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1024,1536,0.006968533496061961
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,7168,0.014634666840235391
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,7168,0.02413546641667684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1024,2048,0.007575466732184092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,768,6144,0.08592000007629394
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,6144,0.013169067104657492
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,6144,0.021637332439422608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,768,5120,0.07347413698832193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,5120,0.011524266997973124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,5120,0.019248000780741372
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1024,768,0.0058112000425656635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,768,4096,0.06275306542714437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,768,4096,0.010794666409492493
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,4096,0.009900800387064616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,4096,0.016658133268356322
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,768,3584,0.05783786773681641
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,3584,0.009165866176287334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1024,256,0.005099733173847198
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,3584,0.01542080044746399
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,768,3072,0.051478401819864905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,3072,0.008246399958928426
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,3072,0.014103466272354126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,768,2560,0.045607467492421463
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,768,10240,0.019885865847269694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,2560,0.008312533299128216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,2560,0.012820266683896384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,768,2048,0.040236798922220866
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,768,8192,0.01685653328895569
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,2048,0.007400533556938172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,2048,0.01160533328851064
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,768,1536,0.036414933204650876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,768,1536,0.00699946681658427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,1536,0.005824000140031179
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,1536,0.010389332969983418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,768,1024,0.03226986726125081
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,768,1024,0.006232533355553945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,768,7168,0.015406933426856995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,1024,0.004870399832725525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,1024,0.009106133381525676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,768,768,0.029849600791931153
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,768,0.004503466685612996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,768,0.008507733543713886
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,768,512,0.027687466144561766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,768,6144,0.013622400164604188
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,512,0.004530133306980133
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,512,0.00783253312110901
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,768,256,0.025541333357493083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,768,256,0.005047466854254405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,256,0.0036127999424934386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,256,0.007299200197060903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,768,128,0.024588799476623534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,768,5120,0.012245333194732666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,128,0.0033642667035261786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,768,3584,0.010050132870674133
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,128,0.006960000097751618
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,768,64,0.024216532707214355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,64,0.0032245332996050516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,64,0.0069578667481740315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,768,2560,0.008360532919565837
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,768,32,0.024119466543197632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,32,0.003202133377393087
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,768,3072,0.009012266993522644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,32,0.007065600156784058
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,65536,0.05376000006993612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,65536,0.1914122740427653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,512,16384,0.15806934038798015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,512,16384,0.026764800151189167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,512,65536,0.6341194788614909
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,16384,0.016744534174601235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,16384,0.04716053406397502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,512,12288,0.020707199970881142
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,512,12288,0.11757013003031414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,12288,0.015727999806404113
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,12288,0.03675093253453572
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,512,10240,0.018049067258834837
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,512,10240,0.10099733670552571
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,10240,0.014454399545987448
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,10240,0.03170880079269409
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,512,8192,0.08513387044270834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,8192,0.012434132893880208
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,8192,0.026548266410827637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,512,7168,0.09499413172403971
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,512,7168,0.013953066865603127
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,7168,0.011752532919247945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,7168,0.024119466543197632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,512,6144,0.0859615961710612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,6144,0.011717333396275838
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,6144,0.0215829332669576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,512,5120,0.07320319811503093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,512,5120,0.0115146666765213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,5120,0.010342400272687275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,5120,0.019244800011316933
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,512,4096,0.062414932250976565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,768,2048,0.007605333129564922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,4096,0.008909866213798523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,4096,0.016658133268356322
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,512,3584,0.05773546695709229
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,768,768,0.005884799857934316
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,3584,0.008181333541870117
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,3584,0.015447466572125753
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,512,3072,0.05116053422292074
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,512,3072,0.008594133456548055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,3072,0.007506133119265239
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,3072,0.014164266983668008
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,512,2560,0.045305601755778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,512,2560,0.007971199850241344
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,768,512,0.005453866720199585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,2560,0.007653333246707916
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,2560,0.012939733266830445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,512,2048,0.04056106805801392
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,512,2048,0.0073066666722297665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,2048,0.006843733290831248
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,2048,0.011558399597803751
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,512,1536,0.03621546824773152
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,512,1536,0.006751999755700429
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,1536,0.0057087997595469155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,1536,0.010334933797518413
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,512,1024,0.03224213322003682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,512,1024,0.005982933441797892
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,1024,0.004931200047334035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,768,128,0.004823466638724009
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,1024,0.009052800138791402
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,512,768,0.029616000254948933
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,768,0.004379733403523763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,768,0.008539733290672303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,512,512,0.02753173311551412
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,512,512,0.005377066632111868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,512,0.003973333289225897
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,512,0.007895466685295106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,512,256,0.025616000096003216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,512,256,0.005070933202902476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,256,0.003492266684770584
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,256,0.0072618668278058365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,512,128,0.024663466215133666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,512,128,0.0047872001926104225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,128,0.0032085334261258446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,128,0.00697920024394989
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,512,64,0.02439893285433451
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,64,0.003066666672627131
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,64,0.007031466563542683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,512,32,0.023517866929372154
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,32,0.0030901332696278887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,32,0.007062399884064992
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,256,65536,0.6372586568196614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,512,65536,0.09872852961222331
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,65536,0.03856000105539958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,65536,0.19325013160705568
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,256,16384,0.15773119926452636
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,256,16384,0.025438932577768962
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,16384,0.01134399970372518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,16384,0.04668480157852173
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,256,12288,0.11815253098805745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,12288,0.0121888001759847
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,12288,0.036713600158691406
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,512,8192,0.015371732910474143
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,256,10240,0.017192532618840538
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,256,10240,0.10050880114237468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,10240,0.011339733004570007
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,10240,0.031750400861104325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,512,6144,0.01267306705315908
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,256,8192,0.014749866724014283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,256,8192,0.08488213221232097
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,8192,0.009890133142471313
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,8192,0.026661332448323565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,256,7168,0.09525120258331299
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,256,7168,0.013332266608874002
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,7168,0.024146133661270143
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,7168,0.009526399771372478
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,256,6144,0.08555733362833659
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,256,6144,0.012204800049463909
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,6144,0.008963200449943542
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,6144,0.021471999088923135
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,256,5120,0.0732373317082723
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,256,5120,0.010999466975529988
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,5120,0.008738133311271667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,5120,0.01919893423716227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,256,4096,0.06218560139338175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,4096,0.008141866823037466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,4096,0.016582399606704712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,256,3584,0.05754559834798177
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,256,3584,0.009176533420880635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,3584,0.00783253312110901
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,3584,0.01532906691233317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,256,3072,0.050843731562296546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,256,3072,0.00823040008544922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,3072,0.007179733117421467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,3072,0.01402239998181661
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,256,2560,0.04499306678771973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,256,2560,0.0075647999842961625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,2560,0.006451199948787689
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,2560,0.012760532895723977
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,256,2048,0.040046934286753336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,512,4096,0.010052266716957092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,2048,0.005757866799831391
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,2048,0.01145919958750407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,256,1536,0.03593920071919759
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,256,1536,0.006513066589832306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,1536,0.005096533397833506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,1536,0.010262399911880493
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,256,1024,0.031668267647425336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,256,1024,0.00576853354771932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,1024,0.004374399781227112
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,1024,0.008982400099436443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,256,768,0.029414399464925127
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,256,768,0.005478399991989136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,768,0.004069333275159201
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,768,0.008406399687131246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,256,512,0.02722346584002177
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,512,0.0036917333801587426
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,512,0.007781333227952321
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,512,3584,0.009483733773231506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,256,256,0.025226666529973345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,256,256,0.004859733581542969
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,256,0.003402666747570038
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,256,0.007208533088366191
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,256,128,0.02429866592089335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,128,0.0031839999059836066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,128,0.006838400165239971
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,256,64,0.02399253249168396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,64,0.003010133405526479
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,64,0.006885333359241486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,256,32,0.023859200874964397
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,32,0.0030464000999927522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,32,0.006942933301130931
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,128,65536,0.6343786875406902
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,128,65536,0.0938208023707072
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,512,768,0.005659733215967814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,65536,0.03371946811676026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,65536,0.19264853795369466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,128,16384,0.1579114596048991
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,128,16384,0.02498133381207784
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,16384,0.012616533041000366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,16384,0.0465119997660319
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,128,12288,0.11765973567962647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,12288,0.010738133390744527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,12288,0.03646613359451294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,128,10240,0.1000864028930664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,128,10240,0.017063466707865398
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,10240,0.009872000416119893
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,10240,0.03144960006078084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,128,8192,0.08446186383565267
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,256,65536,0.09437973499298095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,8192,0.008968533078829447
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,8192,0.02666986584663391
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,128,7168,0.09512853622436523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,128,7168,0.013331199685732523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,7168,0.008526933193206788
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,7168,0.024184532960255942
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,128,6144,0.08617173035939535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,128,6144,0.012227200468381246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,6144,0.008059733112653096
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,6144,0.021663999557495116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,128,5120,0.07333439985911051
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,128,5120,0.010991999506950378
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,5120,0.007565866907437642
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,256,12288,0.01986879905064901
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,5120,0.0193066676457723
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,128,4096,0.06188160181045532
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,128,4096,0.009786666433016459
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,4096,0.007117866476376851
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,4096,0.01662506659825643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,128,3584,0.05601813395818075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,3584,0.007525333265463512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,3584,0.015265066425005594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,128,3072,0.05093866586685181
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,256,4096,0.009754666686058044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,3072,0.007195733487606049
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,3072,0.013979732990264893
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,128,2560,0.04447893301645915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,256,2048,0.006951466699441274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,2560,0.0064640000462532045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,2560,0.012770133217175803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,128,2048,0.03965333302815755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,256,512,0.005205333232879639
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,2048,0.005714133381843567
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,2048,0.011468799908955891
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,128,1536,0.03580586512883504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,256,128,0.004691199958324432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,1536,0.005061333378156027
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,128,12288,0.019780266284942626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,1536,0.010213333368301391
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,128,1024,0.031933865944544476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,1024,0.004368000229199727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,1024,0.008965333302815754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,128,768,0.02932479977607727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,128,8192,0.014681599537531533
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,768,0.004060799876848856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,768,0.008468266328175862
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,128,512,0.02696533401807149
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,128,512,0.005144533514976501
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,512,0.003705599904060364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,512,0.007782400151093801
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,128,256,0.02508693337440491
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,128,256,0.004818133513132731
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,256,0.0033621333539485933
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,256,0.007178666690985362
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,128,128,0.024257065852483113
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,128,128,0.004675200084845225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,128,0.00308693324526151
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,128,0.006820266445477803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,128,64,0.023915733893712363
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,64,0.0029504001140594482
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,64,0.0068800002336502075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,128,32,0.023763199647267662
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,32,0.003019733230272929
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,32,0.00695253312587738
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,128,3584,0.009111467003822326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,65536,0.03065706690152486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,65536,0.19455787340799968
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,64,65536,0.6350026448567708
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,16384,0.010948266585667927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,64,16384,0.1583072026570638
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,16384,0.046632532278696695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,12288,0.00960640013217926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,64,12288,0.11806080341339112
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,128,2560,0.007505066692829132
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,12288,0.036448001861572266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,10240,0.008918399612108868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,64,10240,0.10060053666432697
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,10240,0.031516800324122116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,8192,0.008282666901747386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,64,8192,0.0840224027633667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,8192,0.026444800694783527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,128,3072,0.008205866813659668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,7168,0.0077344000339508055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,64,7168,0.09493652979532877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,7168,0.023990400632222495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,6144,0.006961066524187725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,64,6144,0.08547733624776205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,6144,0.021566933393478392
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,5120,0.007325866818428039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,64,5120,0.0728384017944336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,64,4096,0.05999146699905396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,5120,0.019128533204396565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,4096,0.006942933301130931
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,4096,0.016456533471743265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,64,3584,0.05564693212509155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,3584,0.007380266487598419
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,3584,0.015240533153216043
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,64,3072,0.05182400147120157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,3072,0.007110400001207988
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,3072,0.014017066359519959
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,64,2560,0.04437973499298096
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,2560,0.006309333443641663
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,2560,0.01260586678981781
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,64,2048,0.03990720113118489
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,2048,0.0055744002262751256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,2048,0.011333333452542622
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,64,1536,0.03585813442866008
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,1536,0.004956800242265066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,1536,0.010141866405804952
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,64,1024,0.03179733355840047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,1024,0.004264533519744873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,1024,0.008835200468699138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,64,768,0.029395200808842975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,768,0.003976533313592275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,768,0.008346666892369587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,64,512,0.02723199923833211
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,512,0.0036138666172822317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,512,0.007758933305740357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,64,256,0.025335466861724852
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,256,0.003272533416748047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,256,0.007195733487606049
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,64,128,0.024254933993021647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,128,0.003083733220895131
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,128,0.006769066552321117
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,64,64,0.023669334252675374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,64,0.0029088000456492106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,64,0.006799999872843425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,64,32,0.023564799626668295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,32,0.00290133332212766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,32,0.006929066777229309
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,65536,0.027106134096781413
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,128,2048,0.006894933183987935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,65536,0.19401173591613768
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,16384,0.008964266379674275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,32,16384,0.1581973393758138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,16384,0.04669653177261353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,12288,0.008248533308506011
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,32,12288,0.11709760030110676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,32,65536,0.6353759765625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,12288,0.036473600069681804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,32,10240,0.09990719954172769
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,10240,0.0076000000039736434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,10240,0.031531733274459836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,8192,0.007410133381684621
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,8192,0.02642986575762431
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,32,8192,0.08448426723480225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,7168,0.0070250665148099255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,32,7168,0.0946890672047933
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,7168,0.02392746607462565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,6144,0.006872533261775971
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,32,6144,0.08534719944000244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,5120,0.0072170664866765336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,6144,0.021502933899561563
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,5120,0.01920106609662374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,32,5120,0.07235199610392253
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,32,4096,0.06187093257904053
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,4096,0.00687360018491745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,128,1536,0.006618666648864746
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,4096,0.016438399751981102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,32,3584,0.05706666707992554
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,3584,0.00730560024579366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,3584,0.015250133474667868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,32,3072,0.05025279919306437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,3072,0.006970666845639547
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,3072,0.013925333817799887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,32,2560,0.04434986511866252
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,2560,0.006298666695753734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,2560,0.012823466459910074
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,32,2048,0.0395466685295105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,2048,0.00557226687669754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,2048,0.011406933267911274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,32,1536,0.03582613468170166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,1536,0.004906666775544485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,1536,0.010226133465766906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,32,1024,0.03200533390045166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,768,0.003961600114901861
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,1024,0.004233600199222564
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,1024,0.008865066369374593
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,32,768,0.029319467147191365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,128,1024,0.005673600236574808
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,768,0.008294400076071422
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,512,0.0035648000737031303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,32,512,0.02685760060946147
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,512,0.007733333110809326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,32,256,0.02546773354212443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,256,0.003257599969704946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,256,0.007226666808128357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,32,128,0.024410667022069295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,128,0.003044266750415166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,32,0.0029440000653266907
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,128,0.006749866902828217
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,32,64,0.023995733261108397
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,64,0.0028661333024501802
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,64,0.0068234667181968685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,384,32,32,0.02364693284034729
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,32,0.006853333115577698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,65536,16384,0.464462947845459
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,65536,12288,0.3330762545267741
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,128,768,0.005520000060399374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,65536,16384,0.5155349413553874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,65536,16384,0.9103167851765951
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,65536,12288,0.6243210474650065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,65536,12288,0.4123733202616374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,65536,16384,0.8583391825358072
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,65536,12288,0.6812149047851562
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,65536,10240,0.35657386779785155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,65536,10240,0.5455882390340169
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,65536,10240,0.5811274846394856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,65536,8192,0.2788319905598958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,65536,8192,0.42429227828979493
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,65536,8192,0.4626933415730794
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,65536,7168,0.21164372762044273
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,65536,7168,0.3688458760579427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,65536,7168,0.2678773244222005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,65536,7168,0.38374293645222984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,65536,6144,0.23607892990112306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,65536,6144,0.33732693990071616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,65536,6144,0.32087252934773763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,65536,5120,0.2086890697479248
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,65536,5120,0.27003520329793296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,65536,5120,0.24602986971537272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,65536,4096,0.16151466369628906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,65536,4096,0.22148373921712242
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,65536,4096,0.20673599243164062
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,65536,3584,0.14659627278645831
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,65536,3584,0.1967957337697347
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,65536,3584,0.17359360059102374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,65536,3072,0.13112746874491374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,65536,3072,0.09826239744822184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,65536,3072,0.17260586420694987
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,65536,3072,0.15031572977701824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,65536,2560,0.08490026791890462
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,65536,2560,0.11585280100504558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,65536,2560,0.14698452949523927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,65536,2560,0.11221866607666016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,65536,2048,0.09913280010223388
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,65536,2048,0.12085226376851399
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,65536,2048,0.09529493649800619
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,65536,1536,0.08529813289642334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,65536,1536,0.05879040161768595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,65536,10240,0.2828960100809733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,65536,8192,0.2265557289123535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,65536,1536,0.09481813112894694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,65536,1536,0.0679701328277588
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,65536,1024,0.04248426755269368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,65536,1024,0.07004160086313883
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,65536,1024,0.06690773169199625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,65536,1024,0.052239998181660974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,65536,768,0.06023893356323242
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,65536,768,0.054110932350158694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,65536,768,0.04705919822057088
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,65536,512,0.05410240093866984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,65536,512,0.04114773273468018
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,65536,6144,0.1799957275390625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,65536,512,0.0431818683942159
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,65536,256,0.0229695995648702
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,65536,256,0.050399998823801674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,65536,256,0.02914560039838155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,65536,256,0.04149866501490275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,65536,128,0.04460266828536987
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,65536,5120,0.15165653228759765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,65536,128,0.021322667598724365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,65536,64,0.043429334958394364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,65536,128,0.043408000469207765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,65536,64,0.01584106683731079
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,65536,4096,0.1241045316060384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,65536,64,0.04135040044784546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,65536,32,0.047319467862447104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,65536,32,0.01614400049050649
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,65536,32,0.04837013483047485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,65536,3584,0.11302080154418945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,16384,65536,0.46358613967895507
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,65536,0.8476618448893228
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,65536,0.9823914845784506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,16384,16384,0.19042347272237142
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,16384,65536,0.7824650446573893
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,16384,16384,0.12978773117065429
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,16384,0.21928213437398275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,16384,0.19357226689656576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,65536,2048,0.06936533451080322
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,12288,0.1550389289855957
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,12288,0.16748587290445965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,16384,12288,0.1531328042348226
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,16384,10240,0.13189120292663575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,10240,0.13259200255076092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,10240,0.12993919849395752
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,16384,8192,0.06444480021794638
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,16384,8192,0.10953919887542725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,8192,0.11318933169047038
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,8192,0.09703253110249838
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,16384,7168,0.11946667035420735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,16384,7168,0.05979520082473755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,7168,0.09572266737620036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,7168,0.08504746754964193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,16384,6144,0.05144639809926351
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,6144,0.08416852951049805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,16384,6144,0.10864853064219157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,6144,0.07230400244394938
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,16384,5120,0.09405439694722494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,16384,5120,0.04504426717758179
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,5120,0.07023466428120931
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,5120,0.05430399974187215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,16384,4096,0.037256534894307455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,4096,0.058253868420918786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,16384,4096,0.07890559832255045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,4096,0.046625065803527835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,16384,3584,0.07204480171203613
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,16384,3584,0.035087998708089194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,3584,0.05225280125935873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,3584,0.039477332433064775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,16384,3072,0.0660863995552063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,3072,0.04597973426183065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,3072,0.035070931911468504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,16384,2560,0.058133331934610996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,16384,2560,0.029296000798543293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,2560,0.04039040009180705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,2560,0.030925865968068438
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,16384,2048,0.02213866710662842
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,16384,2048,0.051437866687774655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,2048,0.03328213294347127
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,2048,0.026386133829752606
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,16384,1536,0.04262293179829915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,16384,1536,0.015530666708946228
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,1536,0.026641066869099932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,1536,0.020562134186426797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,16384,1024,0.03621120055516561
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,16384,1024,0.012518399953842163
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,1024,0.01729173262914022
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,1024,0.016402133305867515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,16384,768,0.032781867186228435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,16384,768,0.011169067025184632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,768,0.01237226625283559
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,768,0.014685866236686707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,16384,512,0.030085333188374835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,16384,512,0.00988266666730245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,512,0.00983893374602
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,512,0.01323306659857432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,16384,256,0.02768426736195882
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,16384,256,0.008743466933568318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,65536,768,0.03652799924214681
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,256,0.007434666653474172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,65536,512,0.033453865845998125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,256,0.01184213360150655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,16384,128,0.026745599508285523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,128,0.006308266520500183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,128,0.011115733782450359
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,16384,64,0.02667413353919983
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,64,0.005548800031344095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,64,0.012249599893887837
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,16384,32,0.026982400814692182
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,32,0.005876266459623972
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,32,0.015178666512171427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,65536,128,0.01856106718381246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,12288,65536,0.33327360153198243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,65536,0.6236245473225911
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,12288,65536,0.7089578628540039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,12288,16384,0.18699946403503417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,65536,0.7143893559773763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,16384,0.15856746037801106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,16384,0.1409759998321533
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,12288,12288,0.0707754691441854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,12288,12288,0.14346879323323566
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,12288,0.11973760128021241
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,12288,0.11420053641001385
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,12288,10240,0.06049066781997681
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,12288,10240,0.12435306708017986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,10240,0.09975252946217855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,10240,0.09893653392791749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,12288,8192,0.05095680157343546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,12288,8192,0.10349120299021404
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,16384,12288,0.09183359940846761
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,16384,10240,0.07846186955769857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,8192,0.08193919658660889
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,8192,0.06932799816131592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,7168,0.07142613728841146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,12288,7168,0.11382400194803874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,7168,0.0603925347328186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,12288,6144,0.04000853300094605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,12288,6144,0.1010975996653239
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,6144,0.05280213356018067
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,6144,0.06288640101750692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,12288,5120,0.08914026419321695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,12288,5120,0.03466026782989502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,5120,0.05328960021336874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,5120,0.042414931456247966
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,12288,4096,0.07406400044759115
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,12288,4096,0.029165865977605183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,4096,0.044210131963094076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,4096,0.03429973522822062
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,12288,3584,0.06866559982299805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,12288,3584,0.029912533362706502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,3584,0.039469865957895915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,3584,0.031276800235112506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,12288,3072,0.06218239863713583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,12288,3072,0.02426346739133199
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,3072,0.034968535105387374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,3072,0.02818453311920166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,12288,2560,0.05558293263117472
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,12288,2560,0.018145066499710084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,2560,0.03022293249766032
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,2560,0.026076799631118773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,12288,2048,0.014876799782117209
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,12288,2048,0.0455135981241862
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,16384,3072,0.030770132939020794
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,2048,0.02524159948031108
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,2048,0.02152000069618225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,12288,1536,0.03983146746953328
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,12288,1536,0.012562132875124612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,1536,0.020601600408554077
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,1536,0.016218666235605875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,12288,1024,0.035018666585286455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,12288,1024,0.010620799660682679
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,12288,768,0.03174399932225545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,1024,0.011702400445938111
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,1024,0.013645866513252258
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,12288,768,0.009497599800427754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,768,0.009825066725413004
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,768,0.012437333663304646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,12288,512,0.029850665728251142
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,512,0.008138666550318401
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,512,0.011450666189193725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,12288,256,0.027825067440668743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,12288,256,0.007677866518497467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,256,0.006289066871007283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,256,0.010319999853769938
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,12288,128,0.02647999922434489
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,12288,128,0.007063466807206471
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,128,0.005297066768010458
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,128,0.009790933132171631
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,12288,64,0.02646080056826274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,64,0.004753066599369049
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,64,0.00999679962793986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,12288,32,0.026950399080912273
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,32,0.005087999999523163
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,16384,128,0.007972266773382823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,32,0.0125109334786733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,10240,65536,0.2910869280497233
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,10240,16384,0.1750165303548177
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,65536,0.5430485407511394
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,65536,0.603222401936849
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,10240,16384,0.07925439675649007
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,10240,65536,0.6947370529174804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,16384,0.13709012667338055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,16384,0.1202741305033366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,10240,12288,0.1393781344095866
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,12288,0.10212799708048503
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,10240,12288,0.07008960247039794
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,12288,0.09489920139312744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,10240,10240,0.05561173359553019
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,10240,10240,0.12133013407389323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,10240,0.08570026556650798
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,10240,0.08504640261332194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,10240,8192,0.0443669319152832
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,10240,8192,0.10305813153584797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,8192,0.07171200116475424
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,8192,0.06254400014877319
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,10240,7168,0.040695468584696456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,10240,7168,0.11056426366170247
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,7168,0.06232106685638428
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,12288,16384,0.09205439885457357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,10240,6144,0.09969493548075357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,7168,0.05052266518274943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,10240,6144,0.03592853148778279
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,6144,0.054757332801818846
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,6144,0.04399253527323405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,10240,5120,0.031109333038330078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,10240,5120,0.08816213607788086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,5120,0.04600746631622314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,5120,0.03643519878387451
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,4096,0.03881066640218099
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,10240,4096,0.07314346631368002
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,4096,0.030133332808812457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,10240,3584,0.02411839962005615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,10240,3584,0.06809600194295248
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,3584,0.03439146677652995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,3584,0.027164800961812334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,10240,3072,0.017574399709701538
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,10240,3072,0.06157439947128296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,10240,2560,0.053350400924682614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,3072,0.030824534098307294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,12288,7168,0.045843199888865156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,3072,0.024286933739980063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,2560,0.026562132438023883
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,2560,0.02127573291460673
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,10240,2048,0.04439680178960164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,2048,0.021800533930460612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,2048,0.01774719953536987
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,10240,1536,0.038874665896097824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,1536,0.014651733636856078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,1536,0.014229333400726319
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,10240,1024,0.033869866530100504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,10240,1024,0.009577600161234538
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,1024,0.011377066373825073
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,1024,0.012268799543380737
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,10240,768,0.03168746630350749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,10240,768,0.0087936004002889
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,768,0.009640533725420635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,768,0.011341866850852967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,10240,512,0.029806933800379437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,10240,512,0.007995733122030894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,512,0.008070399860541026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,512,0.010523733496665955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,10240,256,0.02730453411738078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,10240,256,0.007165866593519847
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,256,0.006631466746330261
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,256,0.009687466422716777
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,10240,128,0.026574933528900148
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,128,0.005705599983533224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,128,0.009181867043177288
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,10240,64,0.026313600937525432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,64,0.004737066725889841
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,64,0.009360000491142273
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,10240,32,0.02650559941927592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,32,0.00498879998922348
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,32,0.010998400052388509
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,8192,65536,0.2409503936767578
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,12288,512,0.008611200253168742
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,8192,65536,0.5453919728597005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,65536,0.42173013687133787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,65536,0.47100054423014326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,8192,16384,0.1371498743693034
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,8192,16384,0.06888533433278402
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,16384,0.10874453385670979
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,16384,0.09102933406829834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,8192,12288,0.05172266562779745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,12288,0.08197866280873617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,8192,12288,0.11353920300801594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,12288,0.07473920186360677
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,8192,10240,0.04514666795730591
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,8192,10240,0.09745493729909262
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,10240,0.0681109348932902
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,10240,0.06802879969278972
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,8192,8192,0.08218453725179037
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,8192,0.05437973340352377
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,8192,0.044860800107320145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,8192,7168,0.03378986517588298
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,8192,7168,0.09625066916147867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,7168,0.04924799998601277
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,7168,0.04017386833826701
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,8192,6144,0.08589973449707031
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,10240,4096,0.026143999894460042
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,6144,0.04382719993591309
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,6144,0.03593173424402873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,8192,5120,0.07624959945678711
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,8192,5120,0.025614933172861738
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,5120,0.03705173333485921
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,5120,0.031192533175150555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,8192,4096,0.01925546725591024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,8192,4096,0.06354133288065592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,4096,0.030971733729044597
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,10240,2560,0.014998400211334228
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,4096,0.024590933322906496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,8192,3584,0.05970026652018229
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,8192,3584,0.016422399878501893
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,3584,0.02770986755688985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,3584,0.022895999749501548
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,8192,3072,0.0524288018544515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,3072,0.024911999702453613
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,10240,2048,0.012963199615478515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,10240,1536,0.011191466450691223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,3072,0.02110613385836283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,8192,2560,0.0469706654548645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,2560,0.021220266819000244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,2560,0.017413334051767985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,8192,2048,0.03996053139368693
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,2048,0.014442666371663412
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,2048,0.01461013356844584
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,8192,1536,0.03559146722157796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,8192,1536,0.01032426655292511
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,1536,0.01141973336537679
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,1536,0.012316800157229106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,8192,1024,0.03148373365402222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,8192,1024,0.008638933300971985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,1024,0.008851200342178345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,1024,0.010851200421651204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,8192,768,0.029439999659856157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,8192,768,0.00798399994770686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,768,0.007614933451016744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,768,0.010077866911888122
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,8192,512,0.02772480050722758
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,8192,512,0.007282133400440216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,512,0.006326400240262349
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,512,0.009422933061917622
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,10240,128,0.006723199784755707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,8192,256,0.025763199726740522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,8192,256,0.00664106657107671
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,256,0.005153066913286845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,256,0.008762666583061218
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,8192,128,0.025131734212239583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,8192,128,0.006190933287143707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,128,0.00446720023949941
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,128,0.008312533299128216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,8192,64,0.02480640014012655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,64,0.0040501333773136135
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,64,0.008405333757400513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,8192,32,0.025015467405319215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,32,0.00439680020014445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,32,0.009654399752616883
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,65536,0.3550304094950358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,7168,65536,0.5283199946085613
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,7168,16384,0.13253546555836995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,65536,0.44850880304972335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,7168,16384,0.06191573143005371
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,16384,0.0948479970296224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,16384,0.08639787038167318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,8192,8192,0.0368341326713562
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,7168,12288,0.047762131690979
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,7168,12288,0.10716373125712078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,12288,0.07201813062032064
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,12288,0.0726037343343099
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,7168,10240,0.040776534875233965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,7168,10240,0.09582826296488443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,10240,0.06083413362503052
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,10240,0.06456853151321411
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,7168,8192,0.0800704002380371
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,8192,0.04848106702168782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,8192,0.043150933583577473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,8192,6144,0.029425066709518433
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,7168,7168,0.09288000265757243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,7168,7168,0.030872533718744915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,7168,0.043782401084899905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,7168,0.0398144006729126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,7168,6144,0.08435306549072266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,7168,6144,0.027213867505391436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,6144,0.0389301339785258
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,6144,0.03513280153274536
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,7168,5120,0.07472533384958903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,7168,5120,0.02262399991353353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,5120,0.03406399885813395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,5120,0.030341333150863646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,7168,4096,0.018294399976730345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,7168,4096,0.062337064743041994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,4096,0.027453867594401042
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,4096,0.02485760052998861
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,7168,3584,0.057716266314188636
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,3584,0.024765866994857787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,3584,0.023307732741038003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,7168,3072,0.05176746845245361
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,7168,3072,0.015090133746465048
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,3072,0.02219840089480082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,3072,0.019090133905410766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,7168,2560,0.046087467670440675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,7168,2560,0.013868799805641175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,2560,0.0178656001885732
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,2560,0.016798933347066246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,7168,2048,0.0396938681602478
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,7168,2048,0.012147200107574464
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,2048,0.015145599842071533
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,2048,0.01434879998366038
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,7168,1536,0.03559786478678385
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,7168,1536,0.010812800129254658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,8192,3072,0.014345600207646688
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,1536,0.011613866686820984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,1536,0.012351999680201214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,7168,1024,0.0315338671207428
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,8192,2048,0.011412266890207927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,1024,0.008896000186602275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,8192,2560,0.013397333025932313
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,1024,0.010682666301727295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,7168,768,0.029739733537038165
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,7168,768,0.00823359986146291
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,768,0.007517866790294647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,768,0.010125866532325745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,7168,512,0.027742934226989747
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,7168,512,0.007720533510049183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,512,0.006202666461467743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,512,0.009330133597056072
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,7168,256,0.025882667303085326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,7168,256,0.006830933193365733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,256,0.005037866532802582
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,256,0.008709333340326945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,7168,128,0.025145600239435833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,128,0.0043935999274253845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,128,0.008427733182907104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,7168,64,0.024743467569351196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,64,0.003957333415746689
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,64,0.008309333523114523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,7168,32,0.024887466430664064
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,32,0.004233600199222564
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,32,0.009479467074076335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,65536,0.3145994822184245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,6144,65536,0.5168490727742513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,65536,0.35231253306070964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,7168,65536,0.24301652908325194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,6144,16384,0.1289728005727132
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,6144,16384,0.060268799463907875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,16384,0.0828266700108846
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,16384,0.06922986507415771
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,6144,12288,0.10418240229288737
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,12288,0.06344319979349772
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,12288,0.054953598976135255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,6144,10240,0.09288106759389242
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,6144,10240,0.041306666533152264
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,10240,0.05474346478780111
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,10240,0.0515008012453715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,6144,8192,0.03427626689275105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,6144,8192,0.07783573468526205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,8192,0.04316480159759521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,8192,0.03606933355331421
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,6144,7168,0.030821333328882854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,6144,7168,0.09147306283315024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,7168,0.039049601554870604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,7168,0.03226986726125081
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,6144,6144,0.08196693261464437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,6144,6144,0.02733760078748067
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,6144,0.03437973260879516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,7168,8192,0.033854933579762776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,6144,0.02805440028508504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,6144,5120,0.07294720013936361
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,5120,0.030558933814366657
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,5120,0.023407999674479166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,6144,4096,0.06112639904022217
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,6144,4096,0.01646293302377065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,4096,0.02444159984588623
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,4096,0.018871466318766274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,6144,3584,0.05576320091883341
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,6144,3584,0.015068800250689188
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,3584,0.022057600816090903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,3584,0.018153599898020425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,6144,3072,0.05036160151163736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,3072,0.020191999276479085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,3072,0.015707733233769734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,7168,3584,0.016770132382710776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,6144,2560,0.0443338672320048
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,6144,2560,0.012563199798266093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,2560,0.01613866686820984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,2560,0.013555199901262919
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,6144,2048,0.038652801513671876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,2048,0.01360640029112498
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,2048,0.01234773298104604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,6144,1536,0.0355402668317159
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,6144,1536,0.009763200084368389
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,1536,0.010798933108647664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,1536,0.010797866185506185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,6144,1024,0.0314410666624705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,6144,1024,0.008278400202592214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,1024,0.008461866776148479
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,7168,1024,0.00944533348083496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,1024,0.009460266431172688
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,6144,768,0.02959573268890381
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,7168,128,0.00645653357108434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,768,0.007234133283297221
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,768,0.008948266506195068
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,6144,512,0.027812266349792482
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,6144,512,0.0067104001839955645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,512,0.0057429333527882894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,512,0.0086325337489446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,6144,256,0.025733333826065064
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,6144,256,0.0061471998691558834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,256,0.004794666667779287
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,256,0.008001066744327545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,6144,128,0.02529173294703166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,6144,128,0.005766400198141734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,128,0.004251733422279358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,128,0.007541333138942718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,6144,65536,0.2156533400217692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,6144,64,0.024426666895548503
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,64,0.003804799914360046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,64,0.007811200122038524
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,6144,32,0.024528000752131143
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,32,0.004091733445723852
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,32,0.008697600166002909
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,5120,65536,0.1940661271413167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,65536,0.2880245208740234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,6144,12288,0.04798719882965088
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,5120,65536,0.5087552070617676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,5120,16384,0.12712639967600506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,65536,0.31068267822265627
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,16384,0.08073813120524088
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,16384,0.06557120084762573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,5120,12288,0.10136106808980305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,12288,0.0549397349357605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,12288,0.052856532732645664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,5120,10240,0.03755520184834798
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,5120,10240,0.09079893430074057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,10240,0.05233386754989624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,10240,0.044202665487925216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,5120,8192,0.03086079955101013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,5120,8192,0.07615573406219482
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,8192,0.03758293390274048
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,8192,0.033880531787872314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,5120,7168,0.08929386933644613
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,5120,7168,0.02747093240420024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,7168,0.034381866455078125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,6144,5120,0.020886399348576865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,7168,0.029286400477091475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,5120,6144,0.08033599853515624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,6144,0.03049280047416687
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,6144,0.02662186622619629
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,5120,5120,0.07122240066528321
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,5120,0.028677332401275634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,5120,0.022107734282811483
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,5120,4096,0.05835306644439697
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,6144,3072,0.013820800185203552
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,5120,4096,0.014844800035158793
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,4096,0.021794132391611733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,4096,0.018269866704940796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,5120,3584,0.014018133282661438
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,5120,3584,0.054093865553538
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,3584,0.020731733242670695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,3584,0.017004799842834473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,6144,2048,0.011211733023325603
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,5120,3072,0.04979626735051473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,3072,0.017307732502619425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,3072,0.015544533729553223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,5120,2560,0.04461439847946167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,5120,2560,0.011586133639017742
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,2560,0.014570666352907815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,2560,0.013819733262062072
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,5120,2048,0.039376000563303634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,5120,2048,0.01033066709836324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,2048,0.012414933244387309
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,2048,0.012237866719563801
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,5120,1536,0.03554986715316773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,5120,1536,0.008980266253153483
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,1536,0.010248532891273499
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,1536,0.0107914666334788
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,5120,1024,0.03147626717885335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,5120,1024,0.00774186650911967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,1024,0.007948799928029378
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,1024,0.00955839951833089
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,5120,768,0.029478400945663452
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,5120,768,0.007045333087444305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,768,0.006779733300209046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,768,0.00897706647713979
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,5120,512,0.006426666676998138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,5120,512,0.0277621328830719
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,512,0.00557226687669754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,512,0.008335999647776286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,5120,256,0.02595413327217102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,5120,256,0.005865600208441416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,256,0.004580266773700714
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,256,0.007880533238252004
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,5120,128,0.024948267141977946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,5120,128,0.005467733244101206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,6144,768,0.007454933226108551
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,128,0.004060799876848856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,128,0.007614933451016744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,5120,64,0.024513065814971924
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,64,0.003688533355792364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,64,0.007714133461316426
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,5120,32,0.024521599213282265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,32,0.0038901334007581077
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,32,0.008438400427500407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,4096,65536,0.15338133176167806
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,4096,65536,0.501201057434082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,65536,0.22514239947001138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,4096,16384,0.12362240155537922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,65536,0.2279285271962484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,16384,0.06481279929478964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,16384,0.04710186719894409
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,4096,12288,0.09917226632436117
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,12288,0.054771200815836585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,12288,0.03744959831237793
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,4096,10240,0.08832000096638998
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,4096,10240,0.028987733523050944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,5120,16384,0.055055999755859376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,10240,0.04955306847890218
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,10240,0.03508266607920329
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,4096,8192,0.07388906478881836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,4096,8192,0.02397759954134623
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,8192,0.03391466538111369
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,8192,0.025293866793314617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,5120,12288,0.044120534261067705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,4096,7168,0.021538132429122926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,4096,7168,0.0875157356262207
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,7168,0.03152640064557393
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,7168,0.021117866039276123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,4096,6144,0.07859946886698405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,6144,0.028462932507197065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,6144,0.018602667252222697
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,4096,5120,0.015406933426856995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,4096,5120,0.06977919737497965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,5120,0.026612265904744463
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,5120,0.016852267583211265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,4096,4096,0.057095468044281006
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,4096,0.019475199778874717
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,4096,0.0139765332142512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,4096,3584,0.05363839864730835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,5120,6144,0.020873600244522096
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,4096,3584,0.012382933497428894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,3584,0.017619200547536216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,3584,0.013304533561070761
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,4096,3072,0.048802133401234946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,5120,5120,0.01815999945004781
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,3072,0.0155157337586085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,4096,3072,0.011241599917411804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,3072,0.012459733088811238
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,4096,2560,0.042728531360626223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,4096,2560,0.010525866349538168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,2560,0.013384532928466798
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,2560,0.010986666878064473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,4096,2048,0.037921067078908285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,2048,0.011242666840553283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,2048,0.009815466403961182
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,4096,1536,0.03487786849339803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,4096,1536,0.008190933366616566
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,1536,0.009193600217501322
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,1536,0.008940800031026205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,4096,1024,0.031233066320419313
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,4096,1024,0.007019733389218648
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,5120,3072,0.012504532933235168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,1024,0.007218133409818013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,1024,0.008160000046094257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,4096,768,0.029244800408681233
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,4096,768,0.006537599861621857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,768,0.006218666831652323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,768,0.007760000228881836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,4096,512,0.027509333690007527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,4096,512,0.006053333481152853
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,512,0.005131733417510986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,512,0.007392000158627827
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,4096,256,0.02571306626001994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,256,0.004267733295758565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,256,0.00695253312587738
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,4096,128,0.024868265787760416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,4096,128,0.0052159999807675685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,128,0.0037685332198937735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,128,0.006695466736952464
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,4096,64,0.024439465999603272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,64,0.0034293333689371743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,64,0.007051733136177063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,4096,32,0.024425599972407022
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,32,0.0036949334045251214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,32,0.00746666689713796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3584,65536,0.15697919527689616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,65536,0.19387307167053222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3584,65536,0.4896085421244304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3584,16384,0.12089813550313314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,65536,0.21641386349995934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3584,16384,0.04383893410364787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,16384,0.05753493309020996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,16384,0.045501867930094406
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3584,12288,0.03413013219833374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3584,12288,0.09743359883626303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,12288,0.0470197319984436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,12288,0.03578346570332845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3584,10240,0.08664746284484863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3584,10240,0.029626667499542236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,10240,0.03699839909871419
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,10240,0.03068480094273885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,4096,16384,0.04224640130996704
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3584,8192,0.0722208023071289
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3584,8192,0.024265599250793458
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,8192,0.030140799283981324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,8192,0.02312320073445638
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3584,7168,0.08576853275299072
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3584,7168,0.021887999773025513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,7168,0.02714666724205017
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,4096,12288,0.03314346671104431
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,7168,0.020308266083399452
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3584,6144,0.07714560031890869
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,6144,0.024250666300455727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,6144,0.01890666683514913
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3584,5120,0.06880319913228353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,5120,0.018321067094802856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,5120,0.016619732975959776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3584,4096,0.05716053247451782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3584,4096,0.013859200477600097
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,4096,0.015778133273124696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,4096,0.013954133788744608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3584,3584,0.05331413348515829
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3584,3584,0.013040000200271606
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,3584,0.013916800419489542
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,3584,0.013424000144004822
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3584,3072,0.048545066515604654
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3584,3072,0.011808000008265178
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,3072,0.012781866391499839
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,3072,0.012462932864824932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,4096,6144,0.018518400192260743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3584,2560,0.04246293306350708
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3584,2560,0.010994133353233338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,2560,0.011522133151690166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,2560,0.01107413371404012
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3584,2048,0.03811200062433879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,2048,0.00962773362795512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,2048,0.009962667028109233
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,4096,4096,0.013401599725087485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3584,1536,0.034970664978027345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3584,1536,0.008618666728337606
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,1536,0.007624533275763195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,1536,0.009091200431187947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3584,1024,0.0310261329015096
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,1024,0.006093866626421611
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,1024,0.00841813286145528
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3584,768,0.029074132442474365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3584,768,0.006964266796906789
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,768,0.0054666668176651
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,768,0.008026666442553202
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3584,512,0.027566933631896974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3584,512,0.0063381334145863845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,512,0.004644266764322917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,512,0.007464533547560374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3584,256,0.025608533620834352
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3584,256,0.005688533186912537
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,256,0.003985066711902618
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,256,0.007111466427644093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3584,128,0.02480213244756063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,4096,2048,0.009218133489290873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3584,128,0.005459199845790863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,128,0.0036085332433382668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,128,0.00680213322242101
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3584,64,0.024573866526285806
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,64,0.0032981333633263906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,64,0.007048533360163371
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3584,32,0.02444480061531067
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,32,0.0035455999275048576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,32,0.007582933207352956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3072,65536,0.15796693166097003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,65536,0.16907626787821453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,65536,0.2044895966847738
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3072,16384,0.11913386980692546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,4096,256,0.005553066730499268
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3072,65536,0.48642454147338865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3072,16384,0.04498666524887085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,16384,0.05132586558659872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,16384,0.04572906494140625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3072,12288,0.034740265210469565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3072,12288,0.09557120005289713
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,12288,0.03447893460591634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,12288,0.04236160119374593
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3072,10240,0.028776532411575316
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3072,10240,0.08500373363494873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,10240,0.036133333047231035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,10240,0.02802026669184367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3072,8192,0.02210986614227295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3072,8192,0.07109119892120361
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,8192,0.02890026569366455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,8192,0.022657066583633423
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3072,7168,0.08490986824035644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3072,7168,0.020410666863123574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,7168,0.026374399662017822
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,7168,0.019926400979359944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3072,6144,0.07604053020477294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3072,6144,0.01551040013631185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,6144,0.02256960074106852
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,6144,0.018267732858657838
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3072,5120,0.06776426633199056
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3072,5120,0.013598933815956116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,5120,0.018897066513697304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,5120,0.016561067104339598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3072,4096,0.05661760171254476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,4096,0.014494933684666953
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,4096,0.013943466544151305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3072,3584,0.052116266886393224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3072,3584,0.011116799712181092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,3584,0.012470400333404541
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,3584,0.013453867038091025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3072,3072,0.04617600043614705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3072,3072,0.010024533669153849
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,3072,0.011964799960454305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,3072,0.012213333447774252
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3072,2560,0.04210026661554973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3072,2560,0.009392000238100688
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,2560,0.010216533144315084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,2560,0.011148800452550251
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3072,2048,0.03837333520253499
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3072,2048,0.008333866794904072
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,2048,0.00871573289235433
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,2048,0.010084266463915508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3072,1536,0.03513493140538533
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3072,1536,0.0075541332364082335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,1536,0.007195733487606049
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,1536,0.009196799993515015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3072,1024,0.03125226696332296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3072,1024,0.006583466629187266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,1024,0.005842133363087972
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,1024,0.008477866649627686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3072,768,0.02913493315378825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3072,768,0.006171733140945435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,768,0.0051146666208903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,768,0.007922133306662242
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3072,512,0.02749226689338684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3584,6144,0.018668800592422485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3072,512,0.005761066575845083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,512,0.0044383997718493145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,512,0.0073066666722297665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3072,256,0.0257205327351888
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3072,256,0.005350400010744731
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3584,5120,0.01618133286635081
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,256,0.003950933367013932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,256,0.007032533486684163
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3072,128,0.024884267648061117
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,128,0.003533866753180822
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,128,0.006676266590754191
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3072,64,0.024395734071731567
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,64,0.0032842665910720824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,64,0.007020799815654755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,3072,32,0.02422399918238322
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,32,0.0034495999415715536
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,32,0.007344000041484833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,65536,0.14576853116353353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,65536,0.19733972549438478
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2560,16384,0.11726720333099365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2560,65536,0.48715734481811523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,16384,0.04463680187861125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3584,2048,0.009867733716964722
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,16384,0.042029865582784015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2560,12288,0.0939786672592163
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,12288,0.03708373308181763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,12288,0.032126933336257935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3584,1024,0.007663999994595845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2560,10240,0.02469653288523356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2560,10240,0.08369173208872477
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,10240,0.03379840056101481
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,10240,0.027458133300145467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2560,8192,0.07047680219014486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,8192,0.023280000686645506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,8192,0.021747199694315593
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2560,7168,0.08361706733703614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2560,7168,0.018045866489410402
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,7168,0.02040106654167175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,7168,0.019631999731063842
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2560,6144,0.07490773200988769
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,6144,0.015574399630228677
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,6144,0.017864533265431724
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2560,5120,0.06644800106684366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2560,5120,0.01365226705869039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,5120,0.014460800091425577
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,5120,0.01623040040334066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2560,4096,0.055631999174753824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2560,4096,0.01167466640472412
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,4096,0.01307093302408854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,4096,0.013557333747545877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2560,3584,0.052321068445841466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2560,3584,0.011168000102043153
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,3584,0.011773866415023804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,3584,0.013005866607030233
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2560,3072,0.045821865399678544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2560,3072,0.01015786627928416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,3072,0.010759466886520385
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,3072,0.011907200018564861
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2560,2560,0.042131201426188154
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2560,2560,0.00963200032711029
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,2560,0.009380267063776652
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,2560,0.010730666915575664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2560,2048,0.03774826526641846
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3072,4096,0.011661866307258606
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,2048,0.00798399994770686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,2048,0.00983893374602
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2560,1536,0.03502613306045532
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2560,1536,0.007739733159542084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,1536,0.006793599824110668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,1536,0.009078400333722432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3072,128,0.005093333125114441
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2560,1024,0.030844799677530926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2560,1024,0.006788266698519389
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,1024,0.005618133147557576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,1024,0.008341333270072937
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2560,768,0.029098665714263915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2560,768,0.006291200220584869
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,768,0.005011199911435445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,768,0.00786133309205373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2560,512,0.02738986611366272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,512,0.004347733159859975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,512,0.007321600119272869
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2560,65536,0.12645333607991535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2560,256,0.025628799200057985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2560,256,0.005388799806435903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,256,0.003804799914360046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,256,0.006900266806284587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2560,16384,0.03631680011749268
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2560,128,0.02474453250567118
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,128,0.0034261333445707956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,128,0.006738133231798808
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2560,64,0.024191999435424806
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,64,0.0032448001205921174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,64,0.006913066903750102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2560,12288,0.028384000062942505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2560,32,0.02431360085805257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,32,0.003335466732581457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,32,0.007077333331108093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2560,8192,0.020402133464813232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,65536,0.11918400128682453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,65536,0.12300586700439453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2048,65536,0.48633174896240233
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2048,16384,0.1156885306040446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2048,16384,0.03004266619682312
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,16384,0.03830186525980632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,16384,0.02802773316701253
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2560,6144,0.015924266974131265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2048,12288,0.023336533705393472
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2048,12288,0.09320000012715658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,12288,0.03216320077578227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,12288,0.020054399967193604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2048,10240,0.08333653608957926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,10240,0.029444267352422077
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,10240,0.01628373364607493
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2048,8192,0.069814403851827
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,8192,0.020987733205159505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,8192,0.014005333185195923
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2048,7168,0.08366933663686117
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2048,7168,0.01498240033785502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,7168,0.019994666179021202
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,7168,0.013148799538612366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2048,6144,0.07551573117574056
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2048,6144,0.013503999511400858
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,6144,0.016985599199930826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,6144,0.01223360002040863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2048,5120,0.06713600158691406
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2048,5120,0.011743999520937602
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,5120,0.01441493332386017
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,5120,0.011493333180745443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2048,4096,0.05628266731897989
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2048,4096,0.010332799951235453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,4096,0.010615467031796774
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,4096,0.010095999638239542
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2048,3584,0.05066986481348673
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2048,3584,0.00978773335615794
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,3584,0.010247466961542766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,3584,0.009648000200589497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2048,3072,0.04700160026550293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2048,3072,0.008940800031026205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,3072,0.00879039963086446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,3072,0.00920746624469757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2048,2560,0.0426911989847819
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2048,2560,0.0084906667470932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,2560,0.007707733412583668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,2560,0.008671999971071879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2048,2048,0.03824533224105835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2048,2048,0.007561600208282471
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,2048,0.006915199756622315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,2048,0.007965866724650066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2048,1536,0.0347541332244873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2048,1536,0.00698773314555486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,1536,0.006116266548633576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,1536,0.007566933333873749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2048,1024,0.030963200330734252
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2048,1024,0.006196266909440359
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,1024,0.0050805335243542995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2560,2048,0.008513066172599792
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,1024,0.007095466554164887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2048,768,0.02922346591949463
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,768,0.004705066482226053
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,768,0.006902400155862172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2048,512,0.027321600914001466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2048,512,0.00547626664241155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,512,0.00412266676624616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,512,0.0066101332505544026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2048,256,0.025434666872024538
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2048,256,0.0052255998055140175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,256,0.0037610667447249093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,256,0.006270933151245117
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2048,128,0.024741333723068238
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2048,128,0.004896000027656555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,128,0.003457066665093104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,128,0.006071466704209646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2048,64,0.024129066864649454
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,64,0.0032277333239714304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,64,0.006286933521429698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,2048,32,0.023931733767191567
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,32,0.003373866776625315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,32,0.006478933493296306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1536,65536,0.4847968101501465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1536,65536,0.09897279739379883
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,65536,0.0927669366200765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2560,512,0.005835733314355215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,65536,0.09957226912180582
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2560,128,0.005083733300367991
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1536,16384,0.028692267338434857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,16384,0.031496532758076984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1536,16384,0.11696853637695312
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,16384,0.02483733296394348
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1536,12288,0.02207466761271159
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1536,12288,0.09350826740264892
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,12288,0.027361067136128743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1536,10240,0.08290987014770508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,12288,0.017593600352605186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1536,10240,0.019236266613006592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,10240,0.02381333311398824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,10240,0.015701333681742348
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1536,8192,0.06938560009002685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,8192,0.016150400042533875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,8192,0.01369493305683136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1536,7168,0.08291412989298502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1536,7168,0.014792533715566
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,7168,0.015155200163523355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,7168,0.012813867131868998
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1536,6144,0.07522559960683187
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2048,65536,0.103329070409139
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1536,6144,0.013041067123413085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,6144,0.013204266627629599
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,6144,0.012051199873288471
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1536,5120,0.06640106836954753
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1536,5120,0.011354666948318482
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,5120,0.014793599645296732
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,5120,0.011342933773994446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1536,4096,0.054289066791534425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1536,4096,0.009998933474222819
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,4096,0.010603732864061991
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,4096,0.010021332899729412
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1536,3584,0.05077866713205973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1536,3584,0.009583999713261921
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,3584,0.009994666775067646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,3584,0.00962559978167216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2048,10240,0.020181334018707274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1536,3072,0.04595520099004109
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,3072,0.009197866916656494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,3072,0.009083732962608337
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2048,8192,0.01694933374722799
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1536,2560,0.0421887993812561
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1536,2560,0.008193066716194153
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,2560,0.00783679982026418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,2560,0.008548266688982646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1536,2048,0.0380181352297465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,2048,0.006921599805355072
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,2048,0.007923200229803721
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1536,1536,0.03461120128631592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1536,1536,0.006829866766929626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,1536,0.006119466821352641
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,1536,0.0074538667996724445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1536,1024,0.030894933144251506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,1024,0.006968533496061961
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1536,1024,0.0061941335598627726
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,1024,0.00527999997138977
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1536,768,0.029021867116292316
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1536,768,0.005721599857012431
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,768,0.004904533425966898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,768,0.006820266445477803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1536,512,0.0272053341070811
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,512,0.004233600199222564
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,512,0.006515199939409892
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1536,256,0.02564479907353719
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,256,0.0035487999518712364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,256,0.006289066871007283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1536,128,0.024617600440979003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1536,128,0.004930133124192556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,128,0.0032927999893824257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,128,0.006042666733264923
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1536,64,0.024155733982721965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,64,0.003138133386770884
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,64,0.006450133522351582
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1536,32,0.024196267127990723
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,32,0.003218133250872294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,32,0.006337066491444905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2048,768,0.005851733187834421
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1024,65536,0.4837194760640462
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,65536,0.06876586278279623
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,65536,0.08872426350911458
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1024,16384,0.11659413178761799
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1024,16384,0.02541866699854533
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,16384,0.02568746606508891
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,16384,0.021463465690612794
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1024,12288,0.09389119942982992
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1024,12288,0.01967039903004964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,12288,0.020426666736602782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,12288,0.017560533682505288
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1024,10240,0.08323626518249512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,10240,0.0193066676457723
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,10240,0.015595733126004537
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1024,8192,0.06930027008056641
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1024,8192,0.014571733276049294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,8192,0.013562666376431784
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,8192,0.013582932949066161
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1024,7168,0.08284479777018229
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1536,8192,0.01600320041179657
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,7168,0.013105066617329917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,7168,0.012821333607037863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1024,6144,0.0748650630315145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1024,6144,0.012337066729863485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,6144,0.01383573313554128
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,6144,0.011830400427182515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1024,5120,0.06582933266957601
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1024,5120,0.010717866818110149
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,5120,0.012556800246238708
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,5120,0.011069867014884948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1024,4096,0.05351786613464356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1536,3072,0.008779733379681905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,4096,0.009841066598892213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,4096,0.009802666306495667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1024,3584,0.05053226550420126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1024,3584,0.00893440047899882
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1536,2048,0.007401599983374278
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,3584,0.00921493371327718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,3584,0.009484799702962239
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1024,3072,0.045815467834472656
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,3072,0.00853013296922048
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1536,512,0.0054175997773806255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,3072,0.00895360012849172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1024,2560,0.007766399780909221
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1024,2560,0.04195839961369832
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,2560,0.008333866794904072
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,2560,0.008340266346931458
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1024,2048,0.03768639961878459
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1024,2048,0.007036800185839335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,2048,0.007375999788443248
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,2048,0.007815466821193695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1024,1536,0.034499200185139973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1024,1536,0.006545066833496094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,1536,0.006347733239332835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,1536,0.007388799885908763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1024,1024,0.030726399024327593
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,1024,0.0050570666790008545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1536,256,0.005127466718355815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,1024,0.006958933174610138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1024,768,0.028626133998235066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1024,768,0.005690666536490122
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,768,0.004428799947102865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,768,0.006701866785685222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1024,512,0.02693866689999898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1024,512,0.005321600039800008
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,512,0.0039711999396483105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,512,0.006445866823196411
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1024,256,0.025300266345342
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,256,0.0034826666116714476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1024,65536,0.09060160319010416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,256,0.006257066627343495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1024,128,0.024516266584396363
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,128,0.0032501332461833954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,128,0.006175999840100607
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1024,64,0.02424853245417277
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,64,0.0030517332255840302
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,64,0.00610453337430954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,1024,32,0.024260266621907552
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,32,0.0030933332939942675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,32,0.006221866607666016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,768,65536,0.48515199025472
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,768,65536,0.08914773464202881
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,65536,0.05645440022150675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,65536,0.08850346406300863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,768,16384,0.11704320112864178
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1024,10240,0.017035732666651406
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,16384,0.017011199394861856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,16384,0.021656533082326256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,768,12288,0.0935861349105835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,768,12288,0.01950613260269165
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,12288,0.01458453337351481
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,12288,0.017475199699401856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,768,10240,0.08321812947591146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,768,10240,0.016951467593510947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,10240,0.014189866185188294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,10240,0.015619200468063355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,768,8192,0.06944639682769775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,768,8192,0.014498133460680643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,8192,0.01232319970925649
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1024,7168,0.013451733191808066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,8192,0.01368106702963511
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,768,7168,0.08246719837188721
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,768,7168,0.01338879962762197
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,7168,0.011595732967058818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,7168,0.012744533022244773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,768,6144,0.07384533087412516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,768,6144,0.012083199620246888
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,6144,0.011569066842397054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,6144,0.011708799997965496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,768,5120,0.06553279956181844
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1024,4096,0.009429333607355754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,5120,0.010321066776911417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,5120,0.011029332876205444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,768,4096,0.053497600555419925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1024,3072,0.008261333405971526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,4096,0.008849066495895386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,4096,0.009894399841626485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,768,3584,0.04987733364105225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,3584,0.008171733220418293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,3584,0.009455999732017517
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1024,1024,0.005838933090368906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,768,3072,0.0459882656733195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,3072,0.007428266604741414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,3072,0.008905599514643352
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,768,2560,0.04227199951807658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1024,256,0.005002666513125102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,2560,0.0069909334182739254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,2560,0.00844693382581075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,768,2048,0.037780265013376876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1024,128,0.0048437332113583885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,2048,0.006165333092212677
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,2048,0.00783679982026418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,768,1536,0.03461759885152181
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,768,1536,0.0065536002318064375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,1536,0.005465599894523621
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,1536,0.007426133255163829
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,768,1024,0.030616533756256104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,768,16384,0.02448319991429647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,1024,0.004644266764322917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,1024,0.006914133330186208
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,768,768,0.028892799218495684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,768,768,0.005539200206597646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,768,0.004309333364168803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,768,0.006660266717274983
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,768,512,0.027210666735966997
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,768,512,0.005342933535575867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,512,0.003882666677236557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,512,0.006461866696675618
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,768,256,0.02537493308385213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,768,256,0.004938666522502899
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,256,0.00346666673819224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,256,0.00613013356924057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,768,128,0.024631466468175253
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,768,5120,0.010607999563217164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,128,0.0031914666295051576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,128,0.0059562668204307554
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,768,64,0.024155733982721965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,64,0.0030303999781608583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,64,0.006062933305899302
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,768,32,0.023974400758743287
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,32,0.0030410667260487873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,32,0.00613013356924057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,512,65536,0.48419628143310545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,512,65536,0.08479253451029459
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,65536,0.05132373174031576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,65536,0.082259202003479
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,512,16384,0.1162943998972575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,512,16384,0.023281067609786987
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,16384,0.013010133306185404
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,16384,0.021051732699076335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,512,12288,0.09263146718343099
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,512,12288,0.018528000513712565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,12288,0.01135040024916331
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,12288,0.017114667097727458
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,768,3072,0.008175999919573466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,512,10240,0.08218453725179037
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,10240,0.010771200060844421
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,10240,0.01532373329003652
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,768,4096,0.009405866265296936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,512,8192,0.06893760363260905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,512,8192,0.013938132921854654
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,8192,0.008888533711433411
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,8192,0.013530666629473368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,768,3584,0.008842666943868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,512,7168,0.01297813355922699
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,512,7168,0.08209066390991211
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,7168,0.009589333335558574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,7168,0.012723199526468911
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,512,6144,0.07292266686757407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,6144,0.011900800466537475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,6144,0.011602133512496948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,512,5120,0.06404053370157878
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,768,2560,0.0076341331005096436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,5120,0.00867306689421336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,5120,0.010950400431950887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,512,4096,0.052894934018452966
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,4096,0.00865600009759267
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,4096,0.009827199578285217
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,512,3584,0.05000746647516886
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,512,3584,0.008594133456548055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,3584,0.007889066636562348
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,768,2048,0.007021866738796234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,3584,0.00941439966360728
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,512,3072,0.045533867677052815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,512,3072,0.007978666822115581
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,3072,0.007894399762153625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,3072,0.008880000313123066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,512,2560,0.0418506662050883
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,512,2560,0.007521066566308339
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,2560,0.006793599824110668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,2560,0.008334933718045553
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,512,2048,0.03750826517740886
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,768,1024,0.005881600081920624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,512,2048,0.006867200136184692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,2048,0.005922133227189382
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,2048,0.007739733159542084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,512,1536,0.03428373336791992
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,512,1536,0.006353066861629486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,1536,0.005250133574008942
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,1536,0.007344000041484833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,512,1024,0.03070293267567952
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,512,1024,0.005738666653633118
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,1024,0.004491733511288961
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,1024,0.006900266806284587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,512,768,0.02858453392982483
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,768,0.0042015999555587765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,768,0.00662720004717509
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,512,512,0.026792534192403156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,512,512,0.005102933446566264
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,512,0.003881600002447764
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,512,0.006321066617965698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,512,256,0.025084799528121947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,256,0.0034954667091369627
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,256,0.006128000219662985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,512,128,0.024181334177652995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,768,128,0.004714666803677877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,128,0.0031189332405726117
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,128,0.005930666625499725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,512,64,0.023859200874964397
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,64,0.0030048000315825146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,64,0.006031999985376993
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,512,32,0.023570134242375692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,32,0.0030026666820049284
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,32,0.006056533257166544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,256,65536,0.48372907638549806
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,256,65536,0.08470613161722819
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,65536,0.03854506810506185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,65536,0.08313493728637696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,256,16384,0.11644266446431477
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,256,16384,0.02297066648801168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,16384,0.010797866185506185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,16384,0.020915200312932335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,256,12288,0.09313279787699381
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,256,12288,0.018501333395640054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,12288,0.00955839951833089
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,12288,0.017170133193333943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,256,10240,0.08257599671681723
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,256,10240,0.015961600343386333
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,10240,0.00914026697476705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,10240,0.015364266435305276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,256,8192,0.0688746690750122
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,256,8192,0.01376426617304484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,8192,0.008124800026416778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,8192,0.013382400075594584
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,256,7168,0.08221120039621989
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,512,10240,0.01606613298257192
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,7168,0.0076789334416389465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,7168,0.012506666779518127
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,256,6144,0.0732373317082723
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,512,5120,0.010328533252080281
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,6144,0.009307733178138733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,6144,0.011666133006413778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,256,5120,0.06408533255259195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,256,5120,0.01015786627928416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,5120,0.00867306689421336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,5120,0.010947199662526448
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,256,4096,0.05330986579259237
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,512,6144,0.011738666892051696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,256,4096,0.008982400099436443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,4096,0.007959466675917308
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,4096,0.009768533706665038
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,256,3584,0.04947093327840169
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,3584,0.007613866527875264
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,512,4096,0.009001599748929341
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,3584,0.009175466497739156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,256,3072,0.0455733338991801
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,3072,0.007274666428565979
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,3072,0.008780800302823384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,256,2560,0.04182933171590169
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,512,768,0.005485866467158
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,2560,0.0064522668719291685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,2560,0.008307200173536937
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,256,2048,0.03769173224767049
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,512,256,0.004860800007979075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,2048,0.005765333275000254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,2048,0.007670400043328603
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,256,1536,0.03441066741943359
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,256,1536,0.006275199850400289
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,512,128,0.004713599880536398
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,1536,0.005072000126043955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,1536,0.00724480003118515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,256,1024,0.030550400416056316
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,1024,0.004384000102678935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,1024,0.006826666494210561
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,256,768,0.028758400678634645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,256,7168,0.012770133217175803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,768,0.004010666658480962
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,768,0.006524799764156342
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,256,512,0.02685333291689555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,256,6144,0.0116074671347936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,512,0.003664000084002813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,512,0.006214400132497152
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,256,256,0.025169066588083905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,256,3584,0.008591999610265095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,256,0.003335466732581457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,256,0.006084266801675161
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,256,128,0.02431466579437256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,256,128,0.00469760000705719
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,128,0.003115733216206233
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,128,0.005937066674232483
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,256,64,0.023962666591008507
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,64,0.0030250666042168934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,64,0.005937066674232483
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,256,32,0.023781333367029825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,32,0.003101866692304611
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,32,0.006002133091290792
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,256,3072,0.007843199868996937
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,128,65536,0.4852992057800293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,256,2560,0.0073077330986658735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,65536,0.0844330628712972
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,65536,0.02661120096842448
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,128,16384,0.1164789358774821
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,16384,0.009752532839775086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,16384,0.021211733420689903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,128,12288,0.09298240343729655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,128,12288,0.018490666151046754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,12288,0.0086325337489446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,12288,0.017335466543833413
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,128,10240,0.08275413513183594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,128,10240,0.015974400440851848
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,10240,0.008019199967384339
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,10240,0.015545599659283958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,128,8192,0.06886293093363444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,256,2048,0.0066549330949783325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,8192,0.007540266712506611
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,8192,0.013430399696032205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,128,7168,0.08210026423136393
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,128,7168,0.012793599565823873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,7168,0.007295999924341838
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,7168,0.012622933586438498
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,128,6144,0.07215893268585205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,256,1024,0.0055978665749231975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,128,6144,0.011557333668073018
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,6144,0.007041066884994507
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,6144,0.011639466881752015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,128,5120,0.06329919894536337
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,5120,0.007397333284219106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,5120,0.010956799983978272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,128,4096,0.05281279881795248
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,128,4096,0.008984532952308655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,4096,0.0070592001080513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,4096,0.00979306697845459
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,128,3584,0.04935466845830282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,128,3584,0.008543999989827473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,3584,0.007446399827798207
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,3584,0.009243733684221904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,128,3072,0.045634134610493975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,128,3072,0.007842133442560833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,3072,0.007111466427644093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,3072,0.008808533350626629
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,128,2560,0.04190613428751628
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,128,2560,0.0073183998465538025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,2560,0.006321066617965698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,2560,0.008242133259773254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,128,2048,0.037808001041412354
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,128,2048,0.0067552000284194945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,2048,0.005686399837334951
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,2048,0.0076799998680750535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,128,1536,0.03455253442128499
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,128,1536,0.006235733131567637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,1536,0.004982399940490723
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,1536,0.007206400235493978
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,128,1024,0.030562132596969604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,128,1024,0.005577600002288819
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,1024,0.004307200014591217
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,256,512,0.0051018665234247845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,256,768,0.00544106662273407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,1024,0.006759466727574666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,128,768,0.028601600726445513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,768,0.003969066590070724
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,768,0.0065536002318064375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,128,512,0.02687679926554362
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,128,512,0.005027199784914652
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,512,0.0036202666660149893
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,512,0.0062613333264986675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,128,256,0.02518080075581868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,128,256,0.0047882666190465295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,256,0.0033013333876927694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,256,0.006051200131575266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,128,128,0.024289067586263022
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,128,0.0030858665704727173
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,256,256,0.004801066716512045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,128,0.005864533285299936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,128,64,0.02359893321990967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,64,0.002930133293072383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,64,0.005869866907596588
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,128,32,0.02341973384221395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,32,0.0029450667401154833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,32,0.005873066683610281
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,64,65536,0.4838869412740071
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,65536,0.020660267273585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,65536,0.0839413324991862
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,64,16384,0.11649920145670574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,16384,0.007772799829641979
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,16384,0.021031467119852702
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,64,12288,0.09297173023223877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,12288,0.007208533088366191
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,12288,0.017077332735061644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,64,10240,0.08252586523691813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,10240,0.007355733215808869
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,10240,0.015239466230074564
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,64,8192,0.06919786930084229
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,8192,0.007235200206438701
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,8192,0.013402666648228964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,64,7168,0.08152853647867839
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,7168,0.007161599894364675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,7168,0.01251200040181478
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,6144,0.006917333106199901
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,64,6144,0.07285013198852539
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,6144,0.011490133404731751
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,5120,0.007338666419188182
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,64,5120,0.06330026785532633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,5120,0.010774399836858113
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,64,4096,0.0528714656829834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,4096,0.006875733534495036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,4096,0.009715200463930766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,64,3584,0.049180801709493
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,3584,0.007403733332951863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,3584,0.009264000256856282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,3072,0.007001600166161854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,64,3072,0.04554666678110759
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,3072,0.008643200000127155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,64,2560,0.04184639851252238
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,2560,0.00624533345301946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,2560,0.008146133522192638
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,64,2048,0.03753386735916138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,2048,0.005638400216897329
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,2048,0.0074976002176602675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,1536,0.004949333270390829
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,64,1536,0.03451306819915771
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,1536,0.007136000196139018
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,64,1024,0.030494934320449828
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,1024,0.004275199770927429
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,1024,0.006720000008742015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,64,768,0.02845226724942525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,768,0.003955200066169103
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,768,0.006513066589832306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,64,512,0.026745599508285523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,512,0.0035957333942254386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,512,0.006206933160622915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,64,256,0.024835199117660522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,128,65536,0.08569066524505616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,256,0.003289599965016047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,256,0.0059232001503308615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,128,16384,0.022899200518925987
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,64,128,0.024131200710932412
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,128,0.0030432000756263735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,128,0.005861333509286245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,64,64,0.023770666122436522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,64,0.002918400118748347
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,64,0.005891199906667074
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,64,32,0.023345067103703817
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,32,0.0029397333661715193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,32,0.005874133110046387
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,65536,0.015124266346295675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,65536,0.08390826384226481
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,16384,0.0071370666225751235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,32,16384,0.11659306685129803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,16384,0.02096959948539734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,32,65536,0.486741320292155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,12288,0.007057066758473713
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,32,12288,0.09294826984405517
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,12288,0.017169066270192466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,10240,0.007275733351707459
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,32,10240,0.0826538642247518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,10240,0.015196800231933594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,32,8192,0.0689184029897054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,8192,0.007223466535409291
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,32,7168,0.08178346951802572
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,8192,0.01334826648235321
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,7168,0.007037866612275441
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,7168,0.012460800011952718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,128,8192,0.013709867000579834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,32,6144,0.07239360014597575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,6144,0.006876799960931141
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,6144,0.011594667037328085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,5120,0.007179733117421467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,32,5120,0.06253866751988729
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,5120,0.010663466652234395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,4096,0.006922666728496551
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,32,4096,0.052418132623036705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,4096,0.009437867005666097
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,32,3584,0.04928640127182007
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,3584,0.007241599758466085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,3584,0.009206400314966837
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,32,3072,0.04549013376235962
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,3072,0.0069578667481740315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,3072,0.0086325337489446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,32,2560,0.04190400044123332
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,2560,0.006285866598288219
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,2560,0.008144000172615051
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,2048,0.007594666878382365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,32,2048,0.03757866621017456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,2048,0.005564799904823304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,1536,0.007049599786599477
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,32,1536,0.034227200349171955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,32,1024,0.030514132976531983
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,1536,0.004910933474699656
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,1024,0.0042453333735466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,1024,0.006604800124963124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,32,768,0.02881493369738261
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,768,0.003885866701602936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,768,0.006484266618887584
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,32,512,0.026679466168085735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,512,0.003570133447647095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,512,0.006181333462397257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,32,256,0.02495253284772237
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,128,5120,0.01011306643486023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,256,0.003269333392381668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,256,0.006022400160630544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,32,128,0.024140799045562746
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,128,0.0030378667016824085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,128,0.00589333325624466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,32,64,0.023800534009933472
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,64,0.0028543998797734577
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,64,0.005760000149408976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,256,32,32,0.02335466742515564
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,32,0.0028736000259717304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,32,0.00581333339214325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,65536,16384,0.5008810679117839
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,65536,16384,0.6899050394694011
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,65536,12288,0.3832277297973633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,65536,16384,0.834937604268392
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,65536,16384,0.5809696197509766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,65536,12288,0.4477173487345378
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,65536,12288,0.6348341623942058
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,65536,12288,0.5303413391113281
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,128,128,0.0046079998215039575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,65536,10240,0.4409482638041179
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,65536,10240,0.38508478800455725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,65536,10240,0.5350624084472656
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,65536,8192,0.27197653452555337
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,128,768,0.005347200234731038
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,65536,8192,0.31319573720296223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,65536,8192,0.39761492411295574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,65536,8192,0.33983678817749025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,65536,7168,0.23600319226582847
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,65536,7168,0.2852277437845866
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,65536,7168,0.3506837209065755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,65536,7168,0.29805545806884765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,65536,6144,0.20592320760091148
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,65536,6144,0.3065429369608561
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,65536,6144,0.2579573313395182
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,65536,5120,0.16900374094645182
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,65536,6144,0.24665066401163735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,65536,5120,0.18526612917582194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,65536,5120,0.20273812611897787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,65536,5120,0.25837546984354653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,65536,4096,0.13971519470214844
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,65536,4096,0.17768534024556476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,65536,4096,0.21033387184143065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,65536,4096,0.162447992960612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,65536,3584,0.18689173062642414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,65536,3584,0.11798826853434245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,65536,3584,0.1484384059906006
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,65536,3072,0.10206293265024821
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,65536,3072,0.1308725357055664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,65536,3072,0.16406933466593426
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,65536,3072,0.10762666861216227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,65536,2560,0.11904213428497315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,65536,2560,0.14001280466715496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,65536,2560,0.0940064032872518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,65536,2048,0.07121813297271729
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,65536,2048,0.10222506523132324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,65536,2048,0.07016426722208659
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,65536,2048,0.11604800224304199
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,65536,1536,0.0547541340192159
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,65536,1536,0.05588693221410116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,65536,1536,0.0843221346537272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,65536,1536,0.09172373612721761
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,65536,1024,0.0664469321568807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,65536,1024,0.0406709353129069
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,65536,1024,0.06617600123087565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,65536,1024,0.04471146663029989
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,65536,768,0.05787200133005778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,65536,768,0.034593065579732254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,65536,512,0.05219306548436483
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,65536,768,0.05329279899597168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,65536,768,0.04048320055007935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,65536,512,0.028100266059239702
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,65536,512,0.036562132835388186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,65536,512,0.04546346664428711
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,65536,256,0.046401067574818926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,65536,256,0.019962666432062785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,65536,256,0.025199999411900837
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,65536,256,0.03241066733996074
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,65536,128,0.040217598279317215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,65536,128,0.014145066340764364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,65536,128,0.01790720025698344
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,65536,128,0.02884053389231364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,65536,64,0.039561601479848225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,65536,64,0.01332266628742218
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,65536,64,0.02899199922879537
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,65536,32,0.04445226589838664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,65536,32,0.014450132846832275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,65536,32,0.03778453270594279
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,16384,65536,0.5141983985900879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,65536,0.7678026835123698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,65536,0.8124725341796875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,16384,16384,0.1932085355122884
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,16384,16384,0.13622612953186036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,16384,65536,0.7708650588989258
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,16384,0.1954975922902425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,16384,0.1457856019337972
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,12288,0.11882560253143311
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,16384,12288,0.15427093505859374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,12288,0.15933866500854493
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,16384,12288,0.10500160058339436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,65536,10240,0.32536001205444337
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,16384,10240,0.08727040290832519
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,16384,10240,0.13000426292419434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,10240,0.13442346254984539
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,10240,0.10819199879964192
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,16384,8192,0.0705141305923462
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,8192,0.10141013463338215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,16384,8192,0.10981866518656414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,8192,0.07470719814300537
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,16384,7168,0.0633898655573527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,7168,0.06179733276367187
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,16384,7168,0.11890986760457356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,7168,0.08996373017628988
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,16384,6144,0.05594346523284912
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,16384,6144,0.10679999987284343
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,6144,0.07926719983418783
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,6144,0.05801920096079508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,16384,5120,0.09020480314890543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,5120,0.06722453435262045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,5120,0.0452128012975057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,16384,4096,0.07758506933848063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,4096,0.05709013144175211
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,16384,3584,0.06971306800842285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,4096,0.036957867940266925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,16384,3584,0.03473386764526367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,3584,0.033802668253580734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,3584,0.05429333448410034
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,16384,3072,0.06339733203252157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,3072,0.04349013169606526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,16384,3072,0.030195200443267824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,3072,0.030044800043106078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,16384,2560,0.05824426809946696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,2560,0.03859093189239502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,2560,0.026936533053716023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,65536,3584,0.12073386510213216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,16384,2048,0.022252800067265828
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,16384,2048,0.051402668158213295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,2048,0.030980267127354938
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,2048,0.023143466313680014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,16384,1536,0.015500799814860026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,16384,1536,0.045021867752075194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,1536,0.01708266735076904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,65536,2560,0.0890005350112915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,1536,0.027955200274785357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,16384,1024,0.0367850661277771
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,16384,1024,0.0123989333709081
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,1024,0.012991999586423239
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,1024,0.01388159990310669
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,16384,768,0.03378239870071411
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,16384,768,0.010712533195813497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,768,0.01066986620426178
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,512,0.008553600311279297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,768,0.012634666760762534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,16384,512,0.030598400036493938
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,16384,256,0.02789013385772705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,16384,512,0.009426132837931315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,512,0.011718400319417318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,256,0.006613333523273468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,256,0.01074666678905487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,128,0.005662933488686879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,128,0.010024533669153849
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,16384,128,0.026498132944107057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,16384,64,0.026313600937525432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,16384,128,0.007534933090209961
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,64,0.004926933348178864
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,64,0.010759466886520385
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,16384,32,0.026718932390213012
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,32,0.005347200234731038
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,32,0.013554132978121438
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,65536,0.5640629450480144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,65536,0.6106037139892578
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,12288,65536,0.6944693247477214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,12288,16384,0.09802666505177816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,12288,16384,0.17434666951497396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,16384,0.16278932889302572
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,16384,0.12019306818644207
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,12288,12288,0.1395359992980957
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,12288,0.11356266339619954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,12288,0.09082880020141601
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,12288,10240,0.11757973035176594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,12288,10240,0.05966826677322388
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,10240,0.09578239917755127
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,10240,0.08165333271026612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,12288,8192,0.09860373338063558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,16384,5120,0.046507732073465986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,8192,0.07773760159810385
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,16384,4096,0.03836693366368611
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,8192,0.0605674664179484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,12288,7168,0.0445685346921285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,12288,7168,0.10967679818471272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,7168,0.0683573325475057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,7168,0.049133865038553874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,12288,6144,0.09795520305633545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,6144,0.06073280175526937
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,16384,2560,0.026608000199000042
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,6144,0.04560533364613851
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,12288,5120,0.034433066844940186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,12288,5120,0.0831274668375651
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,5120,0.05178346633911133
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,5120,0.03761920134226481
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,12288,4096,0.028191999594370527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,12288,4096,0.0713088035583496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,4096,0.042662398020426436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,4096,0.03194560011227925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,12288,3584,0.06601493358612061
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,3584,0.038091735045115156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,3584,0.029105067253112793
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,12288,3072,0.023257599274317423
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,12288,3072,0.06100266774495443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,3072,0.033524266878763836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,3072,0.02568853298823039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,12288,2560,0.05467520157496134
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,2560,0.02874773343404134
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,2560,0.022247467438379923
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,12288,2048,0.04679359992345174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,16384,256,0.008232533435026805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,12288,2048,0.01442026694615682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,2048,0.024016000827153525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,2048,0.01753173271814982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,12288,1536,0.040481066703796385
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,12288,1536,0.012363732854525248
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,1536,0.01790293256441752
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,1536,0.014499200383822122
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,12288,1024,0.035336534182230636
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,12288,1024,0.010437333583831787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,1024,0.00969599982102712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,1024,0.012401066223780314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,12288,768,0.032254934310913086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,12288,768,0.009223467111587525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,768,0.008210133512814839
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,768,0.011530666550000509
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,12288,512,0.030189865827560426
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,12288,512,0.008335999647776286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,512,0.006859733164310456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,512,0.010502400000890096
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,12288,256,0.02759786645571391
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,12288,256,0.00749120016892751
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,256,0.005545599758625031
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,256,0.009691733121871948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,12288,128,0.026370133956273394
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,12288,128,0.006810666620731353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,128,0.004790399968624115
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,128,0.009156266848246258
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,12288,64,0.026241066058476765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,64,0.004379733403523763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,64,0.00944640040397644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,12288,65536,0.34097601572672526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,12288,32,0.026866134007771807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,32,0.004671999812126159
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,32,0.011632000406583149
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,65536,0.5049536069234212
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,12288,12288,0.070796799659729
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,10240,65536,0.6630954742431641
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,10240,16384,0.16841386159261068
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,10240,16384,0.08743893305460612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,65536,0.4530943870544434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,16384,0.13871893882751465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,16384,0.09401600360870362
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,10240,12288,0.06401066780090332
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,10240,12288,0.13351252873738606
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,12288,0.10059093634287516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,12288,0.07250560124715169
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,10240,10240,0.05436906814575195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,10240,10240,0.11315840085347492
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,10240,0.0886741320292155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,10240,0.06636586586634317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,10240,8192,0.044513066609700516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,10240,8192,0.09511040051778158
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,8192,0.07177066802978516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,8192,0.04742613236109416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,12288,8192,0.04932693243026733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,10240,7168,0.10708906650543212
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,7168,0.060444800059000645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,7168,0.04193600018819173
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,10240,6144,0.09625600179036459
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,10240,6144,0.03534293174743652
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,6144,0.05388693412144979
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,6144,0.03710506757100423
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,10240,5120,0.08241066932678223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,10240,5120,0.03044053316116333
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,5120,0.04702186584472656
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,5120,0.0324234664440155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,10240,4096,0.02579413255055745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,10240,4096,0.07095146973927816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,4096,0.04071040153503418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,4096,0.026946133375167845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,10240,3584,0.0651253342628479
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,3584,0.03521920045216878
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,12288,6144,0.03938239812850952
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,3584,0.024409600098927817
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,10240,3072,0.05909013350804647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,10240,3072,0.01835093299547831
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,3072,0.029689600070317585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,3072,0.02090453306833903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,12288,3584,0.02651626666386922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,10240,2560,0.05149226586023966
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,2560,0.026021333535512288
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,2560,0.01911146640777588
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,10240,2048,0.04560106595357259
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,2048,0.02155839999516805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,2048,0.01525973379611969
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,10240,1536,0.0396448016166687
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,10240,1536,0.010858666896820069
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,1536,0.01680319905281067
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,1536,0.012499200304349263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,12288,2560,0.018518400192260743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,10240,1024,0.034392531712849936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,10240,1024,0.009398399790128072
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,1024,0.012223999698956807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,1024,0.010988799730936687
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,10240,768,0.031959466139475506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,768,0.010179199775060018
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,768,0.01032960017522176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,10240,512,0.029785599311192828
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,10240,512,0.0077898666262626644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,512,0.006613333523273468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,512,0.009514666597048442
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,10240,256,0.02733760078748067
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,10240,256,0.006940799951553345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,256,0.005831466615200042
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,256,0.008964266379674275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,10240,128,0.02617599964141846
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,10240,128,0.006455466647942861
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,128,0.004686933259169261
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,128,0.008519466718037922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,10240,64,0.026016000906626386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,64,0.004193066557248434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,64,0.008929066856702169
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,10240,32,0.026048000653584796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,32,0.004643199841181437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,32,0.010131200154622395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,8192,65536,0.236407470703125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,8192,65536,0.5076352119445801
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,65536,0.4298399925231934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,8192,16384,0.13022293249766032
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,65536,0.3511029243469238
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,8192,16384,0.06645973523457846
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,16384,0.1072533369064331
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,16384,0.07267413139343262
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,10240,65536,0.31283305486043295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,8192,12288,0.10559146404266358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,8192,12288,0.05247146685918173
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,12288,0.07984106540679932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,12288,0.05950080156326294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,8192,10240,0.09103039900461832
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,10240,0.069268266359965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,10240,0.051438931624094644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,8192,8192,0.07675733566284179
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,8192,8192,0.03719040155410767
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,8192,0.054296533266703285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,8192,0.037829331556955975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,8192,7168,0.08991146882375081
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,7168,0.04841493368148804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,7168,0.033411200841267905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,8192,6144,0.08192000389099122
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,10240,7168,0.04000106652577718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,6144,0.04381866852442424
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,6144,0.03024959961573283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,8192,5120,0.07090240319569906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,8192,5120,0.025987199942270917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,5120,0.03716053167978923
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,5120,0.026636799176534016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,8192,4096,0.061755732695261634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,10240,3584,0.022395733992258707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,8192,4096,0.018718934059143065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,4096,0.029501867294311524
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,4096,0.02097066640853882
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,8192,3584,0.016731733083724977
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,8192,3584,0.05667093197504679
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,3584,0.026317866643269856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,3584,0.019292799631754558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,10240,2560,0.015424000223477683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,8192,3072,0.05170666774113973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,3072,0.023436800638834635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,3072,0.017490132649739584
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,8192,2560,0.04477333227793376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,8192,2560,0.01341973344484965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,10240,2048,0.012993066509564718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,2560,0.018768000602722167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,2560,0.014998400211334228
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,8192,2048,0.04036800066630046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,8192,2048,0.011853866775830587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,2048,0.011632000406583149
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,2048,0.012460800011952718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,8192,1536,0.03607146739959717
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,1536,0.009117866555849712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,1536,0.010899200042088827
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,8192,1024,0.032202666997909545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,10240,768,0.008503466844558716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,1024,0.0073749333620071415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,1024,0.009636267026265462
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,8192,768,0.029601067304611206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,768,0.006502399841944377
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,768,0.009117866555849712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,8192,512,0.02844160000483195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,8192,512,0.006921599805355072
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,512,0.005534933507442474
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,512,0.00862613320350647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,8192,256,0.026316799720128375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,8192,256,0.006205866734186808
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,256,0.004660266637802124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,256,0.008067200084527333
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,8192,128,0.024943999449412026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,8192,128,0.005740800003210703
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,128,0.004155733436346054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,128,0.007694933315118153
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,8192,64,0.024875734249750772
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,64,0.003751466671625773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,64,0.00811413327852885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,8192,10240,0.0451530655225118
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,8192,32,0.025077333052953083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,32,0.004159999887148539
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,32,0.00906986693541209
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,8192,7168,0.03389439980189006
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,7168,65536,0.31562773386637366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,8192,6144,0.0302784005800883
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,7168,65536,0.4827157338460286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,7168,16384,0.12488319873809814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,65536,0.3659306526184082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,65536,0.34808851877848307
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,16384,0.1059381326039632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,16384,0.07385599613189697
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,7168,12288,0.10208213329315186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,12288,0.06983679930369059
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,12288,0.05538133382797241
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,7168,10240,0.04605226516723633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,7168,10240,0.08703893025716146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,10240,0.05883520046869913
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,10240,0.04908906618754069
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,7168,8192,0.038199468453725176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,7168,8192,0.07522133191426596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,8192,0.04730879863103231
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,8192,0.03723733425140381
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,7168,7168,0.03374613523483276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,7168,7168,0.08821226755777994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,8192,3072,0.014735999703407287
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,7168,0.04230826695760091
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,7168,0.032459733883539836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,7168,6144,0.0807808001836141
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,7168,6144,0.02959360082944234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,6144,0.03776106834411621
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,6144,0.029346134265263873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,7168,5120,0.06919253667195638
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,7168,5120,0.02448319991429647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,5120,0.03178880016009013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,5120,0.024546132485071818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,7168,4096,0.059689601262410484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,4096,0.026497066020965576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,4096,0.02017813324928284
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,8192,1536,0.010203733046849569
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,7168,3584,0.055384532610575354
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,7168,3584,0.018308266003926595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,3584,0.023963733514149984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,3584,0.01920959949493408
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,7168,3072,0.050443732738494874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,8192,1024,0.008637866377830506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,3072,0.021458133061726888
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,3072,0.01665386656920115
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,7168,2560,0.043724799156188966
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,7168,2560,0.014392532904942832
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,2560,0.020987733205159505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,2560,0.014532267053922018
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,7168,2048,0.04033493200937907
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,7168,2048,0.012053333719571431
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,8192,768,0.007740800082683563
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,2048,0.016389333208402
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,2048,0.012607999642690024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,7168,1536,0.0352512001991272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,1536,0.013565867145856222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,1536,0.011030399799346923
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,7168,1024,0.03210026621818542
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,7168,1024,0.009199999769528707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,1024,0.008822400371233623
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,1024,0.009717333316802978
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,7168,768,0.030075732866923017
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,7168,768,0.007904000083605449
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,768,0.007853866616884867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,768,0.009086933732032777
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,7168,512,0.027970133225123088
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,7168,512,0.007359999914964039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,512,0.005881600081920624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,512,0.008407466610272725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,7168,256,0.025931733846664428
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,7168,256,0.006495999793211619
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,256,0.004808533191680908
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,256,0.007915733257929485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,7168,128,0.02481493353843689
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,128,0.004347733159859975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,128,0.007593599955240886
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,7168,64,0.024654932816823325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,64,0.004003199934959412
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,64,0.007850666840871174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,7168,32,0.02445546587308248
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,32,0.004273066421349844
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,32,0.008733866612116496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,6144,65536,0.4687616030375163
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,6144,65536,0.18871572812398274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,65536,0.31549228032430016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,65536,0.30906346638997395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,6144,16384,0.12066880067189534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,16384,0.08769493103027344
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,16384,0.06380693515141805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,7168,12288,0.055003734429677334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,7168,16384,0.07618880271911621
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,6144,12288,0.09772266546885172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,12288,0.05998506546020508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,12288,0.050327467918396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,6144,10240,0.03530240058898926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,6144,10240,0.08442560036977133
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,10240,0.05182400147120157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,10240,0.04765866597493489
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,6144,8192,0.07279253005981445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,8192,0.04168959856033325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,8192,0.03515626589457194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,6144,7168,0.08698240121205648
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,6144,7168,0.02639893293380737
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,7168,0.03716906706492106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,7168,0.03231146732966105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,6144,6144,0.07889813582102458
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,6144,6144,0.023102933168411256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,6144,0.03523199955622355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,7168,4096,0.021149865786234536
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,6144,0.027973333994547527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,6144,5120,0.06722026666005453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,6144,5120,0.019285333156585694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,5120,0.028641066948572796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,5120,0.022925867636998495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,7168,3072,0.015879467129707336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,6144,4096,0.058310401439666745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,6144,4096,0.0160778671503067
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,4096,0.01888426740964254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,4096,0.02342826724052429
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,6144,3584,0.05362026691436768
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,3584,0.02087679902712504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,3584,0.01803093353907267
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,6144,3072,0.049004801114400226
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,6144,3072,0.013705600301424661
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,3072,0.021067732572555543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,3072,0.01594986617565155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,6144,2560,0.04253653287887573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,6144,2560,0.012202666203180949
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,7168,1536,0.010477866729100544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,2560,0.017198934157689413
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,2560,0.014446933070818582
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,6144,2048,0.038464001814524335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,2048,0.013912533720334372
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,2048,0.01195733348528544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,6144,1536,0.03563626607259114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,1536,0.01197760005791982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,1536,0.010949333508809406
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,6144,1024,0.03208319942156474
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,6144,1024,0.008736000458399455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,1024,0.008210133512814839
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,1024,0.009716266393661499
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,6144,768,0.029996800422668456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,7168,128,0.006118399898211161
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,6144,768,0.0074442664782206226
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,768,0.006613333523273468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,768,0.009144533673922222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,6144,512,0.027824000517527266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,6144,512,0.006750933329264323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,512,0.005333333214124044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,512,0.008377599716186523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,6144,256,0.02574933369954427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,6144,256,0.006061866879463196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,256,0.0044383997718493145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,256,0.007926400005817413
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,6144,128,0.024740266799926757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,6144,128,0.005666133264700572
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,128,0.003945599993069967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,128,0.007648000121116638
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,6144,64,0.024102399746576943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,64,0.0035818666219711304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,64,0.007766399780909221
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,6144,32,0.024234666426976522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,32,0.0038293334345022834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,32,0.00860479970773061
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,5120,65536,0.4542442639668782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,5120,65536,0.27359040578206384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,6144,16384,0.05134613513946533
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,65536,0.23801066080729166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,65536,0.25839254061381023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,6144,12288,0.040299733479817704
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,5120,16384,0.06648106575012207
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,5120,16384,0.11545813083648682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,16384,0.07407893339792887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,16384,0.05286080042521159
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,5120,12288,0.09492373466491699
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,5120,12288,0.050330666700998936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,12288,0.039325865109761556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,12288,0.056967465082804356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,5120,10240,0.08177279631296794
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,10240,0.04891626834869385
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,10240,0.03594880104064942
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,5120,8192,0.07133546670277914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,5120,8192,0.03519359827041626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,6144,8192,0.029217066367467244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,8192,0.03564586639404297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,8192,0.02704426646232605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,5120,7168,0.08564266363779703
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,7168,0.03493119875590007
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,7168,0.024201599756876628
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,5120,6144,0.07730666796366373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,5120,6144,0.026324266195297243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,6144,0.030586665868759154
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,6144,0.02107093334197998
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,5120,5120,0.06534293492635092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,5120,0.024785067637761435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,5120,0.01753173271814982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,5120,4096,0.05611413319905599
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,5120,4096,0.018247467279434205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,4096,0.02173973321914673
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,4096,0.015019733707110086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,5120,3584,0.0513322671254476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,5120,3584,0.01643946667512258
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,3584,0.017691733439763387
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,3584,0.0141184002161026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,5120,3072,0.047577599684397384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,5120,3072,0.014294399817784628
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,3072,0.014475733041763306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,3072,0.012940800189971924
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,5120,2560,0.041621331373850504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,5120,2560,0.012809600432713827
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,2560,0.014453333616256715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,2560,0.011622400085131327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,5120,2048,0.03806080023447673
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,5120,2048,0.010872532924016316
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,6144,3584,0.014759467045466105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,2048,0.00920960009098053
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,2048,0.010121599833170573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,5120,1536,0.03456319967905681
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,5120,1536,0.009305600325266521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,1536,0.009992532928784688
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,1536,0.009351467092831928
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,5120,1024,0.03160213430722554
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,5120,1024,0.00809279978275299
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,1024,0.007727999985218048
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,1024,0.008377599716186523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,5120,768,0.029525333642959596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,6144,2048,0.011096533139546711
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,768,0.0065536002318064375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,768,0.007931733131408691
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,5120,512,0.027777065833409626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,512,0.0053951998551686605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,6144,1536,0.00959999958674113
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,512,0.007490133245786031
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,5120,256,0.025751467545827227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,256,0.004156800111134847
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,256,0.007076266904671986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,5120,128,0.024733867247899374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,5120,128,0.005688533186912537
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,128,0.003833599885304769
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,128,0.00681279997030894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,5120,64,0.02411306699117025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,64,0.003621333340803782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,64,0.0071168000499407455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,5120,32,0.02453119953473409
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,32,0.0036938667297363283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,32,0.007796266674995422
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,4096,65536,0.1633301258087158
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,4096,65536,0.44202346801757814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,65536,0.21295466423034667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,4096,16384,0.10902933279673259
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,65536,0.20817386309305824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,4096,16384,0.04620693524678548
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,16384,0.05615679820378622
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,16384,0.046266667048136395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,4096,12288,0.03679680029551188
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,4096,12288,0.09177066485087076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,12288,0.04891413450241089
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,12288,0.03715840180714925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,4096,10240,0.07834773063659668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,10240,0.04891733328501384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,5120,10240,0.04348586797714234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,10240,0.03479466835657756
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,4096,8192,0.06943679650624593
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,5120,7168,0.03049386739730835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,8192,0.030013867219289142
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,8192,0.023910399278004965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,5120,5120,0.022629332542419434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,4096,7168,0.01835413376490275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,4096,7168,0.08258773485819498
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,7168,0.021258666117986044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,7168,0.034237865606943765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,4096,6144,0.01607146660486857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,4096,6144,0.07438293298085531
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,6144,0.02722773353258769
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,6144,0.019617066780726115
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,4096,5120,0.06340693235397339
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,5120,0.023779199520746867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,5120,0.016807466745376587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,4096,4096,0.053795198599497475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,4096,0.017066667477289833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,4096,0.014298666516939798
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,4096,3584,0.04998933474222819
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,4096,3584,0.010999466975529988
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,3584,0.015346133708953857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,3584,0.013662933309872945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,4096,3072,0.046683732668558756
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,4096,3072,0.010174933075904845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,3072,0.01390506625175476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,3072,0.012497066458066305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,5120,768,0.007380266487598419
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,4096,2560,0.04144320090611776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,2560,0.012238933642705282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,2560,0.011801600456237793
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,4096,2048,0.03762986660003662
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,5120,512,0.006662400066852569
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,4096,2048,0.008526933193206788
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,5120,256,0.005962666869163513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,2048,0.009841066598892213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,2048,0.010099200407663982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,4096,1536,0.03483413457870484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,4096,1536,0.00761599987745285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,1536,0.008220799763997396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,1536,0.00929813285668691
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,4096,1024,0.03145066698392232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,4096,1024,0.006725333134333293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,1024,0.006435200075308482
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,1024,0.008481066425641377
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,4096,768,0.02940160036087036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,768,0.005605333546797434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,768,0.008044800162315369
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,4096,512,0.027485867341359455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,4096,512,0.005840000013510386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,512,0.0047765334447224935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,512,0.007462400197982788
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,4096,256,0.025498666365941364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,4096,256,0.005425066749254862
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,256,0.004010666658480962
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,256,0.00701333334048589
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,4096,128,0.024537599086761473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,4096,128,0.005098666747411092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,128,0.0036447999378045404
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,128,0.006828799843788147
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,4096,64,0.0238154669602712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,64,0.0033802665770053864
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,64,0.007066666583220164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,4096,32,0.024295467138290405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,32,0.0036501333117485045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,32,0.007447466750939687
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3584,65536,0.157370662689209
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,4096,10240,0.031325866778691605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3584,65536,0.4244223912556966
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,65536,0.18745706876118978
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,65536,0.19972373644510905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3584,16384,0.044750932852427164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3584,16384,0.10817386309305828
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,16384,0.08018026351928711
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,16384,0.04482560157775879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3584,12288,0.034883201122283936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3584,12288,0.08987946510314941
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,12288,0.03930026690165202
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,12288,0.036673065026601157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3584,10240,0.029832533995310467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3584,10240,0.07716266314188638
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,10240,0.033445334434509276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,10240,0.0319925328095754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3584,8192,0.06715520222981772
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,8192,0.022648533185323082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,8192,0.036773331960042316
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3584,7168,0.01834239959716797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3584,7168,0.08109333515167236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,7168,0.03519999980926514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,7168,0.020662399133046468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,4096,8192,0.02333973248799642
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3584,6144,0.07283946673075357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3584,6144,0.01644373337427775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,6144,0.0299615999062856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,6144,0.01932906707127889
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3584,5120,0.06251733303070069
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3584,5120,0.013877333203951518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,5120,0.025617067019144697
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,5120,0.016552533706029257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3584,4096,0.054103465874989834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3584,4096,0.011654399832089742
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,4096,0.01734293301900228
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,4096,0.01434346636136373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3584,3584,0.04907733201980591
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3584,3584,0.010921600461006164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,3584,0.014429866274197897
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,3584,0.013313066959381104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3584,3072,0.04594240188598633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3584,3072,0.010318932930628459
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,3072,0.013502933581670127
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,3072,0.012642133235931396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,4096,5120,0.013817600409189858
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3584,2560,0.041278934478759764
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3584,2560,0.009337600072224934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,2560,0.01295360028743744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,2560,0.011191466450691223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3584,2048,0.03775466680526733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3584,2048,0.008424533406893413
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,4096,4096,0.011714133620262145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,2048,0.009993599851926167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3584,1536,0.034485332171122235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,2048,0.010897066195805867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,1536,0.008513066172599792
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,1536,0.009231999516487122
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3584,1024,0.031126399834950764
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3584,1024,0.00664106657107671
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,1024,0.006675200164318084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,1024,0.008380800485610962
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3584,768,0.02921813329060872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3584,768,0.006138666470845541
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,768,0.0056618665655454
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,768,0.007934933404127757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3584,512,0.027540266513824463
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3584,512,0.0056991999348004665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,512,0.004770133395989736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,512,0.007355733215808869
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3584,256,0.025521065791447955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3584,256,0.005314133564631144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,256,0.004080000023047129
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,256,0.007068799932797749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,4096,2560,0.009425066908200582
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3584,128,0.024451200167338052
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,128,0.003673599908749262
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,128,0.006849066913127899
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3584,64,0.024387200673421226
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,64,0.003470933437347412
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,64,0.006937600175539653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3584,32,0.0245578666528066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,32,0.003685333331425985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,32,0.0072736000021298725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3072,65536,0.23576960563659669
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,65536,0.16079786618550618
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,65536,0.17474667231241864
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3072,16384,0.10517973105112713
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3072,65536,0.4152223904927571
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,16384,0.046906665960947676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,16384,0.04249386787414551
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,4096,768,0.006367999811967213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3072,12288,0.04416000048319499
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3072,12288,0.08835519949595133
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,12288,0.03496106863021851
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,12288,0.03355840047200521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3072,10240,0.07539199988047282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3072,10240,0.03734399875005086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,10240,0.04029653469721477
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,10240,0.029100799560546876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3072,8192,0.02889066735903422
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3072,8192,0.06604586839675904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,8192,0.032842665910720825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,8192,0.02127359906832377
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3072,7168,0.07989546457926432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3072,7168,0.025331199169158936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,7168,0.024220800399780272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,7168,0.01997440059979757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3072,6144,0.0721770683924357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,6144,0.018013866742451985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,6144,0.019364267587661743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3072,5120,0.061341865857442224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,5120,0.02142826716105143
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,5120,0.015752533078193666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3072,4096,0.05367146730422974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,4096,0.01493119994799296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,4096,0.013928533593813578
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3072,3584,0.04890133142471313
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3072,3584,0.014315733313560485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,3584,0.014267733693122864
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,3584,0.012882133324941
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3072,3072,0.04527146816253662
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3072,3072,0.012661332885424295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,3072,0.014083199699719749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,3072,0.012004266182581585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3072,2560,0.04135359923044841
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3072,2560,0.011396267016728719
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,2560,0.011187199751536052
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,2560,0.010957866907119751
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3584,8192,0.021341866254806517
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3072,2048,0.03749653498331706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,2048,0.00921066701412201
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3584,1536,0.00767146646976471
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,2048,0.009786666433016459
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3072,1536,0.03447680075963338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,1536,0.008040533463160197
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,1536,0.008975999553998311
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3072,1024,0.031274666388829545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3072,1024,0.007107200225194295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3584,128,0.005116799970467886
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,1024,0.006487466891606649
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,1024,0.008182399968306223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3072,768,0.029207466046015422
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,768,0.0052928000688552855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,768,0.00784746656815211
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3072,512,0.027525333563486735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3072,512,0.006000000238418579
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,512,0.004682666560014089
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,512,0.007334400216738384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3072,256,0.025525333484013875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3072,256,0.005566933254400889
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,256,0.0039989332358042395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,256,0.006995200117429097
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3072,128,0.024452267090479533
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3072,128,0.005286400020122528
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,128,0.003671466559171677
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3072,16384,0.05647679964701334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,128,0.006716800232728322
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,64,0.003433600068092346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3072,64,0.023814400037129722
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,64,0.00695360004901886
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,3072,32,0.023770666122436522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,32,0.003602133442958196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,32,0.0072170664866765336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2560,65536,0.40901654561360673
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,65536,0.14137280782063802
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3072,6144,0.022826667626698813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3072,4096,0.01594986617565155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,65536,0.12024213473002117
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3072,5120,0.01911679903666178
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2560,16384,0.1034336010615031
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,16384,0.04347413380940755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,16384,0.029366399844487506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2560,12288,0.0225983997186025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2560,12288,0.08621653715769449
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,12288,0.023069866498311362
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,12288,0.03309546709060669
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2560,10240,0.01957119901974996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2560,10240,0.07413439750671387
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,10240,0.028649600346883138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,10240,0.0190175990263621
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2560,8192,0.01592639982700348
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2560,8192,0.06479039986928305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,8192,0.013865600029627482
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,8192,0.02055786649386088
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2560,7168,0.07914559841156006
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,7168,0.017123200496037803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,7168,0.014197333653767904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2560,6144,0.07088320255279541
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,6144,0.015561599532763162
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3072,2048,0.009683199723561605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,6144,0.013829333583513894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2560,5120,0.06134293476740519
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2560,5120,0.011430399616559346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,5120,0.01635840038458506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,5120,0.011817600329717
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2560,4096,0.052978134155273436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2560,4096,0.010029866298039754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,4096,0.013017599781354269
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,4096,0.010419199864069622
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2560,3584,0.04786239862442017
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2560,3584,0.009709866841634114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,3584,0.012035199999809265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,3584,0.009896533687909444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2560,3072,0.04458986520767212
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2560,3072,0.008961066603660583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,3072,0.010857599973678588
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,3072,0.009250133236249288
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2560,2560,0.040982401371002196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2560,2560,0.008183466891447704
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,2560,0.007654400169849395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,2560,0.008719999591509502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2560,2048,0.0370197335879008
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2560,2048,0.007464533547560374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,2048,0.00811413327852885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,2048,0.008063999811808269
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2560,1536,0.034074668089548746
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2560,1536,0.006894933183987935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,1536,0.0069482664267222095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,1536,0.007608533402283986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2560,1024,0.031093333164850873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2560,1024,0.006137600044409434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,1024,0.005602133274078369
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,1024,0.007113599777221679
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2560,768,0.02911253372828166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2560,768,0.0057781333724657696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,768,0.004538666705290476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,768,0.006931200126806895
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2560,512,0.027427200476328534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2560,512,0.005588266750176748
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,512,0.004232533276081085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,512,0.00664106657107671
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2560,256,0.025333333015441894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3072,1536,0.008400000135103862
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3072,768,0.006493866443634033
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,256,0.003818666686614355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,256,0.006325333317120869
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2560,128,0.02437439958254496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,128,0.0034773332377274835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,128,0.0060810665289560955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2560,64,0.024065067370732628
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,64,0.00325546662012736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,64,0.006322133541107178
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2560,32,0.024127999941507973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,32,0.0033151999115943907
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,32,0.00666133314371109
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2048,65536,0.4093781471252441
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,65536,0.11818346977233887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,65536,0.09751466910044351
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2560,65536,0.10072426795959473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2048,16384,0.10301120281219482
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2048,16384,0.028065067529678345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,16384,0.041458133856455484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,16384,0.026970666646957398
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2048,12288,0.08533013661702474
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,12288,0.032636799414952594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2560,16384,0.029024000962575274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,12288,0.018232532342274985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2048,10240,0.018920532862345376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2048,10240,0.07344319820404052
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,10240,0.02964266737302144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,10240,0.016673066218694053
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2048,8192,0.06411093473434448
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,8192,0.018439465761184694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,8192,0.01365013321240743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2048,7168,0.0779093345006307
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2048,7168,0.014244266351064048
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,7168,0.01765120029449463
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,7168,0.01288426617781321
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2048,6144,0.07021973133087159
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2048,6144,0.012542933225631714
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,6144,0.014965333541234336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,6144,0.011989333232243856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2048,5120,0.06082559823989868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2048,5120,0.011300266782442728
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,5120,0.017258665959040322
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2560,7168,0.014391466975212097
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,5120,0.01111466685930888
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2048,4096,0.05140693187713623
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2048,4096,0.009831466277440389
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,4096,0.009883733590443929
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,4096,0.01399786671002706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2048,3584,0.04882239898045858
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,3584,0.011969066659609477
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2560,6144,0.012810666362444559
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,3584,0.009453866879145305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2048,3072,0.008717866738637288
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2048,3072,0.04449386596679687
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,3072,0.01146346628665924
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,3072,0.009014399846394856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2048,2560,0.04082026481628418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2048,2560,0.008056533336639405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,2560,0.009523199995358785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,2560,0.00853653351465861
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2048,2048,0.037240533034006755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2048,2048,0.007300266623497009
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,2048,0.008241066833337148
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,2048,0.007928533355395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2048,1536,0.03425279855728149
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,1536,0.006965333223342895
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,1536,0.0074976002176602675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2048,1024,0.030932267506917317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2048,1024,0.006031999985376993
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,1024,0.005658666789531708
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,1024,0.0070592001080513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2048,768,0.02905279994010925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2048,768,0.005682133138179779
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,768,0.00514026681582133
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,768,0.006857599814732869
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2048,512,0.02722773353258769
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2048,512,0.005445333321889242
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,512,0.0044938668608665465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,512,0.006548266609509785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2560,256,0.005177600185076395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2048,256,0.025255467494328814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,256,0.003651199986537298
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,256,0.006285866598288219
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2560,128,0.004925866425037384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2048,128,0.024409600098927817
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2048,128,0.004850133260091146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,128,0.0033834666013717652
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,128,0.0060586666067441305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2048,64,0.024026666084925333
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,64,0.0032106667757034303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,64,0.006189866860707601
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,2048,32,0.024276266495386757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,32,0.003435733417669932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,32,0.006449066599210103
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1536,65536,0.40869760513305664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,65536,0.09008746941884359
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,65536,0.08275626500447592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1536,16384,0.10304746627807618
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2048,65536,0.09633386929829915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2048,12288,0.02223680019378662
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,16384,0.030713599920272828
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,16384,0.0236735999584198
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2048,8192,0.015568000078201295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1536,12288,0.08533973693847656
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,12288,0.025195733706156416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,12288,0.01762346625328064
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1536,10240,0.07340586980183919
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1536,10240,0.018305067221323648
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,10240,0.02239360014597575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,10240,0.0156960000594457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1536,8192,0.015661866466204325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1536,8192,0.06394879817962647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,8192,0.018437333901723228
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,8192,0.01362666686375936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1536,7168,0.014054399728775025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1536,7168,0.07730666796366373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,7168,0.014197333653767904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,7168,0.012768000364303589
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1536,6144,0.06976959705352784
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,6144,0.01434346636136373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,6144,0.011814399560292562
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1536,5120,0.060725335280100504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1536,5120,0.011107200384140014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,5120,0.012635733683904013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,5120,0.011080533266067505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1536,4096,0.051532800992329916
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1536,4096,0.009802666306495667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,4096,0.010540800293286641
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,4096,0.009913600484530131
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2048,3584,0.009410132964452107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1536,3584,0.047763200600941975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1536,3584,0.00928533375263214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,3584,0.009607467055320739
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,3584,0.009522133072217305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1536,3072,0.044248533248901364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1536,3072,0.008575999736785888
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,3072,0.008105599880218506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,3072,0.008985599875450135
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1536,2560,0.04085439840952555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,2560,0.008248533308506011
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,2560,0.008454400300979614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1536,2048,0.03720426559448242
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1536,2048,0.007321600119272869
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2048,1536,0.006771199901898702
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,2048,0.007212799787521362
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,2048,0.007765333354473114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1536,1536,0.03408960103988647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1536,1536,0.006741333504517872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,1536,0.005970133344332377
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,1536,0.007356800138950348
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1536,1024,0.030722133318583172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,1024,0.004958933095137278
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,1024,0.00702400008837382
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1536,768,0.02878933350245158
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1536,768,0.005685333410898844
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,768,0.004445866743723551
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,768,0.006785066425800323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1536,512,0.02696640094121297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1536,512,0.005386666456858317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,512,0.003995733211437861
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2048,256,0.0051242664456367494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,512,0.006483200192451477
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1536,256,0.025226666529973345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1536,256,0.005021866659323374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,256,0.003588266670703888
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,256,0.006260266900062561
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1536,128,0.024193066358566283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1536,128,0.004846933484077454
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,128,0.0033962666988372804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,128,0.006097066899140676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1536,64,0.024022400379180908
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,64,0.0031829332311948144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,64,0.006192000210285186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1536,32,0.02397973338762919
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,32,0.003197866678237915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,32,0.006372266511122386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1024,65536,0.17287786801656085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1536,65536,0.09451839923858643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1024,65536,0.408677323659261
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,65536,0.06567360162734985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,65536,0.08510613441467285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1024,16384,0.041517865657806394
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,16384,0.023548799753189086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1024,16384,0.10354026953379314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,16384,0.021486934026082358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1024,12288,0.03362773259480794
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1024,12288,0.08556586901346842
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,12288,0.019015467166900633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,12288,0.017318399747212727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1024,10240,0.07337706883748372
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1024,10240,0.02815786600112915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,10240,0.015479466319084168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,10240,0.015677866339683533
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1024,8192,0.06374293168385824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,8192,0.01318826675415039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,8192,0.013569066921869913
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1024,7168,0.02029013236363729
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1024,7168,0.07740159829457602
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,7168,0.013251200318336487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,7168,0.01286720037460327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1024,6144,0.06883839766184488
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1024,6144,0.017539199193318686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,6144,0.01162453293800354
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,6144,0.011865599950154623
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1024,5120,0.05839680035909017
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1024,5120,0.015213867028554281
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,5120,0.01025279959042867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,5120,0.011075199643770853
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1536,12288,0.02103786667188009
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1024,4096,0.05047359863917032
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1024,4096,0.013187199831008911
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,4096,0.008785067001978557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1536,16384,0.027334400018056232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,4096,0.009802666306495667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1024,3584,0.047576534748077395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1024,3584,0.011808000008265178
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,3584,0.008089600006739299
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,3584,0.009462400277455648
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1024,3072,0.044402134418487546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,3072,0.007401599983374278
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,3072,0.008999466896057129
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1024,2560,0.040771198272705075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1024,2560,0.0095551997423172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,2560,0.007169066866238911
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,2560,0.00834986666838328
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1024,2048,0.03705919981002807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,2048,0.0064181332786877945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,2048,0.007844266792138416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1024,1536,0.033964800834655764
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1536,6144,0.012482133507728577
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1024,1536,0.0074976002176602675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,1536,0.005654400090376536
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,1536,0.007454933226108551
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1024,1024,0.006554666658242543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1024,1024,0.03083946704864502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,1024,0.0046858668327331545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,1024,0.006913066903750102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1024,768,0.028806400299072266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1024,768,0.006037333110968272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,768,0.004248533149560293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,768,0.00674773355325063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1024,512,0.026999467611312868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1024,512,0.00559146652619044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,512,0.0038346665600935614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,512,0.006450133522351582
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1024,256,0.02528426647186279
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1024,256,0.005258666475613912
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,256,0.00347626656293869
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,256,0.006202666461467743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1024,128,0.024422399202982583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1024,128,0.004958933095137278
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,128,0.00602346658706665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,128,0.00322026660044988
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1024,64,0.023897600173950196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,64,0.0030464000999927522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,64,0.0060810665289560955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,1024,32,0.023827199141184488
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,32,0.003091199944416682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,32,0.006172800064086914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,768,65536,0.13967572848002116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,65536,0.052818131446838376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,65536,0.08279039859771728
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,768,65536,0.40874665578206376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,768,16384,0.10363626480102539
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,768,16384,0.040226133664449056
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,16384,0.01767680048942566
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1536,2560,0.007890133559703827
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,16384,0.020949333906173706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,768,12288,0.030729599793752033
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,768,12288,0.08531306584676107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,12288,0.01588373382886251
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,12288,0.017190400759379068
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,768,10240,0.07325332959493001
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,768,10240,0.027203200260798137
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,10240,0.0152319997549057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,10240,0.015452800194422403
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,768,8192,0.06282773415247599
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,768,8192,0.02285226583480835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,8192,0.012133333086967468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,8192,0.013470932841300964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1536,1024,0.005981866518656413
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,768,7168,0.07628693580627441
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,768,7168,0.020538665850957236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,7168,0.01297813355922699
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,7168,0.012683733304341634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,768,6144,0.06898773511250814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,768,6144,0.01751040021578471
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,6144,0.01165120005607605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,5120,0.01014293332894643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,6144,0.01167466640472412
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,768,5120,0.058362666765848795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,768,5120,0.015411200126012168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,5120,0.01095466713110606
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,768,4096,0.05056853294372558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,768,4096,0.01286079982916514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,4096,0.008731733759244282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,4096,0.00981226662794749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,768,3584,0.04826879898707072
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,768,3584,0.011331199606259664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,3584,0.008016000191370647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,3584,0.009309867024421692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,768,3072,0.04448533455530802
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,768,3072,0.010191999872525533
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,3072,0.007344000041484833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,3072,0.008877866466840108
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,768,2560,0.040958933035532635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,2560,0.00659093310435613
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,2560,0.008312533299128216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,768,2048,0.037267200152079266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,2048,0.005776000022888183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,2048,0.007751466830571492
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,768,1536,0.03391466538111369
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,768,1536,0.00730453332265218
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,1536,0.0051018665234247845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,1536,0.0073173334201176955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,768,1024,0.030588799715042116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,768,1024,0.006221866607666016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,1024,0.004379733403523763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,1024,0.0068800002336502075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,768,768,0.029037866989771528
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,768,768,0.005796266595522562
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,768,0.004073599974314371
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,768,0.006663466493288676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,768,512,0.0271999994913737
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,512,0.003692800054947535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,512,0.006351999938488007
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1024,8192,0.023553067445755006
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,768,256,0.025174399216969807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,256,0.003365333378314972
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,256,0.006169599791367849
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,768,128,0.02435413400332133
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,768,128,0.004871466755867004
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,128,0.0031221332649389905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,128,0.0060149331887563075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,768,64,0.024069333076477052
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,64,0.002942933390537898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,64,0.006040533383687338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,768,32,0.023827199141184488
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,32,0.002932266642649968
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,32,0.006087466577688853
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,512,65536,0.07992533047993979
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1024,3072,0.010483200351397198
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,65536,0.04100159804026286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,65536,0.07611626784006754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,512,65536,0.40779733657836914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,512,16384,0.022046933571497597
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,16384,0.013767466942469279
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,512,16384,0.10330026944478352
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,16384,0.020814933379491172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,512,12288,0.08479253451029459
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,12288,0.013071999947230021
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,12288,0.01716266671816508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,512,10240,0.07269653479258219
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,512,10240,0.015621333320935567
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,10240,0.011913599570592244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,10240,0.015403733650843302
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,512,8192,0.062338133653004967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,8192,0.01046720047791799
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,8192,0.013364266355832419
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,512,7168,0.07512746651967367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,512,7168,0.012505599856376648
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,7168,0.010010666648546855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,7168,0.012577066818873087
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1024,2048,0.00844266712665558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,512,6144,0.06698559919993083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,512,6144,0.01106666624546051
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,6144,0.009317333499590557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,6144,0.011692800124486287
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,512,5120,0.05845973491668701
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,512,5120,0.009945600231488546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,5120,0.008787199854850769
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,5120,0.010893866419792175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,512,4096,0.05048213402430216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,512,4096,0.008842666943868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,4096,0.008053333560625712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,4096,0.009773866335550944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,512,3584,0.047326934337615964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,512,3584,0.008364799618721008
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,3584,0.007925333579381307
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,3584,0.009275733431180318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,512,3072,0.04413546721140544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,3072,0.007155199845631917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,3072,0.008757332960764568
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,512,2560,0.04018666744232178
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,768,2560,0.009144533673922222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,2560,0.006484266618887584
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,768,2048,0.008065066734949748
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,2560,0.008313600222269695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,512,2048,0.03675946791966756
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,512,2048,0.0066997334361076355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,2048,0.005737600227197012
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,2048,0.007752533257007599
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,512,1536,0.033973332246144614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,512,1536,0.0061951999862988796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,1536,0.005053866902987162
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,1536,0.007287466526031494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,512,1024,0.03054080009460449
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,512,1024,0.005633066594600678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,1024,0.004374399781227112
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,1024,0.006780800223350525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,512,768,0.028552534182866414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,768,512,0.005606399973233541
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,768,256,0.00510506679614385
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,768,0.0040277334551016486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,768,0.006588799754778545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,512,512,0.026923733949661254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,512,512,0.0050240000089009605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,512,0.0036650667587916053
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,512,0.006278400123119354
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,512,256,0.024829866488774617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,256,0.0033546666304270422
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,256,0.006052266558011373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,512,128,0.02400746742884318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,512,128,0.004644266764322917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,128,0.003159466634194056
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,128,0.0059338668982187905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,512,64,0.023754666248957314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,64,0.0029781334102153777
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,64,0.0059232001503308615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,512,32,0.023404800891876222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,32,0.002997333308060964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,32,0.0060138667623202005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,512,12288,0.017772799730300902
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,256,65536,0.07974186738332113
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,512,8192,0.01343786617120107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,256,65536,0.40909334818522136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,65536,0.030819199482599896
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,65536,0.07782080173492431
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,256,16384,0.02188160022099813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,256,16384,0.10320746898651123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,16384,0.020836265881856282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,16384,0.012422399719556172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,256,12288,0.01767680048942566
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,256,12288,0.08473280270894369
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,12288,0.01708266735076904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,12288,0.010632533828417461
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,256,10240,0.01558080017566681
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,256,10240,0.07275199890136719
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,10240,0.00976746678352356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,10240,0.01544533371925354
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,256,8192,0.062205866972605384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,256,8192,0.013396267096201578
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,8192,0.00881599982579549
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,256,7168,0.07419306437174479
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,8192,0.013481600085894265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,256,7168,0.012247467041015625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,7168,0.008718933661778767
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,7168,0.012683733304341634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,256,6144,0.0679754654566447
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,256,6144,0.010998400052388509
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,6144,0.007914666831493378
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,256,5120,0.058766933282216396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,6144,0.011730133493741354
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,256,5120,0.009899733463923137
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,5120,0.007682133217652638
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,5120,0.010886399944623312
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,256,4096,0.050246401627858484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,256,4096,0.008779733379681905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,4096,0.0070613334576288865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,4096,0.00976746678352356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,512,2560,0.007223466535409291
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,256,3584,0.04711146752039592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,256,3584,0.008285866677761078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,3584,0.007565866907437642
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,3584,0.009276800354321798
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,256,3072,0.04381440083185832
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,256,3072,0.007742933432261149
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,3072,0.007342933118343354
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,3072,0.00877333382765452
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,512,3072,0.00788266658782959
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,256,2560,0.04031360149383545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,256,2560,0.007083733379840851
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,2560,0.006442666550477346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,2560,0.008191999793052674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,256,2048,0.03673706849416097
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,256,2048,0.0066453332702318835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,2048,0.005721599857012431
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,2048,0.007639466722806294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,256,1536,0.03339200019836426
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,256,1536,0.00614933321873347
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,1536,0.005013333261013031
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,1536,0.007258666555086772
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,256,1024,0.03017173409461975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,256,1024,0.005454933146635691
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,1024,0.004330666859944662
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,1024,0.006757333377997081
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,256,768,0.028518400589625043
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,256,768,0.005367466807365417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,768,0.004008533308903376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,768,0.0065290664633115125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,256,512,0.026428800821304322
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,256,512,0.0050570666790008545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,512,0.0036256000399589538
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,512,0.006318933268388112
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,256,256,0.025161600112915038
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,256,256,0.0047082667549451195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,128,0.0030954666435718536
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,128,0.005897599955399831
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,256,0.003304533412059148
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,256,0.006072533130645752
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,256,128,0.024016000827153525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,256,64,0.0240447998046875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,64,0.002914133419593175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,64,0.00595413347085317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,256,32,0.02350506583849589
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,32,0.0029877332349618276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,32,0.00592853327592214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,65536,0.023565866549809775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,65536,0.07606720129648845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,128,65536,0.4079765319824219
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,128,16384,0.10307946999867756
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,128,16384,0.02195733388264974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,16384,0.009136000275611877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,16384,0.020899200439453126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,128,12288,0.08470613161722819
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,128,12288,0.01766293247540792
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,12288,0.008189866443475087
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,12288,0.017157334089279174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,128,10240,0.07242666880289714
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,128,10240,0.015556266903877259
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,10240,0.007718400160471599
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,10240,0.015416533748308817
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,128,8192,0.060347731908162436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,8192,0.0073642666141192125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,8192,0.013371733824412027
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,128,7168,0.07508479754130046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,128,7168,0.012281599640846252
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,7168,0.007239466905593872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,7168,0.012565333644549051
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,128,6144,0.06591253280639649
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,128,6144,0.010999466975529988
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,6144,0.007043200234572093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,6144,0.01160533328851064
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,128,5120,0.057726931571960446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,128,5120,0.00990826686223348
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,5120,0.007380266487598419
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,5120,0.010779733459154766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,128,4096,0.05020800034205118
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,4096,0.00692799985408783
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,4096,0.00965119997660319
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,128,3584,0.04715840021769206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,128,3584,0.008246399958928426
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,3584,0.007448533177375793
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,3584,0.00913706620534261
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,128,3072,0.04367573261260986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,3072,0.007008000214894612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,3072,0.008578133583068848
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,128,2560,0.039996798833211264
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,512,768,0.005335466563701629
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,2560,0.006267733375231425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,2560,0.008228266735871632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,128,2048,0.03662933508555095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,512,256,0.0047978664437929785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,128,2048,0.006569600105285645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,2048,0.00763626645008723
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,2048,0.005602133274078369
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,128,1536,0.0329802672068278
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,1536,0.004923733572165171
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,1536,0.007176533341407776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,128,1024,0.03007146716117859
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,128,1024,0.0054517333706219995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,1024,0.00425600012143453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,1024,0.006741333504517872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,128,768,0.028016000986099243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,128,768,0.005224533379077911
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,768,0.0038912000755469004
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,768,0.00653546651204427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,128,512,0.026369067033131917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,256,128,0.004669866462548574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,512,0.0036202666660149893
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,512,0.006236800054709116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,128,256,0.02480213244756063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,128,256,0.0047872001926104225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,256,0.0033333333830038703
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,256,0.00603413333495458
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,128,128,0.02398080031077067
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,128,128,0.004647466540336609
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,128,65536,0.07969386577606201
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,128,0.0030879999200503034
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,128,0.005941333373387655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,128,64,0.023937066396077476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,64,0.0029397333661715193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,64,0.0059114664793014525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,128,32,0.023434666792551677
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,32,0.0029834667841593427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,32,0.0059338668982187905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,65536,0.014070399602254233
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,65536,0.07358506520589193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,64,65536,0.409333324432373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,64,16384,0.10273813406626384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,16384,0.008054399987061818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,16384,0.020747733116149903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,12288,0.007477333148320516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,64,12288,0.0843061367670695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,12288,0.017035732666651406
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,10240,0.007388799885908763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,64,10240,0.07199359734853109
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,10240,0.015221333503723145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,64,8192,0.06185706853866577
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,8192,0.007359999914964039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,8192,0.0132832000652949
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,64,7168,0.075436798731486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,7168,0.0071263998746871945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,7168,0.01244160036245982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,64,6144,0.06838826338450113
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,6144,0.0069472000002861025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,6144,0.011524266997973124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,128,8192,0.013401599725087485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,5120,0.007298133273919423
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,64,5120,0.0572213331858317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,5120,0.010659199953079224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,64,4096,0.05000213384628296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,4096,0.006855466465155284
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,4096,0.009662933150927226
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,128,3072,0.0076896001895268755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,64,3584,0.047116800149281816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,3584,0.007427200178305308
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,3584,0.009128533800443013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,3072,0.007050666709740956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,64,3072,0.04364266792933146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,3072,0.008678399523099263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,128,2560,0.007079466680685679
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,64,2560,0.04012266794840495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,2560,0.006250666578610738
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,2560,0.008097066481908163
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,128,4096,0.008784000078837078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,64,2048,0.036245334148406985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,2048,0.005565866827964783
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,2048,0.007503999769687653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,64,1536,0.032978133360544844
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,1536,0.004922666649023692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,1536,0.00710399995247523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,64,1024,0.030052266518274945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,1024,0.004232533276081085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,1024,0.006705066561698914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,64,768,0.02837013403574626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,768,0.003851733356714249
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,768,0.006515199939409892
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,512,0.003572266548871994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,64,512,0.02643199960390727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,512,0.006185600161552429
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,64,256,0.02485226591428121
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,256,0.0032074667513370516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,256,0.005972266693909963
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,64,128,0.023935999472935995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,128,0.0030346666773160298
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,128,0.005868799984455109
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,64,64,0.023727999130884806
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,128,1536,0.006129066646099091
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,64,0.0028778667251269023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,64,0.0058890665570894875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,64,32,0.023231999079386393
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,32,0.0028927999238173166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,32,0.005851733187834421
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,65536,0.0710752010345459
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,65536,0.013056000073750814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,32,65536,0.40888532002766925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,128,512,0.004907733201980591
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,32,16384,0.10257919629414876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,16384,0.007259733478228251
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,16384,0.020639999707539877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,32,12288,0.08439146677652995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,12288,0.00697386662165324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,12288,0.01699840029080709
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,32,10240,0.07232426802317302
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,10240,0.0072512000799179075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,10240,0.015175466736157736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,32,8192,0.06133546829223633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,8192,0.007188266515731812
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,8192,0.013168000181516013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,32,7168,0.0728928009668986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,7168,0.006940799951553345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,7168,0.012427733341852824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,32,6144,0.06668586730957031
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,6144,0.0068810666600863145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,6144,0.011461333433787028
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,32,5120,0.05719253222147623
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,5120,0.0072543998559316
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,5120,0.01053013304869334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,32,4096,0.05026880105336508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,4096,0.006900266806284587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,4096,0.009403733412424724
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,32,3584,0.04670613209406535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,3584,0.007229866584142049
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,3584,0.009041066964467366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,32,3072,0.04371840159098307
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,3072,0.0069013332327206925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,3072,0.008572799960772197
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,32,2560,0.04017279942830403
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,2560,0.00619946668545405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,2560,0.008075733482837678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,32,2048,0.03622080087661743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,2048,0.005520000060399374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,2048,0.007534933090209961
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,32,1536,0.03296639919281006
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,1536,0.004886400202910105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,768,0.006445866823196411
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,1536,0.007142400244871776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,32,1024,0.02981013258298238
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,1024,0.004203733305136363
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,1024,0.006623999774456024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,32,768,0.02815999984741211
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,768,0.0038880000511805216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,32,512,0.02650986711184184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,512,0.0036085332433382668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,512,0.006126933296521505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,32,256,0.024754132827123007
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,256,0.0031946666538715364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,256,0.006000000238418579
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,32,128,0.024014933904012045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,128,0.002993066608905792
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,128,0.005814399818579356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,32,64,0.023499733209609984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,64,0.002865066627661387
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,64,0.0058218667904535925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,192,32,32,0.023190399010976158
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,32,0.0028661333024501802
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,32,0.005798399945100148
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,65536,16384,0.5569194793701172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,65536,16384,0.8052127838134766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,65536,12288,0.4352245330810547
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,65536,16384,0.8023402531941732
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,65536,12288,0.5925397237141927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,65536,16384,0.6215882619222005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,65536,12288,0.47423680623372394
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,65536,10240,0.37687253952026367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,65536,12288,0.6276063919067383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,65536,10240,0.49553600947062176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,65536,8192,0.30239893595377604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,65536,10240,0.518555736541748
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,65536,8192,0.3945002555847168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,65536,10240,0.4138943990071614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,65536,8192,0.34517545700073243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,65536,8192,0.4446367899576823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,65536,7168,0.2737077395121256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,65536,7168,0.3506634712219238
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,65536,7168,0.32423038482666017
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,65536,7168,0.3868330637613932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,65536,6144,0.2544842720031738
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,65536,6144,0.3180821418762207
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,65536,6144,0.32423359553019204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,65536,5120,0.21293973922729492
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,65536,5120,0.19403519630432128
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,65536,5120,0.2535626729329427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,65536,5120,0.2696917215983073
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,65536,4096,0.17134186426798503
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,65536,4096,0.20699413617451987
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,65536,4096,0.2118826707204183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,65536,3584,0.13742720286051432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,65536,3584,0.18104640642801922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,65536,3584,0.15300480524698895
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,65536,3584,0.18895146052042644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,65536,3072,0.11609493096669514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,65536,3072,0.13912960688273113
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,65536,3072,0.15761173566182454
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,65536,3072,0.16247785886128743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,65536,2560,0.12196053663889568
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,65536,2560,0.09819520314534505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,65536,2560,0.13383359909057618
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,65536,2560,0.1429493268330892
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,65536,2048,0.1098357359568278
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,65536,2048,0.11132160027821858
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,65536,2048,0.10812160174051921
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,65536,1536,0.08828693230946859
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,65536,1536,0.060065066814422606
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,65536,1536,0.0860373338063558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,65536,1536,0.08314879735310873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,65536,1024,0.06980693340301514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,65536,1024,0.041266131401062014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,65536,1024,0.061457065741221104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,65536,1024,0.05703039964040121
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,65536,768,0.05894613265991211
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,65536,768,0.048816001415252684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,65536,768,0.045516800880432126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,65536,512,0.05170773267745972
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,65536,512,0.026682666937510174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,65536,512,0.0371509313583374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,65536,512,0.03607146739959717
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,65536,256,0.017448532581329345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,65536,256,0.04519253174463908
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,65536,256,0.022488532463709514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,65536,256,0.030456533034642536
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,65536,128,0.037357866764068604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,65536,128,0.012847999731699625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,65536,128,0.011597866813341778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,65536,128,0.022977066040039063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,65536,64,0.038839467366536456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,65536,64,0.009563733140627544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,65536,64,0.02858133316040039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,65536,32,0.04317440191904704
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,65536,32,0.009954133629798889
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,65536,32,0.037088000774383546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,65536,0.873255475362142
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,65536,0.8569983800252279
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,16384,65536,0.7697354634602864
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,16384,16384,0.15104959805806478
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,16384,16384,0.20144319534301758
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,16384,0.20402560234069825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,16384,0.2120192050933838
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,16384,12288,0.1589248021443685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,16384,12288,0.12143466472625733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,65536,6144,0.2371328035990397
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,12288,0.14718292554219564
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,12288,0.1528447945912679
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,10240,0.1251370668411255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,16384,10240,0.1395605405171712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,65536,4096,0.15304959615071614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,10240,0.133133864402771
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,16384,8192,0.11636906464894611
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,8192,0.09989866415659586
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,8192,0.11440959771474202
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,16384,7168,0.12425920168558757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,7168,0.08765119711558024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,7168,0.09727786382039388
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,16384,6144,0.10618133544921875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,16384,6144,0.06075733502705892
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,65536,2048,0.08198826313018799
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,6144,0.07819413344065348
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,6144,0.08226986726125082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,16384,5120,0.09107840061187744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,16384,5120,0.05198826789855957
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,5120,0.06619626681009928
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,5120,0.06932799816131592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,16384,4096,0.07779839833577475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,16384,4096,0.04248533248901367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,4096,0.055859200159708654
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,4096,0.05411626497904459
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,16384,3584,0.07096532980600992
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,16384,3584,0.038184531529744464
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,3584,0.04785386721293132
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,65536,768,0.033573333422342935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,3584,0.04825386603673299
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,16384,3072,0.03258986671765645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,16384,3072,0.06554773251215616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,3072,0.042410667737325033
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,3072,0.04315840005874634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,16384,2560,0.05892906586329142
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,16384,2560,0.02855679988861084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,2560,0.036150399843851724
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,2560,0.035417600472768145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,16384,2048,0.05306026538213095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,2048,0.03025599916776021
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,16384,2048,0.02253440022468567
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,2048,0.029286400477091475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,16384,1536,0.04396266539891561
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,1536,0.02437333265940348
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,1536,0.02410773237546285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,16384,1024,0.03797760009765625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,16384,1024,0.012610133488972983
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,1024,0.011627733707427979
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,1024,0.01713599960009257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,16384,768,0.03318933248519897
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,768,0.010069333513577779
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,768,0.014813866217931113
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,16384,512,0.03091413378715515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,16384,512,0.009300266702969868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,512,0.007976533472537994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,512,0.012088533242543538
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,16384,256,0.02811093330383301
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,16384,256,0.008055466910203297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,256,0.006451199948787689
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,256,0.010564266641934713
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,16384,128,0.02645439902941386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,16384,128,0.0074442664782206226
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,128,0.005307733515898387
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,128,0.009909333785374959
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,16384,64,0.026634667317072553
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,64,0.004683733483155568
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,64,0.010605866710344952
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,16384,32,0.02731093366940816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,32,0.005161599814891815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,32,0.013451733191808066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,16384,65536,0.5460597356160481
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,12288,65536,0.3999861399332682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,12288,65536,0.667905044555664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,65536,0.6439317067464192
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,65536,0.7089205423990885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,12288,16384,0.1723840077718099
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,16384,0.16694506009419757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,16384,10240,0.10059093634287516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,16384,0.16742293039957684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,12288,12288,0.07726933161417643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,12288,12288,0.13847467104593914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,12288,0.13062187035878497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,16384,8192,0.07801813284556071
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,12288,0.12303253014882405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,12288,10240,0.06467626492182413
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,12288,10240,0.12207679748535157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,10240,0.09510826269785563
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,10240,0.10570133527119954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,12288,8192,0.09962773323059082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,12288,8192,0.05151360034942627
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,8192,0.07597866853078207
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,8192,0.08684159914652506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,16384,7168,0.0701695998509725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,12288,7168,0.04684693415959676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,7168,0.07557546297709147
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,12288,7168,0.1099509318669637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,7168,0.06862186590830485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,12288,6144,0.04164160092671712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,6144,0.06050879955291748
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,12288,6144,0.0960640033086141
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,6144,0.0649450659751892
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,12288,5120,0.08368213176727295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,5120,0.050977067152659095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,5120,0.05642666816711426
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,12288,4096,0.07243626912434896
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,4096,0.0416320006052653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,4096,0.043442134062449136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,12288,3584,0.026858667532602947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,12288,3584,0.06632746855417887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,3584,0.0367082675298055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,3584,0.04017920096715291
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,12288,3072,0.0628554662068685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,3072,0.032687999804814655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,3072,0.03568106492360433
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,12288,2560,0.054560001691182455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,12288,2560,0.019002666076024376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,2560,0.02807146708170573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,2560,0.03018239935239156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,12288,2048,0.04724693298339844
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,12288,2048,0.014972800016403198
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,2048,0.02355733315149943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,2048,0.02407360076904297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,12288,1536,0.04128640095392863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,16384,1536,0.01722986698150635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,12288,1536,0.012395733594894409
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,1536,0.014682666460673014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,1536,0.019873066743214925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,16384,768,0.010949333508809406
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,12288,1024,0.035957332452138266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,1024,0.008384000261624653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,12288,1024,0.010615467031796774
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,1024,0.014638933539390563
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,12288,768,0.032365866502126056
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,768,0.007482666770617168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,768,0.0131221334139506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,12288,512,0.03009386658668518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,12288,512,0.008227199812730153
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,512,0.0062047998110453285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,512,0.010923733313878376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,12288,256,0.028172800938288372
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,12288,256,0.007264000177383423
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,256,0.009830400347709656
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,256,0.005128533144791921
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,12288,128,0.026368000109990436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,128,0.004481066763401031
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,12288,128,0.006626133124033611
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,128,0.009219200412432352
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,12288,64,0.026270933945973712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,64,0.004035199930270513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,64,0.009241599837938945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,12288,32,0.026071467002232868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,32,0.0044159998496373495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,32,0.011002666751543681
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,10240,65536,0.39676265716552733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,65536,0.5152885437011718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,10240,65536,0.6534101486206054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,10240,16384,0.17137494087219238
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,10240,16384,0.09583893616994223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,65536,0.5716810862223307
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,16384,0.14855252901713054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,12288,16384,0.10378026962280273
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,16384,0.133622407913208
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,10240,12288,0.13172586758931476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,12288,0.09854826927185059
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,12288,0.10214400291442871
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,10240,10240,0.06047253211339315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,10240,10240,0.1171114683151245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,10240,0.08208853403727213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,10240,0.0868170658747355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,10240,8192,0.04805759986241658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,10240,8192,0.09745173454284668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,8192,0.06647786696751913
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,8192,0.07129387060801187
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,7168,0.059324800968170166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,10240,7168,0.10744106769561768
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,7168,0.06152960062026978
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,10240,6144,0.09334613482157389
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,10240,6144,0.03765546480814616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,6144,0.05317759911219279
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,6144,0.05228373209635416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,12288,5120,0.0344703992207845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,10240,5120,0.08161919911702474
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,10240,5120,0.03248106638590495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,5120,0.04445866743723552
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,5120,0.046011734008789065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,10240,4096,0.07019946575164795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,12288,4096,0.029080533981323244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,4096,0.03580373525619507
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,4096,0.03700266679128011
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,10240,3584,0.06514879862467447
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,10240,3584,0.023449599742889404
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,3584,0.032067199548085526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,3584,0.033446399370829265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,10240,3072,0.059676798184712734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,12288,3072,0.022844799359639487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,10240,3072,0.01986133257548014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,3072,0.028707200288772584
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,3072,0.02960639993349711
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,10240,2560,0.05099733273188273
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,10240,2560,0.017033600807189943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,2560,0.025011199712753295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,2560,0.024382932980855306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,10240,2048,0.04502293268839518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,10240,2048,0.013617066542307535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,2048,0.020467199881871543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,2048,0.02087999979654948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,10240,1536,0.0398634672164917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,10240,1536,0.01167146662871043
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,1536,0.011035733421643575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,1536,0.0169706662495931
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,10240,1024,0.034620801607767746
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,10240,1024,0.00947093367576599
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,1024,0.008983467022577922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,1024,0.013336533308029174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,10240,768,0.03179306586583455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,10240,768,0.008628267049789428
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,768,0.00766186664501826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,768,0.011422933141390482
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,10240,512,0.029643734296162922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,512,0.006549333532651265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,512,0.010296533505121868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,10240,256,0.02728533347447713
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,256,0.00550186683734258
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,256,0.008873599767684936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,10240,128,0.026324266195297243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,12288,768,0.009114666779836019
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,128,0.004700799783070883
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,128,0.008397866288820903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,10240,64,0.025857067108154295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,64,0.004159999887148539
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,64,0.008507733543713886
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,10240,32,0.026609067122141523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,32,0.004310399790604909
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,32,0.009796266754468281
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,8192,65536,0.2359018643697103
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,8192,65536,0.4928266525268555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,65536,0.40505278905232744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,8192,16384,0.13266133467356364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,65536,0.4267744064331055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,8192,16384,0.06461759805679321
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,16384,0.09831466674804687
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,16384,0.10690666834513347
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,8192,12288,0.05162986516952515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,8192,12288,0.10516266822814942
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,12288,0.07684266567230225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,12288,0.07991466522216797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,10240,12288,0.07210346857706705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,8192,10240,0.04443200031916301
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,8192,10240,0.09379626909891764
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,10240,0.06477440198262532
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,10240,0.0661354660987854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,8192,8192,0.07922453085581461
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,8192,0.052043732007344565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,8192,0.05697493155797323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,8192,7168,0.09192426999409994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,8192,7168,0.03312106728553772
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,7168,0.046033068497975664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,7168,0.04924906492233276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,8192,6144,0.029656533400217695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,8192,6144,0.0797205368677775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,10240,7168,0.043339733282725015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,6144,0.04103999932607015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,6144,0.04192320108413696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,8192,5120,0.07073813279469808
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,8192,5120,0.02535360058148702
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,5120,0.034782934188842776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,5120,0.03673066695531209
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,8192,4096,0.06174079974492391
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,8192,4096,0.018439465761184694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,4096,0.028739200035731
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,4096,0.029339732726414998
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,8192,3584,0.05677760044733683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,3584,0.025946666797002155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,3584,0.02603093385696411
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,10240,4096,0.026522666215896606
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,8192,3072,0.051330133279164636
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,8192,3072,0.014564266800880432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,3072,0.022821333010991415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,3072,0.023012266556421915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,8192,2560,0.046189868450164796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,2560,0.017666133244832356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,2560,0.019821866353352865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,8192,2048,0.04069973230361938
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,8192,2048,0.011642666657765706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,2048,0.010921600461006164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,2048,0.01684266726175944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,8192,1536,0.036614398161570236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,8192,1536,0.010016000270843506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,1536,0.008785067001978557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,1536,0.014097066720326743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,8192,1024,0.03232746720314026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,8192,1024,0.008573866883913676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,1024,0.006937600175539653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,1024,0.010883200168609618
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,8192,768,0.030103466908137005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,768,0.006323199967543285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,10240,512,0.007714133461316426
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,768,0.009889066219329834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,8192,512,0.027933865785598755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,10240,256,0.006916266679763794
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,8192,512,0.0067669332027435304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,512,0.0051136001944541935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,512,0.008707200487454731
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,8192,256,0.025934932629267375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,8192,256,0.006077866752942403
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,256,0.004496000210444133
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,256,0.007998933394749958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,8192,128,0.024707200129826863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,128,0.0039893334110577905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,10240,128,0.006306133170922597
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,128,0.007707733412583668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,8192,64,0.02454293370246887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,64,0.003667200108369192
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,64,0.00802346666653951
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,8192,32,0.02442453304926554
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,32,0.0040618665516376495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,32,0.009195733070373534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,7168,65536,0.31953385670979817
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,7168,65536,0.46508267720540364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,65536,0.35953388214111326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,7168,16384,0.08302293618520101
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,7168,16384,0.12711466948191324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,65536,0.41909332275390626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,16384,0.08649706840515137
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,16384,0.09944533507029216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,7168,12288,0.06495893398920695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,7168,12288,0.10140053431193034
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,12288,0.0680512030919393
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,12288,0.07472639878590902
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,7168,10240,0.08972906271616618
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,7168,10240,0.053729065259297694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,10240,0.057124265034993495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,10240,0.06542186737060547
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,7168,8192,0.07602880001068116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,7168,8192,0.0426474650700887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,8192,0.0460480014483134
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,8192,0.05327893495559692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,7168,7168,0.03815466562906901
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,7168,7168,0.08928746382395426
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,7168,0.04148906469345093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,7168,0.04710613489151001
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,7168,6144,0.07757439613342285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,7168,6144,0.034178133805592856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,6144,0.03709333340326945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,6144,0.04116053183873494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,7168,5120,0.06909120082855225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,5120,0.03143253326416016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,5120,0.034468265374501546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,7168,4096,0.05973653395970663
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,8192,8192,0.03650453488032023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,7168,4096,0.02318506638209025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,4096,0.025653332471847534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,4096,0.027409066756566364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,7168,3584,0.054746667544047035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,7168,3584,0.020357332626978555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,3584,0.025183999538421632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,3584,0.025163733959198
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,7168,3072,0.018016000588734947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,7168,3072,0.04907199939092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,3072,0.020894932746887206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,3072,0.022083199024200438
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,7168,2560,0.01621333360671997
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,7168,2560,0.04393279949824015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,2560,0.012315733234087627
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,2560,0.01967573364575704
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,7168,2048,0.04010773499806722
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,7168,2048,0.013943466544151305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,2048,0.009991466999053955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,2048,0.016251732905705772
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,7168,1536,0.035689600308736164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,8192,3584,0.016429866353670754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,1536,0.008382933338483174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,1536,0.013944533467292786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,7168,1024,0.032239999373753866
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,7168,1024,0.009777067104975383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,1024,0.006877866884072621
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,1024,0.01076800028483073
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,7168,768,0.029790933926900225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,7168,768,0.007891199986139933
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,8192,2560,0.013492266337076822
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,768,0.006144000093142191
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,768,0.00946666697661082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,7168,512,0.02763306697209676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,512,0.005313066641489664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,512,0.00876800020535787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,7168,256,0.025702399015426636
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,256,0.004465066889921824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,256,0.007998933394749958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,7168,128,0.024752000967661537
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,7168,128,0.005932799975077311
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,128,0.003961600114901861
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,128,0.007549866537253062
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,64,0.0038986665507157645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,7168,64,0.024570665756861367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,64,0.007670400043328603
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,7168,32,0.02469546596209208
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,32,0.003991466760635376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,8192,768,0.007737599809964498
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,32,0.008742400010426839
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,8192,128,0.005714133381843567
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,6144,65536,0.1835861365000407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,6144,65536,0.4497397422790527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,65536,0.3090410550435384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,65536,0.3989013353983561
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,6144,16384,0.12191572984059651
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,16384,0.07663359642028808
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,16384,0.09119679927825927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,6144,12288,0.09648959636688233
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,6144,12288,0.03998719851175944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,12288,0.060083198547363284
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,12288,0.07022293408711752
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,6144,10240,0.08466560045878092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,6144,10240,0.03465706507364909
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,10240,0.05071893135706583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,10240,0.059249067306518556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,6144,8192,0.07321386337280274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,8192,0.040753066539764404
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,8192,0.04958719809850057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,6144,7168,0.08666133085886638
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,6144,7168,0.026153600215911864
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,7168,0.03693759838740031
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,7168,0.04270399808883667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,6144,6144,0.07592106660207112
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,6144,6144,0.022707200050354003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,6144,0.033114665746688844
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,6144,0.03674026727676392
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,6144,5120,0.06755733489990234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,6144,5120,0.01864853302637736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,5120,0.027670399347941084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,5120,0.03222186764081319
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,6144,4096,0.015960533420244852
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,6144,4096,0.05859839916229248
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,4096,0.02328853408495585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,4096,0.025974400838216144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,6144,3584,0.053741868336995444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,3584,0.02055786649386088
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,3584,0.02272426684697469
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,6144,3072,0.04718826611836751
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,6144,3072,0.01378666659196218
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,3072,0.014457600315411887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,7168,5120,0.028615466753641766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,3072,0.02100800077120463
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,6144,2560,0.012184533476829528
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,6144,2560,0.043025068442026776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,2560,0.011333333452542622
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,2560,0.018273067474365235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,6144,2048,0.03952213525772095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,2048,0.009773866335550944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,2048,0.015621333320935567
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,6144,1536,0.03593386809031169
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,6144,1536,0.009508267045021057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,1536,0.008297599852085114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,1536,0.012966400384902954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,6144,1024,0.03195733428001404
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,6144,1024,0.008762666583061218
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,1024,0.00681386689345042
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,1024,0.010338133573532105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,7168,1536,0.011421866218249003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,6144,768,0.029618134101231892
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,768,0.0059690664211908976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,768,0.009437867005666097
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,6144,512,0.02802986701329549
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,6144,512,0.006678399940331777
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,512,0.005123200019200643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,512,0.008628267049789428
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,7168,512,0.007289599875609081
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,6144,256,0.0258240004380544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,6144,256,0.00589333325624466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,256,0.004401066899299621
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,7168,256,0.006404266754786174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,256,0.007981866598129272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,6144,128,0.02474453250567118
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,128,0.003975466638803482
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,128,0.007574399809042613
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,6144,64,0.02436800003051758
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,64,0.0037109332780043284
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,64,0.007620266576608021
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,6144,32,0.024551467100779215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,32,0.003921066721280416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,32,0.008270933230717977
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,5120,65536,0.2946666717529297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,5120,65536,0.43056853612263996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,65536,0.2496778647104899
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,65536,0.290120538075765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,6144,16384,0.05147626797358194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,5120,16384,0.11099840005238851
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,5120,16384,0.07306453386942545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,16384,0.0750432014465332
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,16384,0.0674069325129191
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,5120,12288,0.09061439832051596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,5120,12288,0.05781439940134684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,12288,0.06004800001780192
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,12288,0.05418453216552734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,5120,10240,0.07983040014902751
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,5120,10240,0.04852586587270101
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,10240,0.04454506635665893
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,10240,0.046876800060272214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,5120,8192,0.0688266674677531
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,8192,0.036049067974090576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,8192,0.03949973185857137
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,5120,7168,0.08252906799316406
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,6144,8192,0.028845866521199543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,5120,7168,0.03323733409245809
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,7168,0.033793067932128905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,7168,0.03288106719652812
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,5120,6144,0.07154026826222738
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,5120,6144,0.02909119923909505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,6144,0.029205334186553956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,6144,0.028562132517496747
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,5120,5120,0.024489599466323852
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,5120,5120,0.06279360055923462
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,5120,0.025202133258183795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,5120,0.02534293333689372
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,5120,4096,0.05528746843338013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,5120,4096,0.019441066185633342
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,4096,0.019719467560450236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,4096,0.020707199970881142
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,5120,3584,0.0515445351600647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,5120,3584,0.018198400735855103
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,6144,3584,0.014612266421318054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,3584,0.014869333306948344
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,3584,0.0187285323937734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,5120,3072,0.045660801728566486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,3072,0.01243839959303538
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,3072,0.017155200242996216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,5120,2560,0.0407317320505778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,5120,2560,0.014281599720319112
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,2560,0.011081600189208984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,2560,0.014728533228238425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,5120,2048,0.037360000610351565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,5120,2048,0.011867733796437581
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,2048,0.009734400113423665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,2048,0.012965333461761475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,5120,1536,0.03439893325169881
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,1536,0.008365866541862488
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,1536,0.010924800237019857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,5120,1024,0.031413332621256514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,5120,1024,0.008431999882062276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,1024,0.006836266815662384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,1024,0.00904960036277771
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,5120,768,0.02945386568705241
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,5120,768,0.007500799993673961
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,6144,2048,0.011156266927719117
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,768,0.006066133578618368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,768,0.008355200290679932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,5120,512,0.027449599901835126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,512,0.005166933437188466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,512,0.007609599828720092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,5120,256,0.025668267409006757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,256,0.004457599918047587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,256,0.007203199962774913
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,5120,128,0.024679466088612875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,5120,128,0.005640533566474914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,128,0.003941333293914795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,128,0.006855466465155284
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,6144,768,0.007479466497898102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,5120,64,0.02440320054690043
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,64,0.0035818666219711304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,64,0.007100800176461537
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,5120,32,0.024201599756876628
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,32,0.003669333209594091
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,32,0.00767680009206136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,6144,128,0.005584000051021576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,4096,65536,0.3880832036336263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,65536,0.2051402727762858
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,65536,0.27627627054850257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,4096,16384,0.108787202835083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,4096,16384,0.046019200483957926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,16384,0.05928320089975993
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,16384,0.06399573485056559
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,4096,12288,0.08622612953186035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,4096,12288,0.036442665259043376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,12288,0.04794986645380656
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,12288,0.052621865272521974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,4096,10240,0.07635413010915121
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,4096,10240,0.030742400884628297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,10240,0.04177600145339966
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,10240,0.04209386507670085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,4096,8192,0.023053866624832154
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,4096,8192,0.0668351968129476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,8192,0.03556053241093953
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,8192,0.03514773448308309
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,4096,7168,0.019722666343053183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,4096,7168,0.07985066572825114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,7168,0.03251840074857076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,7168,0.031250133117039995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,4096,6144,0.0702122688293457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,4096,6144,0.016525866587956746
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,6144,0.029475200176239013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,5120,8192,0.03871999979019165
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,6144,0.02694080074628194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,4096,5120,0.0616159995396932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,5120,0.02599253257115682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,5120,0.023025067647298177
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,4096,4096,0.0524895985921224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,4096,0.015737600127855935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,5120,3072,0.015848533312479655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,4096,0.019411200284957887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,5120,1536,0.010406399766604107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,4096,3584,0.0487829327583313
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,4096,3584,0.011567999919255573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,3584,0.017796266078948974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,3584,0.013762133320172629
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,4096,3072,0.0452458659807841
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,4096,3072,0.010270933310190838
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,3072,0.012872533003489176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,3072,0.01607253352801005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,4096,2560,0.04128746589024861
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,4096,2560,0.00969493289788564
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,2560,0.011085866888364156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,2560,0.014051199952761332
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,4096,2048,0.037596801916758224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,4096,2048,0.008718933661778767
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,2048,0.00920960009098053
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,2048,0.011904000242551168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,5120,512,0.006748799979686737
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,4096,1536,0.03445440133412679
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,4096,1536,0.007979733248551685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,1536,0.007874133189519246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,1536,0.010473600029945374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,4096,1024,0.03128959933916728
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,5120,256,0.006016000111897787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,4096,1024,0.006810666620731353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,4096,768,0.02919573386510213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,4096,768,0.006391466657320659
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,1024,0.006217599908510844
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,1024,0.008855467041333515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,768,0.0053269331653912864
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,768,0.008351999521255492
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,4096,512,0.027462400992711383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,4096,512,0.005914666752020518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,512,0.004594133297602335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,512,0.007531733314196269
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,4096,256,0.025645865996678667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,256,0.00403413325548172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,256,0.007073066631952922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,4096,128,0.024425599972407022
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,4096,128,0.005220266679922739
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,128,0.003623466690381368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,128,0.006769066552321117
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,4096,64,0.02418880065282186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,64,0.003419733295838038
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,64,0.007077333331108093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,4096,65536,0.16350186665852864
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,4096,32,0.024088533719380696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,32,0.003634133438269297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,32,0.007387733459472657
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3584,65536,0.15588693618774413
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,65536,0.17984320322672526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3584,16384,0.10220692952473957
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,65536,0.2766453425089518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3584,65536,0.3942591985066732
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,16384,0.05297706524531046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,16384,0.06352213223775229
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3584,12288,0.0833237330118815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3584,12288,0.034780800342559814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,12288,0.04304320017496745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,12288,0.04930986563364665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3584,10240,0.029230932394663494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3584,10240,0.07454079786936442
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,10240,0.038302934169769286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,10240,0.03988693157831828
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3584,8192,0.065228799978892
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,8192,0.025621332724889118
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,8192,0.03297599951426188
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3584,7168,0.01827413241068522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3584,7168,0.07958079973856608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,7168,0.023679999510447185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,7168,0.02996480067571004
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3584,6144,0.016013866662979125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3584,6144,0.06900266806284586
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,6144,0.020950400829315187
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,6144,0.0264789342880249
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3584,5120,0.06132160027821859
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3584,5120,0.01458026667435964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,5120,0.013357866803805032
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,5120,0.022357332706451415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3584,4096,0.05139733155568441
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3584,4096,0.012497066458066305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,4096,0.01088533302148183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,4096,5120,0.015082666277885437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,4096,0.019243733088175455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3584,3584,0.04802026748657227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3584,3584,0.011141332983970641
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,3584,0.00983679989973704
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,3584,0.016705065965652466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,3072,0.008964266379674275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,4096,4096,0.012411733468373615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,3072,0.015373866756757101
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3584,3072,0.04393279949824015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3584,2560,0.04073706865310669
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,2560,0.0079925333460172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,2560,0.01392213304837545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3584,2048,0.03749013344446818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,2048,0.007179733117421467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,2048,0.011921067039171855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3584,1536,0.0343178669611613
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3584,1536,0.00790293316046397
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,1536,0.006200533111890157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,1536,0.010626133282979329
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3584,1024,0.03095146616299947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3584,1024,0.006806399921576183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,1024,0.005256533126036326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,1024,0.009001599748929341
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3584,768,0.02910933295885722
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3584,768,0.00619946668545405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,768,0.004762666424115499
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,768,0.008409600456555684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3584,512,0.027562665939331054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3584,512,0.005922133227189382
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,512,0.004274133344491323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,512,0.007667199770609538
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3584,256,0.025563732782999678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3584,256,0.0053941334287325535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,256,0.0038176000118255613
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,256,0.007201066613197327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3584,128,0.024660267432530723
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3584,128,0.005188266436258951
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,128,0.003516799956560135
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,128,0.006819200019041698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3584,64,0.024246400594711302
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,64,0.003500800083080927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,64,0.007014399766921997
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3584,32,0.02431360085805257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,32,0.0035946667194366455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,32,0.0072618668278058365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3072,65536,0.37274347941080727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,65536,0.15965973536173503
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,4096,256,0.005475200215975444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3072,16384,0.10192853609720867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,65536,0.20657173792521158
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3072,16384,0.06057279904683431
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,16384,0.04787413279215495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,16384,0.05302293300628662
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3072,12288,0.08032533327738443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,12288,0.03924800157546997
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,12288,0.04000213146209717
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3072,10240,0.07212266921997071
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3072,10240,0.03980160156885783
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3584,16384,0.04403200149536133
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,10240,0.03400533199310303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,10240,0.03518400192260742
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3072,8192,0.06300693353017171
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,8192,0.028615466753641766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,8192,0.02801919976870219
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3584,8192,0.02187626759211222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3072,7168,0.07711466948191324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3072,7168,0.0276309331258138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,7168,0.028703999519348145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,7168,0.025834665695826216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3072,6144,0.06711146831512452
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3072,6144,0.02439360022544861
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,6144,0.022759467363357544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,6144,0.022869332631429037
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3072,5120,0.05948266585667929
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3072,5120,0.02090239922205607
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,5120,0.01862506667772929
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,5120,0.019966934124628702
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3072,4096,0.050735998153686526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,4096,0.01548373301823934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,4096,0.016375466187795003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3072,3584,0.04783039887746175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3072,3584,0.015288533767064414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3584,3072,0.010557867089907328
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,3584,0.01649493376413981
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,3584,0.015147733688354491
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3072,3072,0.04362560113271077
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3072,3072,0.013532800475756326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,3072,0.008939733107884724
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,3072,0.013977600137392679
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3584,2560,0.009880533814430237
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3072,2560,0.04042453368504842
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,2560,0.007979733248551685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,2560,0.01228053371111552
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3072,2048,0.0372106671333313
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3072,2048,0.010513066252072652
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,2048,0.0070602665344874065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,2048,0.010890666643778484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3072,1536,0.03373653491338094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3072,1536,0.009050666292508443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3584,2048,0.00864746669928233
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,1536,0.00626453310251236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,1536,0.00970240036646525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3072,1024,0.030937600135803222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3072,1024,0.0073749333620071415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,1024,0.005191466708978017
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,1024,0.008451199531555176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3072,768,0.02908266584078471
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3072,768,0.006792533397674561
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,768,0.008113066852092742
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,768,0.004702933132648468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3072,512,0.027186133464177448
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,512,0.0042026668787002565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,512,0.007449600100517273
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3072,256,0.025321600834528606
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3072,256,0.005637333293755849
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,256,0.0037471999724706015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,256,0.006922666728496551
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3072,128,0.024358399709065757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3072,128,0.005277866621812185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,128,0.003433600068092346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,128,0.006647466619809468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3072,64,0.02408426602681478
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,64,0.0033952000240484873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,64,0.006908800204594929
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,3072,32,0.023964800437291465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,32,0.003505066782236099
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,32,0.007189333438873291
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2560,65536,0.37177066802978515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2560,65536,0.14608960151672362
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,65536,0.13735040028889972
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,65536,0.14521387418111165
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3072,65536,0.24945492744445802
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2560,16384,0.09857172966003418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2560,16384,0.036356266339619955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,16384,0.041076266765594484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,16384,0.039079467455546066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2560,12288,0.029242666562398274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2560,12288,0.07901973724365234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,12288,0.03340906699498494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,12288,0.029275733232498168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2560,10240,0.07084480126698812
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2560,10240,0.023830399910608927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,10240,0.029803733030954998
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,10240,0.02315946618715922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2560,8192,0.06200106541315714
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2560,8192,0.019793067375818887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,8192,0.025163733959198
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,8192,0.02077546715736389
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3072,12288,0.04842880169550578
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2560,7168,0.017998933792114258
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2560,7168,0.07632213433583578
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,7168,0.021783467133839926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,7168,0.019090133905410766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2560,6144,0.016051200032234193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2560,6144,0.0662069320678711
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,6144,0.01970240076382955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,6144,0.01685439944267273
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2560,5120,0.05906879901885986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2560,5120,0.013421866297721862
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,5120,0.016321067015329996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,5120,0.015027200182278952
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2560,4096,0.050443732738494874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3072,8192,0.030852266152699787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,4096,0.013619200388590495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,4096,0.012779733538627625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2560,3584,0.047543466091156006
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,3584,0.012005333105723064
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,3584,0.011915733416875202
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2560,3072,0.04358933369318645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,3072,0.011867733796437581
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,3072,0.011141332983970641
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2560,2560,0.04035840034484863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2560,2560,0.008894933263460796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,2560,0.007878399888674418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,2560,0.009958400328954061
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2560,2048,0.03696426550547282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2560,2048,0.007595733304818471
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,2048,0.006983466446399689
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,2048,0.009036800265312195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2560,1536,0.033614933490753174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2560,1536,0.007047466437021891
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3072,4096,0.01691306630770365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,1536,0.006090666850407918
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,1536,0.008291199803352356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2560,1024,0.03078826665878296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2560,1024,0.006113066772619883
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,1024,0.005109333495299021
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,1024,0.0073642666141192125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2560,768,0.028756266832351683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2560,768,0.0058559998869895935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,768,0.004655999938646952
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,768,0.007083733379840851
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2560,512,0.02720746596654256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2560,512,0.005559466779232025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3072,2560,0.012291199962298075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,512,0.004165333261092504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,512,0.006718933085600535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2560,256,0.025244800249735515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2560,256,0.005187200009822845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,256,0.0037920000652472176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,256,0.006374399860699971
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2560,128,0.024284799893697105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2560,128,0.004905599852403005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,128,0.0034656000634034475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,128,0.006140799820423126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2560,64,0.024024534225463866
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,64,0.0032768001159032187
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,64,0.006278400123119354
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2560,32,0.023959465821584067
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,32,0.0033258666594823206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,32,0.006533333162466686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3072,512,0.006177066763242086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2048,65536,0.14296426773071289
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,65536,0.1075711965560913
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,65536,0.137173334757487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2048,65536,0.37182931900024413
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2048,16384,0.0974783976872762
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2048,16384,0.03418879906336467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,16384,0.035076268513997394
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,16384,0.03546773195266724
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2048,12288,0.0782912015914917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2048,12288,0.02600533366203308
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,12288,0.028625067075093585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2048,10240,0.0700821320215861
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,12288,0.026499199867248534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2048,10240,0.02265920042991638
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,10240,0.025539199511210125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,10240,0.022215465704600014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2048,8192,0.062047998110453285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2048,8192,0.01785279909769694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,8192,0.023363200823465984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,8192,0.019462400674819948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2048,7168,0.07600853443145753
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2048,7168,0.01634666621685028
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,7168,0.020992000897725425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,7168,0.017605332533518474
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2048,6144,0.06621973514556885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2048,6144,0.014967466394106546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,6144,0.01811306675275167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,6144,0.015335466464360556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2048,5120,0.05774826606114706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2048,5120,0.012724266449610392
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,5120,0.01575146714846293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2048,4096,0.04985599915186564
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,5120,0.01343786617120107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2048,4096,0.011275733510653179
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,4096,0.013211733102798462
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,4096,0.011615999539693196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2048,3584,0.04676160017649333
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2048,3584,0.010468266407648722
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,3584,0.012086400389671325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,3584,0.011053867141405741
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2048,3072,0.04378453493118286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2048,3072,0.009324799974759419
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,3072,0.010865066448847454
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,3072,0.010356266299883525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2048,2560,0.04028586546579997
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2048,2560,0.008553600311279297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,2560,0.0076789334416389465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,2560,0.009341866771380106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2048,2048,0.036857601006825766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,2048,0.00647680014371872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,2048,0.008453333377838134
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2048,1536,0.03341653347015381
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2048,1536,0.006878933310508728
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,1536,0.005694933235645294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,1536,0.007795199751853943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2048,1024,0.03062719901402791
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,1024,0.004851200183232625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,1024,0.007218133409818013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2048,768,0.028743465741475422
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2048,768,0.005724800129731497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,768,0.004434133569399515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,768,0.0069013332327206925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2048,512,0.027113600571950273
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2048,512,0.005508266886075338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,512,0.00396373321612676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,512,0.006618666648864746
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2048,256,0.025225599606831867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,256,0.003626666714747747
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,256,0.006316799918810527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2048,128,0.02436800003051758
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2048,128,0.004949333270390829
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,128,0.003340800106525421
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2560,4096,0.011582932869593303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,128,0.006100266675154368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2048,64,0.024166399240493776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,64,0.0031615999837716425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,64,0.006259199976921081
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2560,3584,0.010790399710337321
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,2048,32,0.023439999421437582
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,32,0.0032885332902272543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,32,0.0064064001043637585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1536,65536,0.10342506567637126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2560,3072,0.010079999764760334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,65536,0.08513279755910239
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,65536,0.09453012943267822
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1536,16384,0.09863466421763102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1536,65536,0.3715082804361979
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1536,16384,0.02821439901987712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,16384,0.02866453329722087
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,16384,0.02340266704559326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1536,12288,0.07889599800109863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,12288,0.02240426739056905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,12288,0.02084160049756368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1536,10240,0.07007893721262613
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1536,10240,0.01924053430557251
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,10240,0.02667199969291687
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,10240,0.01801813244819641
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1536,8192,0.015187199910481772
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1536,8192,0.06103893518447876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,8192,0.021492266654968263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1536,7168,0.07530240217844644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,8192,0.015250133474667868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1536,7168,0.013797332843144735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,7168,0.018686934312184652
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1536,6144,0.06609066724777221
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,7168,0.0145087997118632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1536,6144,0.012494933605194092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,6144,0.017051732540130614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,6144,0.013128532965977987
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1536,5120,0.011262933413187664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1536,5120,0.057085867722829184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,5120,0.012221866846084594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,5120,0.014146133263905843
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1536,4096,0.009872000416119893
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1536,4096,0.050596264998118076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,4096,0.012174933155377706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,4096,0.010788266857465107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1536,3584,0.04658453464508057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1536,3584,0.009372799595197042
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,3584,0.01111466685930888
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,3584,0.010007466872533162
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1536,3072,0.04350719849268596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1536,3072,0.008603733777999879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,3072,0.00999679962793986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,3072,0.009393067161242167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1536,2560,0.040267733732859294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,2560,0.007231999933719635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,2560,0.008707200487454731
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1536,2048,0.036722131570180255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1536,2048,0.007341866691907247
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,2048,0.0061471998691558834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,2048,0.008123733103275299
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1536,1536,0.03365866740544637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,1536,0.005460266768932342
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,1536,0.007594666878382365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1536,1024,0.030753066142400105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,1024,0.004660266637802124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,1024,0.007101866602897644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1536,768,0.028518400589625043
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1536,768,0.005709866682688395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,768,0.004345599810282389
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,768,0.006913066903750102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1536,512,0.02709226608276367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1536,512,0.005373866856098175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,512,0.003922133396069208
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2048,2048,0.007586133480072021
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,512,0.0065760001540184024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1536,256,0.02524159948031108
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1536,256,0.005096533397833506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,256,0.003587199995915095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,256,0.006293333570162455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2048,1024,0.005985066791375478
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1536,128,0.024250666300455727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1536,128,0.004913066824277243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,128,0.0032277333239714304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,128,0.0060032000144322716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1536,64,0.023986132939656575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,64,0.0030933332939942675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,64,0.006142933170000712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1536,32,0.02383786638577779
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,32,0.006277333199977875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,32,0.003171200056870779
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,65536,0.06308799982070923
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1024,65536,0.19417813618977864
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2048,256,0.005130666494369507
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1024,65536,0.3709941228230794
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,65536,0.08594773610432943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1024,16384,0.09807039896647135
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1024,16384,0.04629546801249186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,16384,0.02292799949645996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,16384,0.02372373342514038
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1024,12288,0.07809066772460938
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1024,12288,0.037435734272003175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,12288,0.023113600413004556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,12288,0.019215999046961467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1024,10240,0.0695530652999878
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1024,10240,0.032459733883539836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,10240,0.019671465953191122
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,10240,0.016781866550445557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1024,8192,0.06002879937489828
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,8192,0.01693333387374878
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,8192,0.014452266693115234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1024,7168,0.0741919994354248
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1024,7168,0.0221781333287557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,7168,0.013514666755994161
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,7168,0.013843199610710144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1024,6144,0.06568426688512166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1024,6144,0.01967039903004964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,6144,0.012475732962290447
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,6144,0.011966933806737263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1024,5120,0.05666773319244385
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1024,5120,0.017173333962758382
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1536,12288,0.021347200870513915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,5120,0.01123199959595998
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,5120,0.011424000064531963
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1024,4096,0.049609601497650146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,4096,0.010121599833170573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,4096,0.010261332988739014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1024,3584,0.046453332901000975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1024,3584,0.012759466965993246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,3584,0.008909866213798523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,3584,0.009742933511734008
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1024,3072,0.04328960180282593
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,3072,0.00801813354094823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,3072,0.00906880001227061
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1024,2560,0.04012053410212199
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1024,2560,0.010315733154614766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,2560,0.007158400118350982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,2560,0.00895360012849172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1024,2048,0.03681813478469849
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1024,2048,0.009059199690818786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,2048,0.006513066589832306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,2048,0.00795413355032603
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1024,1536,0.03341653347015381
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1024,1536,0.0081535999973615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,1536,0.005547733108202616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,1536,0.007474133372306823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1024,1024,0.030536532402038574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1024,1024,0.006818133095900218
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,1024,0.004721066852410635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1536,2560,0.008155733346939087
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,1024,0.007010133564472198
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1024,768,0.028548266490300494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,768,0.004283733169237773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1536,1536,0.006740266581376393
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,768,0.006792533397674561
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1024,512,0.02699306607246399
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1024,512,0.005817600091298421
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,512,0.00383786658445994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,512,0.006493866443634033
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1536,1024,0.006062933305899302
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1024,256,0.02507839997609456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,256,0.0034656000634034475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,256,0.0061749334136645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1024,128,0.024306132396062215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,128,0.003201066702604294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,128,0.006037333110968272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1024,64,0.023909332354863484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,64,0.0030986666679382324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,64,0.0061482667922973635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,1024,32,0.023357866207758586
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,32,0.003155199935038885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,32,0.006151466568311056
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,768,65536,0.3710325241088867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,65536,0.05152426560719808
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,65536,0.07512426376342773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,768,16384,0.09837333361307779
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,768,16384,0.04315413236618042
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,16384,0.024105600516001382
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,16384,0.021193599700927733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,768,12288,0.07807892958323161
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,768,12288,0.034024532636006674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,12288,0.0172160009543101
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,12288,0.017387733856836955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,768,10240,0.06922986507415771
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,768,10240,0.029447466135025024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,10240,0.015446399648984274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,10240,0.01562346617380778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,768,8192,0.059480532010396325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,768,8192,0.02485439976056417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,8192,0.013157332936922709
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,8192,0.013621333241462707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,768,7168,0.07393386363983154
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,768,7168,0.02207466761271159
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,7168,0.012607999642690024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,7168,0.012915199995040894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,768,6144,0.0636629343032837
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,768,6144,0.01850773294766744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,6144,0.011558399597803751
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,6144,0.01188800036907196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,768,5120,0.056542932987213135
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,768,5120,0.016106667121251424
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,5120,0.010166399677594503
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,5120,0.011092266440391541
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1024,8192,0.02566186587015788
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,768,4096,0.04942506551742554
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,768,4096,0.014064000050226847
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,4096,0.00879253347714742
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,4096,0.010075733065605164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,768,3584,0.046359467506408694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,768,3584,0.012724266449610392
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,3584,0.00805866668621699
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,3584,0.009569066762924194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,768,3072,0.04326719840367635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,3072,0.007412266731262207
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,3072,0.00904319981733958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,768,2560,0.0400000015894572
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1024,4096,0.013945600390434265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,768,2560,0.00965119997660319
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,2560,0.006647466619809468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,2560,0.00843946635723114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,768,2048,0.03687893152236939
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,768,2048,0.00870293378829956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,2048,0.005965866645177205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,2048,0.007863466441631318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,768,1536,0.0338485320409139
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,768,1536,0.007812266548474629
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,1536,0.005196799834569296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,1536,0.007421866556008657
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,768,1024,0.030585600932439165
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1024,3072,0.011450666189193725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,768,1024,0.006584533552328746
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,1024,0.004487466812133789
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,1024,0.007000533243020375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,768,768,0.028597333033879596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,768,768,0.006011733412742614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,768,0.004198400179545084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,768,0.006723199784755707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,768,512,0.02691733241081238
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,768,512,0.005765333275000254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,512,0.003819733361403147
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,512,0.006403199831644694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,768,256,0.02531733314196269
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,256,0.003455999990304311
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,256,0.006215466558933258
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,768,128,0.024436267217000325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,768,128,0.0049333333969116214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,128,0.003139200061559677
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,128,0.006025599936644236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,768,64,0.024101332823435465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,64,0.003019733230272929
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,64,0.006055466830730438
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,768,32,0.02393280069033305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,32,0.0030410667260487873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,32,0.006111999849478403
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1024,768,0.006284800171852112
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,512,65536,0.3714666684468587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,512,65536,0.07747306823730468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,65536,0.039844266573588055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1024,256,0.00528106689453125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,65536,0.08068693478902181
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,512,16384,0.09809707005818685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,16384,0.015620266397794088
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1024,128,0.004862933357556661
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,16384,0.021000534296035767
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,512,12288,0.0775935967763265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,12288,0.014588800072669984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,12288,0.017461333672205606
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,512,10240,0.06886719862620036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,512,10240,0.015520000457763672
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,10240,0.013489066561063131
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,10240,0.015497600038846334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,512,8192,0.05922773281733194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,512,8192,0.013035733501116434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,8192,0.011901866396268208
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,8192,0.013924266894658408
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,512,7168,0.07194666862487793
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,7168,0.012769066294034324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,7168,0.01048533320426941
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,512,6144,0.06383253335952759
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,512,6144,0.010845866799354554
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,6144,0.010334933797518413
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,6144,0.011749333143234253
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,512,5120,0.05596586863199869
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,5120,0.009521067142486572
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,768,65536,0.1516106605529785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,5120,0.01092693308989207
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,512,4096,0.04967360099156697
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,4096,0.008338133494059246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,4096,0.00993066628774007
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,512,3584,0.046616534392038986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,512,3584,0.008388266960779826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,3584,0.008081066608428954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,3584,0.009447466333707173
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,512,3072,0.043246932824452716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,512,3072,0.007814399898052216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,3072,0.007709866762161255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,3072,0.008959999680519104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,512,2560,0.03973759810129802
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,768,3072,0.011222400267918905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,2560,0.006426666676998138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,2560,0.008342400193214417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,512,2048,0.0365664005279541
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,512,2048,0.0066655998428662615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,2048,0.005694933235645294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,2048,0.007786666850248973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,512,1536,0.03309760093688965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,768,256,0.0052138666311899815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,1536,0.005042133231957754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,1536,0.007354666789372762
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,512,1024,0.03026240070660909
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,512,1024,0.0056757330894470215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,1024,0.004349866509437561
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,1024,0.0068800002336502075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,512,768,0.02839999993642171
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,512,768,0.005364266534646353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,768,0.0040394666294256846
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,768,0.006679466863473256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,512,512,0.026588799556096394
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,512,512,0.005093333125114441
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,512,0.0036533333361148832
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,512,0.006297599772612255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,512,256,0.024924800793329874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,512,256,0.004837333162625631
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,256,0.0033386667569478357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,256,0.006065066655476888
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,512,128,0.024050132433573405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,512,128,0.0046410664916038515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,128,0.003068800022204717
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,128,0.00588266650835673
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,512,16384,0.021640533208847047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,512,64,0.023758933941523234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,64,0.0029845332105954488
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,64,0.005937066674232483
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,512,32,0.022923733790715536
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,32,0.0029493334392706556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,32,0.005964800218741099
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,512,12288,0.017526400089263917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,256,65536,0.07732693354288736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,65536,0.029167999823888142
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,256,65536,0.37179520924886067
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,65536,0.07569920221964518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,256,16384,0.09758613109588624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,256,16384,0.021333332856496176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,16384,0.010680533448855082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,16384,0.02095253268877665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,512,7168,0.012071466445922852
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,512,5120,0.009839999675750732
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,256,12288,0.07742400169372558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,256,12288,0.01741546591122945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,12288,0.009341866771380106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,12288,0.017170133193333943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,10240,0.008825600147247314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,256,10240,0.06805439790089926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,10240,0.015503999590873719
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,256,8192,0.05928959846496582
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,256,8192,0.0128330667813619
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,512,4096,0.008935466408729553
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,8192,0.007947733501593272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,8192,0.01358080009619395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,256,7168,0.07145386536916097
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,256,7168,0.01199893355369568
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,7168,0.007608533402283986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,7168,0.01274133324623108
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,256,6144,0.06250880161921182
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,6144,0.007322666545708974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,6144,0.01184213360150655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,256,5120,0.05590720176696777
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,5120,0.00824426660935084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,5120,0.010985599954922994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,256,4096,0.04903680086135864
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,512,2560,0.007195733487606049
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,4096,0.007787733276685078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,4096,0.009801600376764934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,256,3584,0.04586986700693767
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,256,3584,0.008269866804281871
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,3584,0.007518933216730754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,3584,0.009367466966311137
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,256,3072,0.042829867204030356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,256,3072,0.007602133353551229
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,3072,0.007204266885916392
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,3072,0.008919466535250347
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,256,2560,0.039426132043202715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,2560,0.006379733482996623
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,2560,0.008312533299128216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,256,2048,0.036296534538269046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,256,2048,0.006612266600131989
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,2048,0.005737600227197012
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,2048,0.007774933179219563
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,256,1536,0.03281066616376241
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,256,1536,0.006097066899140676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,1536,0.00505920002857844
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,1536,0.007292800148328145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,256,1024,0.030010666449864703
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,256,1024,0.005576533575852713
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,1024,0.004318933188915253
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,1024,0.0068800002336502075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,256,768,0.028155734141667683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,256,768,0.005285333096981049
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,768,0.004018133382002512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,768,0.006677333513895671
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,256,512,0.026405332485834758
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,256,512,0.005077333251635233
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,512,0.003673599908749262
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,512,0.006270933151245117
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,256,256,0.024959999322891235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,256,256,0.00476800004641215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,256,0.003340800106525421
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,256,0.006061866879463196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,256,128,0.024094933271408082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,256,128,0.0046293333172798155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,128,0.003142400085926056
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,128,0.005907199780146281
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,256,64,0.023780266443888344
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,64,0.0029056000212828318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,64,0.005967999994754791
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,256,32,0.022848000129063926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,32,0.0029834667841593427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,32,0.005942399799823761
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,128,65536,0.07737920284271241
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,65536,0.019048533837000527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,128,65536,0.3704575856526693
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,65536,0.07298666636149088
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,128,16384,0.021363200743993123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,128,16384,0.09716053009033203
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,16384,0.009168000022570292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,16384,0.02085226575533549
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,128,12288,0.07713812987009684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,128,12288,0.017409066359202065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,12288,0.008350933591524761
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,12288,0.017110399405161538
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,512,1536,0.006155733267466227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,128,10240,0.06770239671071371
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,10240,0.0077354664603869125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,10240,0.01530239979426066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,128,8192,0.057608532905578616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,128,8192,0.012753066420555115
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,8192,0.007449600100517273
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,8192,0.013326932986577352
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,128,7168,0.07174719969431559
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,128,7168,0.011753599842389424
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,7168,0.007302399973074596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,7168,0.012635733683904013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,128,6144,0.062370133399963376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,128,6144,0.01069546639919281
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,6144,0.006968533496061961
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,6144,0.011643733580907185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,128,5120,0.055820798873901366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,128,5120,0.00962666670481364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,5120,0.007380266487598419
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,5120,0.010863999525705974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,128,4096,0.04907093445460002
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,4096,0.007079466680685679
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,4096,0.009641599655151368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,128,3584,0.04577920039494832
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,128,3584,0.008241066833337148
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,3584,0.007472000022729237
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,3584,0.009164800246556599
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,128,3072,0.04260480006535848
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,128,3072,0.007601066430409749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,3072,0.007101866602897644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,3072,0.008679466446240743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,128,2560,0.039418665568033855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,128,2560,0.007017600039641063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,2560,0.006289066871007283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,2560,0.008096000055472057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,128,2048,0.03594880104064942
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,128,2048,0.006524799764156342
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,2048,0.0055861334005991616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,2048,0.007546666761239369
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,128,1536,0.032858665784200033
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,128,1536,0.006110933423042297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,1536,0.0049002667268117275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,1536,0.007067733506361644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,128,1024,0.02982826630274455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,128,1024,0.005496533215045929
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,1024,0.0042463997999827065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,1024,0.006723199784755707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,128,768,0.027815467119216918
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,128,768,0.005364266534646353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,768,0.0039007999002933502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,768,0.006528000036875407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,128,512,0.026258132855097455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,128,512,0.004965333143870035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,512,0.0035616000493367515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,512,0.006213333209355672
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,128,256,0.024794665972391765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,256,10240,0.015289599696795145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,256,0.003223466624816259
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,256,0.0059797331690788266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,128,128,0.023891200621922813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,128,128,0.004574933151404063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,128,0.003045333425203959
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,128,0.005852800110975901
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,256,5120,0.009742933511734008
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,128,64,0.023515733083089192
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,64,0.0029130667448043824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,64,0.005861333509286245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,256,4096,0.008727467060089112
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,32,0.0029098667204380036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,128,32,0.02291733423868815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,32,0.005846400062243144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,65536,0.013379200299580892
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,256,6144,0.010868266224861145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,65536,0.0701258659362793
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,16384,0.0077450667818387345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,64,16384,0.09755626519521078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,64,65536,0.3703637440999349
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,16384,0.02066346605618795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,12288,0.007454933226108551
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,64,12288,0.07764480113983155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,12288,0.01686506668726603
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,64,10240,0.06719573338826498
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,10240,0.0073066666722297665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,10240,0.015217066804567973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,64,8192,0.05711359977722168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,8192,0.0073642666141192125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,8192,0.013271466890970866
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,7168,0.007067733506361644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,64,7168,0.07110826969146729
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,7168,0.012442666292190551
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,64,6144,0.06272000074386597
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,6144,0.0069013332327206925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,256,2560,0.007057066758473713
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,6144,0.01143999993801117
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,64,5120,0.05591573317845663
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,5120,0.007331199944019318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,5120,0.010565333565076192
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,64,4096,0.048682665824890135
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,4096,0.006900266806284587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,4096,0.009493333101272584
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,64,3584,0.04570560057957967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,3584,0.007320533196131389
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,3584,0.009019733468691508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,64,3072,0.04314026832580566
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,3072,0.007004799942175548
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,3072,0.008644266923268636
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,64,2560,0.03913493156433105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,2560,0.006275199850400289
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,2560,0.0081045334537824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,64,2048,0.035807998975118
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,2048,0.005527466535568237
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,2048,0.0076000000039736434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,64,1536,0.03261973261833191
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,1536,0.004913066824277243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,1536,0.007117866476376851
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,64,1024,0.029796266555786134
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,1024,0.004250666499137879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,1024,0.006760533154010773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,64,768,0.02772480050722758
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,768,0.003912533322970072
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,768,0.0063967997829119366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,64,512,0.026099199056625368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,512,0.003550933301448822
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,512,0.006212266782919565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,64,256,0.024734934171040855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,256,0.0032277333239714304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,256,0.005967999994754791
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,64,128,0.024170666933059692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,128,0.003028266628583272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,128,0.005804799993832906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,64,64,0.02333866755167643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,64,0.002924799919128418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,64,0.005847466488679251
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,64,32,0.022924800713857017
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,32,0.0028575999041398365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,32,0.00594346672296524
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,65536,0.01317759950955709
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,65536,0.06804587046305338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,32,65536,0.3715157190958659
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,16384,0.00729066679875056
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,32,16384,0.09757226308186849
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,16384,0.020827732483545938
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,12288,0.0070816000302632645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,32,12288,0.0770143985748291
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,12288,0.01700693368911743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,32,10240,0.06659200191497802
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,10240,0.007288533449172974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,10240,0.015129599968592325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,32,8192,0.05647146701812744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,8192,0.007191466788450877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,8192,0.013225600123405457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,128,10240,0.015238400300343832
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,7168,0.00695253312587738
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,32,7168,0.07167786757151286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,7168,0.012376532951990763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,5120,0.007186133166154225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,32,6144,0.06237333218256632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,6144,0.006795733173688252
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,6144,0.011430399616559346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,32,5120,0.05591573317845663
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,5120,0.010738133390744527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,32,4096,0.04910720189412435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,4096,0.006863999863465626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,4096,0.00951039989789327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,32,3584,0.04592853387196859
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,3584,0.007259733478228251
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,3584,0.009147733449935913
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,32,3072,0.042903467019399004
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,3072,0.006817066669464111
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,3072,0.008565333485603333
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,32,2560,0.039383467038472494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,2560,0.006223999957243601
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,2560,0.008132266501585644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,32,2048,0.03602666854858398
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,2048,0.005512533088525137
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,2048,0.007474133372306823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,32,1024,0.029783467451731365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,1536,0.004911999901135763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,32,1536,0.03306986689567566
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,128,4096,0.00862613320350647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,1536,0.007030400137106578
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,1024,0.004171733558177948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,1024,0.006759466727574666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,32,768,0.027980800469716387
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,768,0.0038453333079814913
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,768,0.006434133152167003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,32,512,0.026320000489552815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,512,0.003570133447647095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,512,0.006230400005976359
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,128,0.005767466624577841
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,32,256,0.024621866146723428
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,256,0.0031829332311948144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,64,0.005840000013510386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,256,0.006007466713587443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,32,128,0.023861332734425863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,128,0.0029685333371162414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,32,64,0.02335360050201416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,64,0.0028575999041398365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,160,32,32,0.022690133253733317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,32,0.002856533229351044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,32,0.005904000004132589
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,65536,16384,0.4026048024495442
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,65536,16384,0.4039786656697591
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,65536,16384,0.8195669174194335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,65536,16384,0.4570773442586263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,65536,12288,0.38238614400227866
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,65536,12288,0.5885781606038412
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,65536,12288,0.3161888122558594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,65536,10240,0.25504533449808753
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,65536,10240,0.2637621402740479
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,65536,8192,0.2485055923461914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,65536,10240,0.3131701469421387
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,65536,8192,0.20734613736470542
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,65536,10240,0.48801492055257156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,65536,8192,0.2038719971974691
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,65536,8192,0.3943711916605631
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,128,256,0.0046741331617037455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,65536,7168,0.25045973459879556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,65536,7168,0.34338347117106116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,65536,7168,0.1797760009765625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,65536,6144,0.15063573519388834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,65536,6144,0.22911680539449059
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,65536,6144,0.1564095973968506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,65536,6144,0.2997023900349935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,65536,5120,0.18720320065816243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,65536,5120,0.12853653430938722
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,65536,5120,0.13082773685455323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,65536,5120,0.2508394718170166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,65536,4096,0.10752747058868409
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,65536,4096,0.15115413665771485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,65536,4096,0.10866560141245525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,65536,4096,0.20181760787963868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,65536,3584,0.09524266719818116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,65536,3584,0.1381589412689209
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,65536,3584,0.178330659866333
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,65536,3584,0.09440853595733642
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,65536,3072,0.08369812965393067
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,65536,3072,0.15476479530334472
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,65536,3072,0.12961920102437335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,65536,3072,0.08182400067647298
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,65536,2560,0.07220053672790527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,65536,2560,0.11279679934183758
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,65536,2560,0.07075626850128174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,65536,2560,0.13057493368784587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,65536,2048,0.06019733349482218
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,65536,2048,0.10725759665171306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,65536,2048,0.10048000017801921
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,65536,2048,0.05877120097478231
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,65536,1536,0.08604053656260172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,65536,1536,0.0481877326965332
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,65536,1536,0.08298239707946778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,65536,1536,0.04733333190282186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,65536,1024,0.07183679739634195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,65536,1024,0.05951040188471476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,65536,1024,0.0363317330678304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,65536,768,0.06494719982147217
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,65536,768,0.029740800460179646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,65536,768,0.04641493161519368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,65536,768,0.032654933134714764
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,65536,512,0.05909333229064941
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,65536,512,0.0288266658782959
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,65536,512,0.03450239896774292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,65536,256,0.012403200070063274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,65536,256,0.0501418670018514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,65536,256,0.01909439961115519
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,65536,256,0.01993280053138733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,65536,128,0.040106666088104245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,65536,128,0.01053546667098999
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,65536,128,0.009847467144330341
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,65536,128,0.016577066977818807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,65536,64,0.03978026707967122
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,65536,64,0.02304746707280477
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,65536,64,0.008206933240095774
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,65536,32,0.03975253502527873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,65536,32,0.00872213343779246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,65536,32,0.025490132967631023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,16384,65536,0.40021333694458006
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,65536,0.41662079493204757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,65536,0.8460863749186197
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,16384,65536,0.5849685033162435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,16384,16384,0.10736213525136311
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,16384,16384,0.15173226992289227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,65536,12288,0.2850133260091146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,16384,0.19094079335530598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,16384,12288,0.07828053633371988
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,16384,0.1124725341796875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,16384,12288,0.12791893482208253
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,12288,0.14648319880167643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,12288,0.08624640305836996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,16384,10240,0.06970880031585694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,16384,10240,0.1096000035603841
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,10240,0.12298773129781086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,10240,0.07131199836730957
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,16384,8192,0.05794560114542643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,16384,8192,0.09333013693491618
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,8192,0.05744106769561767
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,8192,0.09841386477152506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,16384,7168,0.1024725357691447
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,16384,7168,0.05133119821548462
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,7168,0.08699093659718832
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,7168,0.050987732410430905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,16384,6144,0.09275306860605875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,6144,0.044700801372528076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,6144,0.08124372959136963
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,16384,5120,0.08170026938120524
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,16384,5120,0.037656533718109134
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,65536,7168,0.17252906163533527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,5120,0.06835520267486572
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,5120,0.03893866539001465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,16384,4096,0.07006933689117431
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,16384,4096,0.03201919992764791
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,4096,0.05551040172576904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,4096,0.03242986599604289
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,16384,3584,0.06487253506978354
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,3584,0.049428268273671465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,3584,0.029303467273712157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,16384,3072,0.05957653522491455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,16384,3072,0.026717867453893023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,3072,0.041365333398183185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,3072,0.026257065931955974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,16384,2560,0.053795198599497475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,2560,0.0399402658144633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,2560,0.022974934180577597
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,16384,2048,0.013187199831008911
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,16384,2048,0.049158398310343424
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,2048,0.02940906683603923
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,2048,0.015929599603017174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,16384,1536,0.039236267407735184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,16384,1536,0.010405332843462626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,1536,0.02360639969507853
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,1536,0.01307413379351298
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,16384,1024,0.03399786551793416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,1024,0.010795733332633973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,1024,0.011095466216405232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,16384,768,0.030477867523829145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,16384,768,0.008281599978605907
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,768,0.008945066730181377
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,768,0.010089600086212158
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,16384,512,0.028640000025431316
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,65536,1024,0.035884801546732584
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,16384,512,0.00780266672372818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,512,0.0071285332242647815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,512,0.009425066908200582
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,16384,256,0.026936533053716023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,16384,256,0.006702933212121327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,256,0.005713066458702088
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,256,0.00871573289235433
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,16384,128,0.02532373269399007
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,16384,128,0.006340266764163971
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,128,0.004817066589991251
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,128,0.008258133133252462
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,16384,64,0.02563733259836833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,65536,512,0.023912533124287923
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,64,0.004382933179537455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,64,0.009574400385220845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,16384,32,0.02537920077641805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,32,0.010175999999046326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,32,0.004818133513132731
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,65536,0.3183637301127116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,65536,0.5921642939249675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,12288,65536,0.501363213857015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,12288,16384,0.13518400192260743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,12288,16384,0.08337173461914063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,16384,0.0843061367670695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,16384,0.15242986679077147
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,12288,12288,0.06394773324330648
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,12288,12288,0.1102677345275879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,12288,0.06487893263498942
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,12288,0.1254815975824992
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,12288,10240,0.09761706988016763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,10240,0.09224212964375814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,10240,0.05574933290481567
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,12288,8192,0.08200000127156576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,12288,8192,0.04546773433685303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,8192,0.07400746345520019
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,8192,0.045517865816752115
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,12288,7168,0.040377600987752275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,12288,7168,0.09399893283843994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,7168,0.0655135989189148
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,7168,0.04020479917526245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,12288,6144,0.03567786614100139
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,12288,6144,0.08435306549072266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,6144,0.057861332098642984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,6144,0.03565333286921184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,12288,5120,0.03087573250134786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,12288,5120,0.07477973302205404
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,5120,0.049080534776051836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,5120,0.03135040005048116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,12288,4096,0.06470079819361368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,12288,4096,0.026002132892608644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,4096,0.040667732556660965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,16384,6144,0.043891199429829914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,4096,0.026265599330266315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,12288,3584,0.023785599072774253
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,12288,3584,0.06019413471221924
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,3584,0.03583360115687052
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,3584,0.02373759945233663
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,12288,3072,0.055325865745544434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,3072,0.03155519962310791
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,3072,0.01849493384361267
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,12288,2560,0.04717439810434977
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,12288,2560,0.013124266266822815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,2560,0.027180800835291546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,16384,3584,0.029154133796691895
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,2560,0.015615999698638916
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,12288,2048,0.04262933333714803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,2048,0.023414399226506552
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,16384,2560,0.02294720013936361
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,2048,0.013146666685740152
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,12288,1536,0.03731946547826131
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,1536,0.01362986663977305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,1536,0.011307733257611592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,12288,1024,0.03184213240941365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,16384,1024,0.00879039963086446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,1024,0.008749866485595703
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,1024,0.009704533219337463
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,12288,768,0.030057599147160847
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,12288,768,0.007478400071461995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,768,0.007816533247629803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,768,0.008968533078829447
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,12288,512,0.02826133370399475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,12288,65536,0.29933013916015627
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,512,0.006358399987220764
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,512,0.008538666367530822
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,12288,256,0.0269269327322642
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,256,0.005226666728655497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,256,0.008026666442553202
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,12288,128,0.025778132677078246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,12288,128,0.005805866420269012
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,128,0.0042698666453361515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,128,0.007565866907437642
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,12288,64,0.02560960054397583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,12288,10240,0.05391573508580526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,64,0.003964799890915553
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,64,0.00824533353249232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,12288,32,0.025596799453099568
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,32,0.004388266801834106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,32,0.008798933029174805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,10240,65536,0.25997440020243323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,10240,65536,0.4659903844197591
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,65536,0.5001301447550456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,65536,0.3108757336934408
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,10240,16384,0.12637440363566083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,10240,16384,0.07840747038523356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,16384,0.13079573313395182
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,16384,0.0806272029876709
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,10240,12288,0.055783466498057044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,10240,12288,0.10438186327616375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,12288,3072,0.01893333395322164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,12288,0.09303253491719564
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,12288,0.06192320187886556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,10240,0.07825067043304443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,10240,10240,0.09297920068105062
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,10240,0.05244693358739218
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,10240,8192,0.03898986577987671
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,12288,2048,0.010994133353233338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,10240,8192,0.07794346809387206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,8192,0.06325973272323608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,8192,0.042558932304382326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,10240,7168,0.03575786749521891
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,10240,7168,0.09075199762980143
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,12288,1536,0.009629866480827332
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,7168,0.055947732925415036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,7168,0.037562668323516846
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,10240,6144,0.08166720072428385
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,6144,0.04907413323720296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,6144,0.033745066324869795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,12288,1024,0.008218666911125183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,10240,5120,0.0725717306137085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,5120,0.041705600420633954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,5120,0.02966826756795247
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,10240,4096,0.0628554662068685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,10240,4096,0.023605332771937052
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,4096,0.03418240149815877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,4096,0.023188267151514688
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,10240,3584,0.017426133155822754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,10240,3584,0.05834986766179403
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,3584,0.030610134204228718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,3584,0.018525866667429607
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,10240,3072,0.05178346633911133
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,3072,0.027211733659108478
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,3072,0.016126933693885803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,12288,512,0.0068341334660847975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,10240,2560,0.04518506526947021
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,10240,2560,0.011476266384124755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,2560,0.01413653294245402
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,2560,0.023348265886306764
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,10240,2048,0.010072533289591472
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,10240,2048,0.04057173331578572
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,2048,0.019257599115371705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,2048,0.012353066603342693
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,10240,1536,0.03514773448308309
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,1536,0.011038933197657268
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,1536,0.01114453375339508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,10240,1024,0.03165973424911499
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,10240,1024,0.00766293356815974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,1024,0.008550399541854858
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,1024,0.009603200356165568
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,10240,768,0.0298634668191274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,10240,768,0.0070271998643875126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,768,0.007382399837176005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,768,0.008919466535250347
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,10240,512,0.028169600168863933
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,10240,512,0.006443733473618825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,512,0.00592853327592214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,512,0.008503466844558716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,10240,256,0.026394667228062947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,10240,256,0.005857066810131073
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,256,0.0048543999592463175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,12288,256,0.0060917332768440245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,256,0.007878399888674418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,10240,128,0.02518613338470459
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,10240,128,0.005525333185990652
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,128,0.004110933343569437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,128,0.0075530668099721265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,10240,64,0.02502826650937398
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,64,0.0038165333370367683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,64,0.008203733464082081
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,10240,32,0.024817067384719848
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,32,0.00420053352912267
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,32,0.008774399757385254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,8192,65536,0.2005631923675537
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,8192,65536,0.372546132405599
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,65536,0.42009280522664383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,65536,0.21459627151489258
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,8192,16384,0.10443413257598877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,8192,16384,0.058633601665496825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,16384,0.05977173248926798
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,16384,0.11345813274383545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,8192,12288,0.044609065850575766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,8192,12288,0.08697600364685058
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,12288,0.07452053229014079
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,12288,0.047006932894388835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,8192,10240,0.03793600002924601
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,8192,10240,0.077348264058431
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,10240,0.06334506670633952
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,10240,0.04047040144602458
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,8192,8192,0.03145920038223267
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,8192,8192,0.06512213150660197
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,8192,0.050817068417867026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,8192,0.033291733264923094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,8192,7168,0.0789514700571696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,8192,7168,0.02835200031598409
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,7168,0.045330135027567546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,7168,0.02959360082944234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,8192,6144,0.07166933218638102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,8192,6144,0.02537493308385213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,6144,0.04020693302154541
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,6144,0.02672959963480632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,8192,5120,0.06378133296966552
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,10240,10240,0.047806934515635176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,5120,0.03376533190409343
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,5120,0.022935465971628825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,8192,4096,0.05445546706517538
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,8192,4096,0.013770666718482972
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,4096,0.028411734104156493
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,4096,0.014876799782117209
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,8192,3584,0.04582080046335856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,8192,3584,0.012666666507720947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,3584,0.014037332932154336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,3584,0.027002666393915815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,8192,3072,0.040752001603444415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,3072,0.022282665967941283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,3072,0.01259519954522451
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,8192,2560,0.036294400691986084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,8192,2560,0.010356266299883525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,10240,6144,0.031548800071080525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,2560,0.01749546726544698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,2560,0.011518933375676473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,8192,2048,0.03301333387692769
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,8192,2048,0.00906880001227061
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,10240,5120,0.02725759943326314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,2048,0.009860266248385112
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,2048,0.010212266445159912
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,8192,1536,0.03059733311335246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,1536,0.008280533552169799
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,1536,0.009432533383369445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,8192,1024,0.02829013268152873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,8192,1024,0.007036800185839335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,1024,0.006365866462389629
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,1024,0.008330666522185007
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,8192,768,0.027229867378870648
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,10240,3072,0.012594133615493774
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,768,0.005589333176612854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,768,0.007811200122038524
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,8192,512,0.025282132625579833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,512,0.005020800232887268
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,512,0.007335466643174489
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,8192,256,0.024461867411931355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,10240,1536,0.008945066730181377
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,256,0.004265599946180979
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,256,0.006960000097751618
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,8192,128,0.023463465770085654
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,8192,128,0.005220266679922739
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,128,0.0038453333079814913
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,128,0.006677333513895671
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,8192,64,0.02321386734644572
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,64,0.0034815999368826545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,64,0.007526400188604991
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,8192,32,0.02272426684697469
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,32,0.003881600002447764
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,32,0.0076341331005096436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,7168,65536,0.3477706591288249
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,7168,65536,0.18342612584431967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,65536,0.36860373814900715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,65536,0.22098347345987954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,7168,16384,0.09864640235900879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,7168,16384,0.05233706633249918
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,16384,0.09557013511657715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,16384,0.05731093486150106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,7168,12288,0.04034880002339681
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,7168,12288,0.08239253362019858
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,12288,0.04360213279724121
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,12288,0.07602240244547526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,7168,10240,0.07409706910451254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,7168,10240,0.034805333614349364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,10240,0.037484800815582274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,10240,0.056608001391092934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,7168,8192,0.062257067362467444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,8192,0.0454367995262146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,8192,0.030998400847117108
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,8192,5120,0.021618133783340453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,7168,7168,0.07634560267130533
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,7168,7168,0.02616106669108073
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,7168,0.040703999996185306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,7168,0.027741867303848266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,7168,6144,0.022962133089701332
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,7168,6144,0.06904106934865316
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,6144,0.02523306608200073
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,6144,0.03600213527679443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,7168,5120,0.061484801769256595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,7168,5120,0.01681706706682841
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,5120,0.03067520062128703
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,5120,0.018770132462183634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,8192,3072,0.01144533356030782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,7168,4096,0.04867306550343831
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,4096,0.024921600023905435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,4096,0.01508799990018209
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,7168,3584,0.012785067160924276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,7168,3584,0.04374186595280965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,3584,0.02230186661084493
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,3584,0.013022933403650919
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,7168,3072,0.03927040100097656
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,7168,3072,0.01164479951063792
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,3072,0.01960853338241577
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,3072,0.012556800246238708
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,7168,2560,0.0365664005279541
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,2560,0.011573333541552227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,2560,0.011794132987658183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,8192,1536,0.008220799763997396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,7168,2048,0.03384213447570801
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,2048,0.009148800373077392
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,2048,0.010263466835021972
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,8192,768,0.006562133133411407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,7168,1536,0.03095146616299947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,8192,512,0.006164266665776571
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,1536,0.007607466479142506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,1536,0.009318400422732036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,7168,1024,0.028934399286905926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,7168,1024,0.007534933090209961
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,1024,0.006177066763242086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,1024,0.008358400066693623
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,8192,256,0.005498666564623515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,7168,768,0.027133866151173906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,768,0.005274666845798493
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,768,0.00793280005455017
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,7168,512,0.025101866324742633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,7168,512,0.006389333307743073
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,512,0.004624000191688538
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,512,0.007397333284219106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,7168,256,0.02372693419456482
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,7168,256,0.005747200051943461
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,256,0.00403413325548172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,256,0.007051733136177063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,7168,128,0.02281173268953959
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,7168,128,0.005388799806435903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,128,0.0035775999228159585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,128,0.006799999872843425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,7168,64,0.022832000255584718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,64,0.003370666752258936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,64,0.0072170664866765336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,7168,32,0.023076266050338745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,32,0.0037237333754698435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,32,0.007603199779987335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,6144,65536,0.18467520078023275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,6144,65536,0.32570133209228513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,65536,0.30893014272054037
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,65536,0.20094827016194663
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,6144,16384,0.09357759952545167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,6144,16384,0.051539198557535804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,16384,0.08293866316477458
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,16384,0.053515732288360596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,6144,12288,0.04098453521728516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,6144,12288,0.07832213242848715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,12288,0.05702079931894938
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,12288,0.042701868216196696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,6144,10240,0.07056639989217123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,6144,10240,0.035019731521606444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,10240,0.04851839939753215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,10240,0.03731199900309245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,6144,8192,0.05926826794942221
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,6144,8192,0.029269333680470782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,8192,0.038819201787312824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,8192,0.029755733410517377
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,6144,7168,0.07358720302581787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,6144,7168,0.026345600684483845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,7168,0.03486933310826619
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,7168,0.02714879910151164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,7168,8192,0.02900693416595459
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,6144,6144,0.06632533470789591
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,6144,0.0313045342763265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,6144,0.021385600169499717
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,6144,5120,0.05851200024286905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,6144,5120,0.012968533237775168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,5120,0.026766933997472125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,5120,0.017489065726598106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,6144,4096,0.0464842677116394
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,6144,4096,0.011179733276367187
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,4096,0.021912533044815063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,7168,4096,0.013646933436393737
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,4096,0.013868799805641175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,6144,3584,0.042667734622955325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,6144,3584,0.010557867089907328
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,3584,0.017670400937398276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,3584,0.012711466352144877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,6144,3072,0.03854933182398478
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,3072,0.011442133784294128
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,3072,0.01199893355369568
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,6144,2560,0.03576426506042481
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,6144,2560,0.009268266956011455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,7168,2560,0.010733866691589355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,2560,0.00974826713403066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,2560,0.01120746632417043
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,6144,2048,0.03262826601664225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,6144,2048,0.008214400211970011
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,2048,0.00829013337691625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,2048,0.010247466961542766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,6144,1536,0.007484800120194752
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,6144,1536,0.030392533540725707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,7168,2048,0.009634133179982502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,1536,0.0068234667181968685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,1536,0.009372799595197042
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,6144,1024,0.028117332855860395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,1024,0.0055637334783871974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,1024,0.00814933329820633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,7168,1536,0.008429867029190064
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,6144,768,0.02651306589444478
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,6144,768,0.006177066763242086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,768,0.00524479995171229
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,7168,768,0.006883200009663899
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,768,0.00775786687930425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,6144,512,0.025257599353790284
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,6144,512,0.005763199925422668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,512,0.004549333453178405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,256,0.003920000046491623
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,512,0.007439999779065449
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,6144,256,0.02375040054321289
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,256,0.007045333087444305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,6144,128,0.023014400402704874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,128,0.003788800040880839
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,128,0.0067210664351781205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,6144,64,0.022860799233118692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,64,0.003268266717592875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,64,0.007249066730340321
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,6144,32,0.02271040081977844
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,32,0.0036544000109036768
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,32,0.007572266459465027
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,65536,0.26580373446146643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,5120,65536,0.3079327901204427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,65536,0.19832639694213866
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,5120,16384,0.08975786368052165
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,5120,16384,0.040779733657836915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,16384,0.070360533396403
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,16384,0.051710931460062656
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,5120,12288,0.03167146643002828
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,5120,12288,0.07479039827982584
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,12288,0.057323733965555825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,12288,0.040651734670003256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,5120,10240,0.06732266743977865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,5120,10240,0.027139200766881304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,10240,0.049073068300882976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,10240,0.035813331604003906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,5120,8192,0.02267520030339559
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,5120,8192,0.05656533241271973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,8192,0.033896533648173015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,8192,0.02890239953994751
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,5120,7168,0.07228586673736573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,5120,7168,0.01835626761118571
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,7168,0.030422399441401165
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,7168,0.022754132747650146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,5120,6144,0.015077333648999533
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,5120,6144,0.06459413369496664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,6144,0.02693439920743306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,6144,0.018195199966430663
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,5120,5120,0.05548799832661947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,5120,5120,0.01332373321056366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,5120,0.02317439913749695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,5120,0.016080000003178916
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,5120,4096,0.04341119925181071
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,5120,4096,0.011485866705576579
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,6144,6144,0.02002133329709371
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,4096,0.018386133511861167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,4096,0.014335999886194864
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,5120,3584,0.041484800974527995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,5120,3584,0.010803199807802836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,3584,0.01120960017045339
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,3584,0.012962133685747782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,5120,3072,0.03834773302078247
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,5120,3072,0.009891200065612792
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,3072,0.010012800494829815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,3072,0.012066133817036946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,5120,2560,0.0357151985168457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,5120,2560,0.009194667140642803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,2560,0.008728532989819845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,2560,0.010773332913716634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,5120,2048,0.03247359991073608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,5120,2048,0.008357333143552144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,2048,0.007463466624418895
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,2048,0.010132267077763876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,5120,1536,0.03017599980036418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,5120,1536,0.007430399954319
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,1536,0.006458666423956554
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,1536,0.009146666526794434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,5120,1024,0.027898667256037395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,5120,1024,0.006725333134333293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,1024,0.005261866748332978
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,1024,0.008373333017031352
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,6144,3072,0.009825066725413004
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,5120,768,0.026426666975021364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,768,0.004885333279768625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,768,0.007716266810894013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,5120,512,0.024836266040802003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,5120,512,0.005843199789524078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,512,0.004247466723124186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,512,0.007366399963696797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,5120,256,0.023702400922775268
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,5120,256,0.005278933544953664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,256,0.003772799919048945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,256,0.006866133213043213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,5120,128,0.02262826760609945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,128,0.003432533393303553
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,128,0.006701866785685222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,5120,64,0.02281600038210551
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,64,0.0032479998966058097
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,6144,1024,0.006548266609509785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,64,0.007038933535416921
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,5120,32,0.02206933299700419
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,32,0.0033642667035261786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,32,0.007192533214886982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,4096,65536,0.11550186475118
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,65536,0.2045642693837484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,6144,256,0.005346133311589559
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,4096,65536,0.28053439458211266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,65536,0.11595839659372967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,4096,16384,0.08011519908905029
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,16384,0.035062400499979655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,16384,0.06067306598027548
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,6144,128,0.005030400057633718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,4096,12288,0.06707839965820313
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,4096,12288,0.026570665836334228
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,12288,0.045459198951721194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,12288,0.028150399525960285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,4096,10240,0.06067306598027548
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,4096,10240,0.02302079995473226
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,10240,0.03412373463312785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,4096,8192,0.051178665955861416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,10240,0.02444159984588623
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,4096,8192,0.015929599603017174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,8192,0.028011733293533327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,8192,0.014934399724006652
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,4096,7168,0.06681280136108399
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,7168,0.025227733453114826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,7168,0.013019733627637229
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,4096,6144,0.05936746597290039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,4096,6144,0.01267519990603129
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,6144,0.02251733342806498
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,6144,0.012261333068211873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,5120,65536,0.13742186228434244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,4096,5120,0.05189119974772135
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,4096,5120,0.011356799801190694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,5120,0.017332265774408974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,5120,0.011090133587519329
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,4096,4096,0.042813865343729655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,4096,4096,0.009949866930643718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,4096,0.011117866635322571
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,4096,0.009777067104975383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,4096,3584,0.0402890682220459
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,3584,0.009596799810727436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,3584,0.009366400043169658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,4096,3072,0.03782080014546712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,4096,3072,0.008733866612116496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,3072,0.009423999985059103
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,3072,0.009382399916648864
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,4096,2560,0.034680533409118655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,4096,2560,0.00812799980243047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,2560,0.008267733454704284
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,2560,0.008444799979527792
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,4096,2048,0.03161919911702474
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,4096,2048,0.0073845331867535904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,2048,0.0067071999112765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,2048,0.00786240001519521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,4096,1536,0.029833600918451948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,1536,0.006107733150323232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,1536,0.007446399827798207
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,4096,1024,0.02723520000775655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,5120,768,0.006258133550484974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,1024,0.005053866902987162
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,1024,0.0070933332045873005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,4096,768,0.02598186731338501
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,4096,768,0.005861333509286245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,768,0.004576000074545542
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,768,0.006842666864395141
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,4096,512,0.024718934297561647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,4096,512,0.0056426664193471275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,512,0.004090666770935059
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,512,0.006592000027497609
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,5120,128,0.005009066561857859
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,4096,256,0.023520000775655112
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,4096,256,0.0051925331354141235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,256,0.0036661334335803984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,256,0.00626986672480901
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,4096,128,0.022940800587336222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,4096,128,0.00480320006608963
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,128,0.0034304000437259674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,128,0.006108800073464712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,4096,64,0.022471467653910317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,64,0.0064746667941411335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,64,0.003222399950027466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,4096,32,0.022078933318456014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,32,0.003642666588226954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,32,0.006540800134340922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3584,65536,0.1046336015065511
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,4096,16384,0.03379840056101481
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,65536,0.1800576051076253
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3584,65536,0.2678624153137207
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,65536,0.10571946303049724
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3584,16384,0.07693333625793457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,16384,0.05332906643549601
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,16384,0.03206826647122701
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3584,12288,0.024727465709050496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3584,12288,0.06507093509038289
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,12288,0.03641706705093384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,12288,0.025856000185012818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3584,10240,0.019449599583943687
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3584,10240,0.058592001597086586
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,10240,0.03068480094273885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,10240,0.019758933782577516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3584,8192,0.015230933825174967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3584,8192,0.04936746756235759
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,8192,0.025135999917984007
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,8192,0.0136245330174764
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3584,7168,0.013734400272369385
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3584,7168,0.06477440198262532
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,7168,0.022436267137527464
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,7168,0.012729600071907043
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,4096,7168,0.014064000050226847
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3584,6144,0.012483200430870056
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3584,6144,0.05700159867604574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,6144,0.020665599902470907
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,6144,0.011865599950154623
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3584,5120,0.049746131896972655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,5120,0.010909866293271382
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3584,5120,0.011187199751536052
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3584,4096,0.04107733170191447
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,5120,0.01243839959303538
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3584,4096,0.00981333355108897
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,4096,0.010050132870674133
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,4096,0.009649067123730978
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3584,3584,0.03977706829706828
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3584,3584,0.009411199887593587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,3584,0.009287466605504353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,3584,0.009348266323407491
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,4096,3584,0.009403733412424724
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3584,3072,0.03680853446324666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3584,3072,0.008719999591509502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,3072,0.008498133222262064
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,3072,0.008922666311264038
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3584,2560,0.03399680058161418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3584,2560,0.008085333307584127
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,2560,0.008330666522185007
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,2560,0.008302933474381765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3584,2048,0.03135040005048116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3584,2048,0.0073290665944417315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,2048,0.008567466338475545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,2048,0.007788800199826558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3584,1536,0.029265065987904865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,4096,1536,0.006806399921576183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,1536,0.006375466783841451
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,1536,0.00740479975938797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3584,1024,0.027063467105229694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3584,1024,0.005990399916966756
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,1024,0.005783466498057047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,1024,0.00692799985408783
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3584,768,0.026206932465235394
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3584,768,0.005717333157857259
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,4096,1024,0.006089599927266439
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,768,0.004982399940490723
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,768,0.0067669332027435304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3584,512,0.024528000752131143
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3584,512,0.005580799778302511
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,512,0.004328533510367076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,512,0.006559999783833821
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3584,256,0.023337600628534953
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3584,256,0.00510506679614385
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,256,0.003755733370780945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,256,0.0062826668222745255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3584,128,0.02249280015627543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,128,0.003587199995915095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,128,0.006021333237489065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3584,64,0.022711465756098427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,64,0.0033301333586374915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,64,0.0063498665889104204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3584,32,0.022038400173187256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,32,0.0034143999218940735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,32,0.006409599880377452
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,65536,0.15497706731160482
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,65536,0.10618133544921875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3072,65536,0.2612501303354899
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3072,16384,0.02921813329060872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3072,16384,0.07404159704844157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,16384,0.046598398685455324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,16384,0.03009493350982666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3072,12288,0.06293226480484009
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,12288,0.021820799509684245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,12288,0.036380799611409505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3072,10240,0.05719146728515625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3072,10240,0.01795519987742106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,10240,0.01735466718673706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,10240,0.03145706653594971
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3072,8192,0.014879999558130899
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3072,8192,0.04780266682306926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,8192,0.022717867294947305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,8192,0.013511466979980468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3072,7168,0.06363840103149414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3072,7168,0.013724799950917563
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,7168,0.020805333058039346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,7168,0.012497066458066305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3072,6144,0.0557365338007609
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3072,6144,0.012195199728012085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,6144,0.014548266927401224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,6144,0.01172693371772766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3072,5120,0.04991999864578247
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3072,5120,0.011036800344785054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,5120,0.016798933347066246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,5120,0.010851200421651204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3072,4096,0.04199680089950562
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,4096,0.013530666629473368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,4096,0.009630933403968811
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3072,3584,0.03899306853612264
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3072,3584,0.00928106705347697
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,3584,0.011694932977358501
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,3584,0.00939626693725586
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3072,3072,0.0366922656695048
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3072,3072,0.008504533767700195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,3072,0.010922666390736897
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3584,16384,0.03146986762682597
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,3072,0.008819199601809184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3072,2560,0.03452479839324951
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3072,2560,0.007963733375072479
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,2560,0.008132266501585644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,2560,0.008298666775226593
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3072,2048,0.03153066635131836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,2048,0.007293866574764251
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,2048,0.007831466694672901
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3072,1536,0.029371732473373414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,1536,0.006342400113741558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,1536,0.00735040009021759
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3072,1024,0.027973333994547527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3072,1024,0.005970133344332377
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,1024,0.005343999962011973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,1024,0.006874666611353557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3072,768,0.026228266954421996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3072,768,0.005730133255322774
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,768,0.004849066833655039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,768,0.006795733173688252
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3072,512,0.02474986712137858
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3072,512,0.005541333556175232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,512,0.004266666869322458
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,512,0.006499200065930684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3072,256,0.023809067408243813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3072,256,0.005022933085759481
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,256,0.0037578667203585305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,256,0.006252799928188324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3072,128,0.02278719941775004
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3072,128,0.004844800134499868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,128,0.0034688000877698263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,128,0.006030933558940887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3072,64,0.022631466388702393
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3584,1536,0.006945066650708516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,64,0.0031669333577156065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,64,0.006243200103441874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,3072,32,0.022660267353057862
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,32,0.0033621333539485933
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,32,0.0063391998410224915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2560,65536,0.09213866392771403
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2560,65536,0.2545034726460775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,65536,0.1324031988779704
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,65536,0.09081599712371827
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2560,16384,0.0720085302988688
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,16384,0.040566400686899824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3584,128,0.004840533435344696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,16384,0.026891734202702838
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2560,12288,0.06067306598027548
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2560,12288,0.0202293336391449
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,12288,0.03201813300450643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,12288,0.019100799163182577
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2560,10240,0.054871467749277744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2560,10240,0.017661867539087932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,10240,0.025091199080149333
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,10240,0.015253333250681558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2560,8192,0.04644266764322917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3072,65536,0.09612267017364502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,8192,0.01982933282852173
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,8192,0.013024000326792398
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3072,12288,0.021720532576243082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2560,7168,0.06097066799799601
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2560,7168,0.013692800203959146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,7168,0.015780267119407655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,7168,0.012289067109425861
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2560,6144,0.05362773338953654
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,6144,0.013611732920010885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,6144,0.011583999792734782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2560,5120,0.0475221316019694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2560,5120,0.010987733801205952
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,5120,0.011852799852689107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,5120,0.010667733351389567
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2560,4096,0.04067200024922689
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3072,4096,0.009655466675758362
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,4096,0.009979733824729919
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,4096,0.009496532877286275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2560,3584,0.038558932145436604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2560,3584,0.009190400441487629
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,3584,0.009195733070373534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,3584,0.009071999788284301
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2560,3072,0.0365610678990682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2560,3072,0.00849173367023468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,3072,0.008423466483751934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,3072,0.008709333340326945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2560,2560,0.03405760129292806
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2560,2560,0.008030933141708375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,2560,0.00766186664501826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,2560,0.00829013337691625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3072,2048,0.007165866593519847
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2560,2048,0.031175466378529866
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2560,2048,0.007216000060240428
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3072,1536,0.00666133314371109
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,2048,0.006818133095900218
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,2048,0.007717333237330119
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2560,1536,0.02961706717809041
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2560,1536,0.006681600213050842
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,1536,0.006000000238418579
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,1536,0.007339733342329661
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2560,1024,0.027564799785614012
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2560,1024,0.005905066430568695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,1024,0.00510506679614385
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,1024,0.006863999863465626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2560,768,0.02611306707064311
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2560,768,0.005703466633955637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,768,0.004553600152333578
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,768,0.006640000144640605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2560,512,0.024809600909550984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2560,512,0.005504000186920166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,512,0.004071466624736786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,512,0.006421333551406861
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2560,256,0.02373653252919515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2560,256,0.005055999755859375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,256,0.0036906667053699495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,256,0.006159999966621399
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2560,128,0.022999467452367146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2560,128,0.004819199939568837
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,128,0.003352533280849457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,128,0.005910400052865346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2560,64,0.022665599981943764
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,64,0.0031839999059836066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,64,0.006237866481145223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2560,32,0.021851734320322672
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,32,0.0033013333876927694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,32,0.006332799792289734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2048,65536,0.08304426670074463
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,65536,0.10875626405080159
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,65536,0.08295893669128418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2048,65536,0.25140372912089026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2048,16384,0.07021546363830566
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2048,16384,0.024473599592844644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,16384,0.03426880041758219
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,16384,0.023379200696945192
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2048,12288,0.01841813325881958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2048,12288,0.05946026643117269
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,12288,0.026574933528900148
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,12288,0.016489600141843162
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2048,10240,0.05354559818903605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2048,10240,0.015702399611473083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,10240,0.0231989324092865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,10240,0.014896000425020853
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2048,8192,0.04453973372777303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2048,8192,0.013503999511400858
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,8192,0.01416000028451284
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,8192,0.013000532984733582
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2560,16384,0.027705599864323933
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2048,7168,0.05971200068791708
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,7168,0.012640000383059184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,7168,0.012232533097267151
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2048,6144,0.054281600316365565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2048,6144,0.011331199606259664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,6144,0.01592639982700348
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,6144,0.01138879954814911
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2048,5120,0.04649920066197713
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2048,5120,0.010258133212725323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,5120,0.010578133662541707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,5120,0.013797332843144735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2048,4096,0.04125653505325318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2048,4096,0.009105066458384197
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,4096,0.011043199896812439
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,4096,0.009527466694513957
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2048,3584,0.03923199971516927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2048,3584,0.008611200253168742
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2560,8192,0.014789332946141561
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,3584,0.010408533612887065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,3584,0.009081600109736125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2048,3072,0.0362826665242513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2048,3072,0.007990399996439617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,3072,0.009521067142486572
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,3072,0.008649599552154542
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2048,2560,0.03380586703618367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2560,6144,0.012181333700815837
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,2560,0.007349333167076111
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,2560,0.00819946676492691
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2048,2048,0.031243733565012616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,2048,0.006520533561706543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,2048,0.007717333237330119
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2048,1536,0.02913600007692973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2048,1536,0.006685866912206013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,1536,0.005658666789531708
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,1536,0.0071829333901405334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2048,1024,0.02741866707801819
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2048,1024,0.005827199916044871
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,1024,0.004785066843032837
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,1024,0.0067893331249554946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2048,768,0.025769599278767902
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,768,0.0043381333351135256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2560,4096,0.009662933150927226
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,768,0.006656000018119812
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2048,512,0.02456000049908956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,512,0.0038922667503356934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,512,0.006393600006898243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2048,256,0.02372693419456482
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2048,256,0.004965333143870035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,256,0.003479466587305069
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,256,0.006180266539255777
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2048,128,0.004740266501903534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2048,128,0.022498132785161336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,128,0.0031839999059836066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,128,0.005929600199063619
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2048,64,0.02209706703821818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,64,0.006109866499900818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,64,0.003048533449570338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,2048,32,0.021724800268809
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,32,0.0031061333914597826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,32,0.006197333335876465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1536,65536,0.0804853359858195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,65536,0.0839733362197876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1536,65536,0.2452234745025635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,65536,0.07315413157145181
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1536,16384,0.06773227055867513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1536,16384,0.02332800030708313
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,16384,0.02638826568921407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,16384,0.020145066579182944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1536,12288,0.05706026554107666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1536,12288,0.01821333368619283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,12288,0.019307732582092285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,12288,0.01650879979133606
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1536,10240,0.05089173316955566
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1536,10240,0.015787733594576518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,10240,0.01478506624698639
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,10240,0.014945066968599954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1536,8192,0.043404801686604814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1536,8192,0.013351466258366904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,8192,0.012782933314641318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,8192,0.013061333696047464
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1536,7168,0.05696426630020142
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1536,7168,0.01245013376077016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,7168,0.011788800358772278
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,7168,0.012225066622098286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1536,6144,0.05159466663996378
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1536,6144,0.01120960017045339
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,6144,0.011755733688672384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,6144,0.011357866724332174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1536,5120,0.010174933075904845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1536,5120,0.04617600043614705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,5120,0.010312533378601075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,5120,0.010544000069300334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1536,4096,0.04109013477961222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1536,4096,0.009031466643015544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,4096,0.008909866213798523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,4096,0.00958720048268636
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1536,3584,0.039332266648610434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,3584,0.008191999793052674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,3584,0.009147733449935913
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1536,3072,0.036457598209381104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1536,3072,0.007973333199818928
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,3072,0.007451733450094859
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,3072,0.008638933300971985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1536,2560,0.03413653373718262
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1536,2560,0.007423999905586243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,2560,0.0071381335457166035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,2560,0.008242133259773254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1536,2048,0.031319467226664226
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1536,2048,0.006852266689141591
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,2048,0.006739200154940288
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,2048,0.007659733295440674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1536,1536,0.02932586669921875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2048,7168,0.012378666798273723
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,1536,0.005671466886997223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,1536,0.007203199962774913
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1536,1024,0.02691413362820943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1536,1024,0.00572266678015391
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,1024,0.004814933240413666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,1024,0.006751999755700429
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1536,768,0.025604265928268432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2048,2560,0.007445333401362102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,768,0.004296533266703288
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,768,0.0065738668044408154
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1536,512,0.02440320054690043
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1536,512,0.005241600175698599
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2048,2048,0.00682773341735204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,512,0.0038719999293486277
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,512,0.006374399860699971
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1536,256,0.023614933093388878
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,256,0.0034634667138258614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,256,0.006140799820423126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1536,128,0.02241706649462382
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1536,128,0.004744533201058706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,128,0.003186133255561193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,128,0.005985066791375478
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1536,64,0.022010666131973267
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,64,0.003035733352104823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,64,0.006040533383687338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1536,32,0.02227946718533834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,32,0.00308693324526151
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2048,768,0.005497600138187409
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,32,0.006181333462397257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1024,65536,0.07646079858144125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2048,512,0.005390933156013489
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,65536,0.060054401556650795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1024,65536,0.24044906298319496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,65536,0.07534826596577962
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1024,16384,0.06525760094324748
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1024,16384,0.022081067164738975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,16384,0.014500266313552857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,16384,0.019924267133076986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1024,12288,0.05497920115788778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1024,12288,0.01723519961039225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,12288,0.011430399616559346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,12288,0.016377600034077962
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1024,10240,0.049219199021657306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1024,10240,0.014856533209482829
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,10240,0.011171199878056844
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,10240,0.014798933267593383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1024,8192,0.041620266437530515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1024,8192,0.012828800082206725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,8192,0.00904746651649475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1024,7168,0.05707519849141439
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,8192,0.01288106640179952
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1024,7168,0.011826133728027344
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,7168,0.012124799688657125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,7168,0.009789866209030152
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1024,6144,0.050977067152659095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1024,6144,0.010753066341082255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,6144,0.011690666278203327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,6144,0.01125440001487732
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1024,5120,0.04585173527399699
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1024,5120,0.00974506636460622
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,5120,0.00876586635907491
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1024,4096,0.04071573416392009
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,5120,0.010451199611028035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1024,4096,0.008746666709582011
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,4096,0.008596266309420269
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,4096,0.009397332866986591
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1024,3584,0.03880000114440918
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1024,3584,0.008327466746171314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,3584,0.0086517333984375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1024,3072,0.03613013426462809
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,3584,0.008966400225957235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1024,3072,0.007738666733105977
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,3072,0.008134399851163227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,3072,0.00855466624101003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1024,2560,0.03368639945983887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1024,2560,0.007152000069618225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,2560,0.006656000018119812
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1024,2048,0.006681600213050842
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,2560,0.008081066608428954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1024,2048,0.03100266655286153
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,2048,0.005906133353710175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,2048,0.007659733295440674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1024,1536,0.02892373402913411
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,1536,0.0052704001466433205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,1536,0.007157333195209503
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1024,1024,0.02690560022989909
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,1024,0.004440533121426901
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1536,3584,0.008563199639320373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,1024,0.0067007998625437425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1024,768,0.025597866376241046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1024,768,0.005377066632111868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,768,0.004166399935881296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,768,0.006504533191521962
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1024,512,0.024201599756876628
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1024,512,0.00522986650466919
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,512,0.0037589333951473236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,512,0.006276266773541768
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1024,256,0.022827732563018798
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1024,256,0.004802133142948151
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,256,0.0034506666163603462
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,256,0.006077866752942403
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1024,128,0.022358399629592896
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1024,128,0.004635733366012573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1536,1536,0.006341333190600078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,128,0.003126399964094162
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,128,0.005906133353710175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1024,64,0.022175999482472737
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,64,0.0030239999294281008
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,64,0.006011733412742614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,1024,32,0.021731199820836385
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,32,0.003094399968783061
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,32,0.006009600063165029
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,768,65536,0.23484907150268555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,65536,0.047326934337615964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,65536,0.07020586331685384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,768,16384,0.06331199804941813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,768,16384,0.021342933177947998
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,16384,0.011401599645614624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,16384,0.01991999944051107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1536,768,0.005407999952634176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,768,12288,0.053711998462677005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,12288,0.012258133292198181
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,12288,0.01629866659641266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,768,10240,0.014879999558130899
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,768,10240,0.04788586695988973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,10240,0.011351466178894043
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,10240,0.01483519971370697
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,768,8192,0.04054400126139323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,768,8192,0.012689066926638284
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,8192,0.009985066453615824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,8192,0.012797866264979044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,768,7168,0.055986134211222324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,768,7168,0.011718400319417318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,7168,0.012054399649302164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,7168,0.009566932916641235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,768,6144,0.05142826636632284
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,768,6144,0.010714667042096455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1536,256,0.004891733328501383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,6144,0.008898133039474487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,6144,0.011337600151697795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,5120,0.008803199728329976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,768,5120,0.04614400068918864
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,5120,0.010595200459162395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,768,4096,0.040438401699066165
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,768,4096,0.008646399776140849
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,4096,0.008147199948628742
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,4096,0.009401599566141765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,768,3584,0.038753068447113036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,768,3584,0.008217599987983704
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,3584,0.00786240001519521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,3584,0.008989866574605305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,768,3072,0.036651734511057535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,768,3072,0.007696000238259633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,3072,0.007179733117421467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,3072,0.008559999863306682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,768,2560,0.033964800834655764
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,768,2560,0.0071168000499407455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,2560,0.0065184002121289565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,2560,0.008075733482837678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,768,2048,0.03101759950319926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,768,2048,0.006607999900976817
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,2048,0.0075328002373377485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,2048,0.005760000149408976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,768,1536,0.029473066329956055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,1536,0.005032533407211303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,1536,0.00714026689529419
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,768,1024,0.027030400435129803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,768,1024,0.005598933498064677
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,1024,0.0043381333351135256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,1024,0.006751999755700429
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,768,768,0.025464532772699992
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,768,768,0.005366399884223938
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,768,0.003988266736268997
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,768,0.006513066589832306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,768,512,0.024141865968704223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,768,512,0.0051466668645540874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,512,0.00365226666132609
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,512,0.006285866598288219
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,768,256,0.022871466477711995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,768,256,0.004789333542188009
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,256,0.003306666761636734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,256,0.006087466577688853
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,768,128,0.022300799687703453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,768,128,0.004650666813055674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,128,0.003052799900372823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,128,0.005939200023810069
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,768,64,0.022083199024200438
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,64,0.002958933264017105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,64,0.005901866654555003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,768,32,0.022142932812372843
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,32,0.0029546665648619336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,32,0.005919999877611796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,512,65536,0.23230293591817222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1024,1536,0.006105599800745646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,65536,0.038865065574645995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,65536,0.07064747015635173
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1024,1024,0.005580799778302511
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,512,16384,0.06255679925282796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,512,16384,0.021206400791803994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,16384,0.011180800199508668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,16384,0.019883733987808228
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,512,12288,0.052629331747690834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,12288,0.009802666306495667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,12288,0.01644159952799479
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,512,10240,0.047882668177286786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,512,10240,0.014806399742762247
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,10240,0.010243200262387593
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,10240,0.014756266276041666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,512,8192,0.03970133463541667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,512,8192,0.012658133109410604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,8192,0.008646399776140849
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,8192,0.012812800208727517
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,512,7168,0.055744000275929774
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,512,7168,0.011748266220092774
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,7168,0.008461866776148479
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,7168,0.012052266796429952
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,512,6144,0.05076266527175903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,512,6144,0.010618666807810467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,6144,0.009288533528645834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,6144,0.011212799946467083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,512,5120,0.045797332127888994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,512,5120,0.009684266646703084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,5120,0.008649599552154542
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,5120,0.010390399893124899
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,512,4096,0.040439466635386154
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,512,4096,0.008526933193206788
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,4096,0.007926400005817413
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,4096,0.009378133217493693
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,512,3584,0.03861226638158162
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,3584,0.007946666578451793
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,3584,0.008922666311264038
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,512,3072,0.03635733524958293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,512,3072,0.007573333382606506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,3072,0.007208533088366191
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,3072,0.008488532900810242
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,768,65536,0.07579519748687744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,512,2560,0.033794132868448894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,2560,0.006423466900984447
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,2560,0.008056533336639405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,512,2048,0.030937600135803222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,512,2048,0.0064522668719291685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,2048,0.005781333148479462
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,768,12288,0.017222400506337485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,2048,0.00747626672188441
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,512,1536,0.006005333364009857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,512,1536,0.029041065772374468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,1536,0.005050666630268097
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,1536,0.007175466914971669
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,512,1024,0.026867200930913288
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,512,1024,0.0055285334587097164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,1024,0.004343466460704803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,1024,0.0066890666882197065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,512,768,0.02541653315226237
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,512,768,0.0053247998158137
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,768,0.004008533308903376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,768,0.006520533561706543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,512,512,0.0241376002629598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,512,0.0036533333361148832
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,512,0.0061749334136645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,512,256,0.022768000761667885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,512,256,0.004726399978001913
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,256,0.003335466732581457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,256,0.00597866674264272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,768,5120,0.009672533472379048
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,512,128,0.02232746680577596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,128,0.003067733347415924
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,128,0.005816533168156942
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,512,64,0.022012799978256226
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,64,0.0030239999294281008
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,64,0.005861333509286245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,512,32,0.02134079933166504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,32,0.0032074667513370516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,32,0.005901866654555003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,256,65536,0.22607146898905434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,256,65536,0.07573760350545247
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,65536,0.026744532585144042
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,65536,0.06936213175455729
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,256,16384,0.06175359884897867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,768,1536,0.006118399898211161
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,256,16384,0.021150932709376017
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,16384,0.009851732850074768
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,16384,0.019963733355204263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,256,12288,0.0516543984413147
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,256,12288,0.017198934157689413
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,12288,0.008665600419044494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,12288,0.016437333822250367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,256,10240,0.04585280021031697
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,10240,0.008106666803359985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,10240,0.014719999829928079
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,256,8192,0.040651734670003256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,512,65536,0.07617812951405843
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,256,8192,0.012648533781369528
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,8192,0.007527466615041096
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,8192,0.012774399916330972
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,256,7168,0.05687040090560913
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,7168,0.007277866701285045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,7168,0.011991467078526814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,512,12288,0.01730560064315796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,256,6144,0.05039680004119873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,6144,0.007063466807206471
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,6144,0.011177600423494975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,256,5120,0.04576853513717651
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,256,5120,0.009630933403968811
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,5120,0.007493333518505096
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,5120,0.010366933544476827
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,256,4096,0.040268798669179276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,256,4096,0.008563199639320373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,4096,0.007079466680685679
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,4096,0.009342933694521587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,256,3584,0.03906453450520833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,256,3584,0.008075733482837678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,3584,0.0074538667996724445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,3584,0.008835200468699138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,256,3072,0.03629226684570312
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,512,3584,0.008155733346939087
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,3072,0.00705813318490982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,3072,0.008508800466855367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,256,2560,0.033794132868448894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,256,2560,0.007083733379840851
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,2560,0.006308266520500183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,512,2560,0.006994133194287618
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,2560,0.007973333199818928
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,256,2048,0.03086400032043457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,256,2048,0.006448000172773997
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,2048,0.005634133517742157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,2048,0.007482666770617168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,256,1536,0.02895786762237549
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,256,1536,0.006049066781997681
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,1536,0.004940799872080485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,1536,0.007112533350785573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,256,1024,0.026995199918746948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,256,1024,0.005478399991989136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,1024,0.0042581334710121155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,1024,0.00664106657107671
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,256,768,0.02523626685142517
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,256,768,0.005268266797065735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,768,0.003970133264859518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,768,0.0064629331231117245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,256,512,0.02393280069033305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,256,512,0.005011199911435445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,512,0.0036159999668598174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,512,0.006209066510200501
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,512,512,0.005110399921735128
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,256,256,0.0231605331103007
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,256,0.0032885332902272543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,256,0.005965866645177205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,256,128,0.022551467021306358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,256,128,0.004554666578769684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,128,0.0030613332986831666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,128,0.005830400188763936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,256,64,0.02204373280207316
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,64,0.0029002666473388673
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,64,0.005904000004132589
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,512,128,0.0045621335506439206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,32,0.0029696000119050344
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,256,32,0.02137920061747233
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,32,0.005826133489608765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,128,65536,0.0745247999827067
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,65536,0.014750933647155762
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,128,65536,0.22662933667500815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,65536,0.06669013500213623
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,128,16384,0.061242667833964023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,128,16384,0.021180800596872964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,16384,0.0077919999758402515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,16384,0.019640533129374187
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,128,12288,0.05010026693344116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,12288,0.007141333321730297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,12288,0.016316800316174825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,128,10240,0.04460053443908692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,10240,0.007342933118343354
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,10240,0.014490666985511779
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,128,8192,0.03810986677805583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,256,10240,0.014806399742762247
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,8192,0.00724480003118515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,8192,0.012745599945386252
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,128,7168,0.055233065287272134
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,128,7168,0.01165120005607605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,7168,0.007134933272997539
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,7168,0.01192639966805776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,128,6144,0.05048640171686808
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,128,6144,0.010521599650382995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,6144,0.006950399776299794
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,6144,0.011107200384140014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,128,5120,0.045858132839202884
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,128,5120,0.009604266285896302
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,5120,0.007357866565386455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,5120,0.01029866635799408
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,128,4096,0.040693334738413495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,128,4096,0.008505599697430928
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,4096,0.006897066533565521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,256,7168,0.011707733074824016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,4096,0.009169066945711773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,128,3584,0.03859200080235799
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,128,3584,0.00803306649128596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,3584,0.007377066711584728
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,3584,0.00877333382765452
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,128,3072,0.036345601081848145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,128,3072,0.0075541332364082335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,3072,0.0070602665344874065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,3072,0.008355200290679932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,128,2560,0.03351786533991496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,128,2560,0.006970666845639547
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,2560,0.006289066871007283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,2560,0.00787306676308314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,128,2048,0.030794666210810347
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,256,6144,0.010636799534161885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,128,2048,0.006434133152167003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,2048,0.005636266867319743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,2048,0.007420800129572551
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,128,1536,0.028934399286905926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,128,1536,0.005963733295599619
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,1536,0.004962133367856344
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,1536,0.006940799951553345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,128,1024,0.026616533597310383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,128,1024,0.005393066505591074
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,1024,0.00425600012143453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,1024,0.006555733581384023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,128,768,0.02531413237253825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,128,768,0.005155199766159057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,768,0.003923200070858002
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,768,0.006357333560784657
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,128,512,0.024027733008066814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,512,0.003575466573238373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,128,512,0.0049792001644770306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,512,0.006107733150323232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,128,256,0.02291733423868815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,256,3072,0.007566933333873749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,128,256,0.004774400095144907
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,256,0.0032799998919169106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,256,0.005888000130653381
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,128,128,0.022398932774861654
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,128,128,0.004556799928347269
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,128,0.0030271999537944795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,128,0.005798399945100148
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,128,64,0.021995733181635536
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,64,0.0028864001234372456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,64,0.00580266664425532
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,128,32,0.02128960092862447
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,32,0.0029258665939172106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,32,0.0058378666639328
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,65536,0.011628799637158712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,65536,0.06398186683654786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,64,65536,0.2261045296986898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,64,16384,0.06094186703364054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,16384,0.00705813318490982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,64,12288,0.051045334339141844
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,16384,0.019637332359949747
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,12288,0.0070250665148099255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,12288,0.016198399662971496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,64,10240,0.044973866144816084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,10240,0.007238399982452392
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,10240,0.014520532886187234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,8192,0.007130666573842366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,64,8192,0.038038400808970134
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,7168,0.006919466455777486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,8192,0.012651733557383218
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,64,7168,0.05553493499755859
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,7168,0.011953066786130269
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,6144,0.011047466595967611
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,64,6144,0.04994026819864909
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,5120,0.007143466671307881
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,6144,0.007067733506361644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,64,5120,0.045651201407114664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,64,4096,0.04020053148269653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,5120,0.010266666611035664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,4096,0.006804266571998596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,4096,0.009262933333714803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,64,3584,0.03802026510238647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,256,256,0.004724266628424326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,3584,0.007190399865309398
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,3584,0.008750933408737182
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,64,3072,0.03598933219909668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,3072,0.006942933301130931
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,3072,0.008321066697438557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,64,2560,0.03337386846542358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,2560,0.0062613333264986675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,2560,0.00788800021012624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,64,2048,0.03061760067939758
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,2048,0.005585066477457682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,2048,0.007322666545708974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,64,1536,0.02890239953994751
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,1536,0.0048885335524876915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,1024,0.006567466755708058
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,1536,0.006997333467006683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,64,1024,0.026947200298309326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,1024,0.004227200150489807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,768,0.0039264000952243805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,64,768,0.025149865945180254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,768,0.006347733239332835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,64,512,0.024545067548751832
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,512,0.0035637333989143372
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,512,0.006166400015354156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,64,256,0.022872533400853476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,256,0.0032672000428040824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,256,0.005970133344332377
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,64,128,0.02235306700070699
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,128,0.003009066730737686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,128,0.005824000140031179
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,64,64,0.021683200200398763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,128,12288,0.017174400885899863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,64,0.0029728000362714132
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,128,10240,0.014813866217931113
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,64,0.00576853354771932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,64,32,0.021246933937072755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,32,0.0028661333024501802
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,32,0.005832533538341522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,128,8192,0.012573867042859396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,65536,0.011507200201352437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,32,65536,0.22546985944112144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,65536,0.062023464838663736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,32,16384,0.06059946616490682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,16384,0.007065600156784058
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,16384,0.019636267423629762
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,32,12288,0.05000746647516886
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,12288,0.006989866495132446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,12288,0.016064000129699708
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,32,10240,0.04407680034637451
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,10240,0.007090133428573608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,10240,0.014437333742777506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,32,8192,0.0379093329111735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,8192,0.0069578667481740315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,8192,0.012661332885424295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,7168,0.0068896000583966565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,32,7168,0.054926931858062744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,7168,0.01176533301671346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,32,6144,0.049959464867909746
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,6144,0.00680213322242101
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,6144,0.010871466994285584
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,32,5120,0.04567360083262126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,5120,0.00717439999183019
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,5120,0.01015786627928416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,32,4096,0.040109864870707196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,4096,0.00660693347454071
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,4096,0.00902826686700185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,32,3584,0.03845226764678955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,3584,0.007124266525109608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,3584,0.00879253347714742
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,32,3072,0.03622186581293742
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,3072,0.0067210664351781205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,3072,0.008338133494059246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,2560,0.006209066510200501
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,32,2560,0.0337066650390625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,2560,0.007893333335717519
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,32,2048,0.0309173325697581
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,2048,0.005514666438102722
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,2048,0.007439999779065449
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,32,1536,0.02872213323911031
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,1536,0.004867200056711833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,1536,0.006962133447329204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,32,1024,0.027360000212987262
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,1024,0.004173866907755534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,1024,0.006585599978764851
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,32,768,0.02522346576054891
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,768,0.0038592000802357995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,768,0.006340266764163971
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,32,128,0.022374399503072104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,32,512,0.023957333962122598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,512,0.0035189333061377203
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,512,0.006120533247788747
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,32,256,0.02283733288447062
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,256,0.0032298666735490165
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,256,0.005871999760468801
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,128,0.0029919999341169994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,128,0.005793066819508871
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,32,64,0.02178666591644287
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,64,0.0028266665836175283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,64,0.00576853354771932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,128,32,32,0.02104960083961487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,32,0.0029045333464940387
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,32,0.00573333352804184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,65536,16384,0.6702069600423177
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,65536,16384,0.44875307083129884
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,65536,12288,0.4915146509806315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,65536,16384,0.8139488220214843
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,65536,12288,0.3817248026529948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,65536,12288,0.5806336085001628
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,65536,10240,0.2516970634460449
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,65536,10240,0.308841609954834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,65536,10240,0.42212053934733074
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,65536,8192,0.2631125291188558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,65536,10240,0.4839392026265462
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,65536,7168,0.1694591999053955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,65536,8192,0.4104309399922689
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,65536,8192,0.3555957476298014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,65536,7168,0.25498560269673665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,65536,7168,0.33856318791707357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,65536,6144,0.15491627057393392
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,65536,7168,0.3065600077311198
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,65536,6144,0.204474671681722
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,65536,6144,0.25296212832132975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,65536,6144,0.29411945343017576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,65536,5120,0.12588586807250976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,65536,5120,0.18273919423421223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,65536,5120,0.25966080029805505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,65536,5120,0.21722346941630044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,65536,4096,0.15147093137105305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,65536,4096,0.2101290702819824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,65536,4096,0.16880319913228353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,65536,3584,0.093941330909729
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,65536,3584,0.13765974044799806
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,65536,3584,0.17440533638000488
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,65536,3584,0.1377621332804362
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,65536,3072,0.12618133227030437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,65536,3072,0.08118399779001871
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,65536,3072,0.15152427355448406
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,65536,3072,0.11522666613260905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,65536,2560,0.11111679871877034
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,65536,2560,0.12791679700215658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,65536,2560,0.1109610637029012
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,65536,2048,0.09935466448465982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,65536,2048,0.05868693192799886
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,65536,2048,0.10460586547851562
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,65536,2048,0.08630293210347494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,65536,1536,0.04710719982783
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,65536,1536,0.08571840127309163
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,65536,1536,0.08012693723042806
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,65536,1536,0.06344639857610067
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,65536,1024,0.03452053467432658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,65536,1024,0.07065707047780355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,65536,1024,0.05595200061798096
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,65536,1024,0.04140053192774455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,65536,768,0.028493867317835493
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,65536,768,0.06440533399581909
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,65536,768,0.04429866472880046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,65536,768,0.03586560090382894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,65536,512,0.05816533168156942
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,65536,512,0.03213013410568237
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,65536,512,0.02453119953473409
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,65536,512,0.02619839906692505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,65536,256,0.0449290672938029
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,65536,256,0.0125408003727595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,65536,256,0.017729065815607705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,65536,256,0.018620800971984864
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,65536,128,0.04024853308995564
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,65536,128,0.010040533542633057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,65536,128,0.008080000181992848
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,65536,128,0.016670932372411094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,65536,64,0.03927893241246541
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,65536,64,0.006975999971230824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,65536,64,0.0252074658870697
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,65536,32,0.038713598251342775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,65536,32,0.007332266867160797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,65536,32,0.026233599583307905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,16384,65536,0.3957269350687663
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,65536,12288,0.28091519673665366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,65536,16384,0.37163947423299154
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,16384,65536,0.5298890749613444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,65536,0.8339199701944986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,16384,16384,0.10039573510487873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,16384,16384,0.14643200238545734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,16384,12288,0.07703680197397868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,16384,0.16992212931315104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,16384,12288,0.12047573725382488
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,65536,0.6518154780069987
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,16384,0.20175253550211586
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,65536,8192,0.19427307446797687
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,12288,0.1428895950317383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,12288,0.11078186829884847
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,16384,10240,0.1024959961573283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,16384,10240,0.06532906691233317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,10240,0.1275007963180542
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,10240,0.09325760205586751
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,16384,8192,0.08754239877065023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,16384,8192,0.054960000514984134
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,8192,0.09718613624572754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,8192,0.08213760058085123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,16384,7168,0.1019701321919759
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,7168,0.08539413611094157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,7168,0.06828587055206299
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,16384,6144,0.089302396774292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,6144,0.07898773352305094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,6144,0.05846933523813883
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,16384,5120,0.07930453618367514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,5120,0.06639573176701864
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,5120,0.051989332834879554
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,16384,4096,0.03231679995854696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,16384,4096,0.06837226549784342
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,4096,0.05190080006917318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,4096,0.040158931414286295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,16384,3584,0.06346453428268432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,3584,0.04799786806106567
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,3584,0.03726826508839925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,16384,3072,0.025496532519658405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,16384,3072,0.059164798259735106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,3072,0.040250666936238605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,3072,0.03205546736717224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,16384,2560,0.02490239938100179
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,16384,2560,0.05473599831263224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,2560,0.034144000212351484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,2560,0.02769706646601359
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,65536,4096,0.10645120143890381
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,16384,2048,0.047083731492360434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,65536,2560,0.07045439879099527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,2048,0.028307199478149414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,2048,0.02336960037549337
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,16384,1536,0.03775786558787028
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,16384,1536,0.01206933359305064
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,1536,0.022759467363357544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,1536,0.01919999917348226
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,16384,1024,0.033514666557312014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,1024,0.010060800115267436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,1024,0.013991467157999673
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,16384,768,0.031093333164850873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,16384,768,0.008550399541854858
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,768,0.008148266871770223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,768,0.012518399953842163
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,16384,512,0.028830933570861816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,16384,512,0.007758933305740357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,512,0.0065194666385650635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,512,0.010215466221173603
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,16384,256,0.026386133829752606
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,16384,256,0.006489600241184235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,256,0.0052906667192777
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,256,0.008653866251309712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,16384,128,0.02490026752154032
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,16384,128,0.0059349333246548975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,128,0.004379733403523763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,128,0.00817386656999588
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,16384,64,0.02547840078671773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,64,0.0040394666294256846
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,64,0.009901866316795349
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,16384,32,0.025864533583323163
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,32,0.004340266684691111
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,32,0.00970240036646525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,12288,65536,0.29780801137288415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,12288,65536,0.45902185440063475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,65536,0.5804927825927735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,65536,0.4794101397196452
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,12288,16384,0.12788266340891521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,12288,16384,0.09280959765116373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,16384,0.15906666119893392
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,16384,0.11738879680633545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,16384,7168,0.048631465435028075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,12288,12288,0.06280213197072347
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,12288,12288,0.106277330716451
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,12288,0.10673173268636067
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,12288,0.08480106989542643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,16384,6144,0.042845865090688065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,12288,10240,0.05342186689376831
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,12288,10240,0.09082773526509604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,10240,0.09373439947764078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,10240,0.07234880129496256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,12288,8192,0.07795306841532389
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,12288,8192,0.04438079992930095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,16384,5120,0.03719040155410767
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,8192,0.07618880271911621
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,8192,0.060697599252065026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,12288,7168,0.09097066720326742
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,12288,7168,0.04204479853312175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,7168,0.06531200011571249
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,7168,0.05275733470916748
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,16384,3584,0.028538666168848675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,12288,6144,0.03502293427785237
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,12288,6144,0.08154346942901611
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,6144,0.045337601502736406
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,6144,0.057697065671285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,12288,5120,0.030371199051539104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,12288,5120,0.07240213553110758
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,5120,0.04807466665903727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,5120,0.04040426810582479
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,12288,4096,0.06337386767069499
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,12288,4096,0.02511253356933594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,4096,0.041092268625895184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,4096,0.03151359955469767
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,12288,3584,0.05899733304977417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,16384,2048,0.014653866489728292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,12288,3584,0.023425066471099855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,3584,0.03501439889272054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,3584,0.029951999584833782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,12288,3072,0.05646400054295858
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,12288,3072,0.020169599850972494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,3072,0.03099306623140971
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,3072,0.026169600089391072
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,12288,2560,0.04817493359247844
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,12288,2560,0.01295146644115448
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,2560,0.026616533597310383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,2560,0.022457599639892578
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,12288,2048,0.04169600009918213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,12288,2048,0.011506133278210958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,16384,1024,0.00932373305161794
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,2048,0.023013333479563393
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,2048,0.017061332861582436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,12288,1536,0.009821866949399311
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,12288,1536,0.03591146469116211
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,1536,0.015467733144760132
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,1536,0.010637866457303365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,12288,1024,0.03293760021527608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,12288,1024,0.008276266853014629
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,1024,0.00755626658598582
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,1024,0.011784533659617107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,12288,768,0.030961066484451294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,12288,768,0.0077898666262626644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,768,0.00659093310435613
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,768,0.010345600048700969
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,12288,512,0.02848213315010071
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,256,0.004715733230113983
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,512,0.005604266623655955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,512,0.008727467060089112
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,12288,256,0.026305067539215087
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,256,0.007876266539096833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,12288,128,0.024915200471878052
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,128,0.004084266722202301
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,128,0.007541333138942718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,12288,64,0.025300266345342
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,64,0.003692800054947535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,64,0.008482133348782856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,12288,32,0.02453226645787557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,32,0.004020266731580098
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,32,0.00879253347714742
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,65536,0.43263893127441405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,65536,0.49526185989379884
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,10240,65536,0.4160885175069173
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,10240,16384,0.07649706999460856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,10240,16384,0.12108053366343181
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,16384,0.15764800707499188
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,16384,0.10231359799702962
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,10240,12288,0.09959466457366943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,10240,12288,0.055061332384745275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,12288,0.0905407985051473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,12288,0.08013013203938803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,10240,10240,0.04673279921213786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,10240,10240,0.08589119911193847
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,10240,0.07634666760762533
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,10240,0.067030398050944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,10240,8192,0.038457600275675456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,10240,8192,0.07368533611297608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,8192,0.06145493189493815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,8192,0.05581119855244955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,10240,7168,0.08760213057200114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,7168,0.054671998818715414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,7168,0.04901653528213501
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,10240,6144,0.031181865930557252
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,10240,6144,0.07921919822692872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,6144,0.04878079891204834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,6144,0.04239999850591024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,10240,5120,0.07077866395314535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,10240,5120,0.02664533257484436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,5120,0.04091946681340535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,5120,0.03696639935175578
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,10240,4096,0.02223893404006958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,10240,4096,0.06391573349634806
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,4096,0.03545706669489543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,4096,0.02834560076395671
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,10240,3584,0.016498133540153503
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,10240,3584,0.05654186805089315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,3584,0.030077866713205975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,3584,0.028604799509048463
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,10240,3072,0.04996159871419271
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,10240,3072,0.013514666755994161
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,3072,0.02707093358039856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,3072,0.022788266340891518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,10240,2560,0.04527786572774251
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,10240,2560,0.012435199817021687
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,2560,0.02303253412246704
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,2560,0.020860799153645835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,10240,2048,0.04084479808807373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,2048,0.017941333850224814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,2048,0.017574399709701538
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,10240,1536,0.03517226775487264
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,10240,1536,0.009353599945704142
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,1536,0.0091839998960495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,1536,0.014708266655604044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,10240,1024,0.03239573240280151
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,10240,1024,0.007913599908351897
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,1024,0.007276799778143566
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,1024,0.010987733801205952
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,10240,768,0.030081067482630414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,10240,768,0.007206400235493978
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,768,0.0063178668419520065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,768,0.010696533322334289
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,10240,512,0.028152533372243244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,10240,512,0.006520533561706543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,512,0.005343999962011973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,512,0.008640000224113464
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,10240,256,0.02614826758702596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,10240,256,0.0058773333827654515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,256,0.004555733501911163
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,256,0.0077909335494041445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,10240,128,0.0251093327999115
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,10240,128,0.005411200225353241
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,128,0.0039989332358042395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,128,0.007417599856853485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,10240,64,0.025115732351938886
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,64,0.003647999962170919
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,64,0.008116266628106435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,10240,32,0.025302400191624958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,32,0.0037952000896135964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,32,0.008505599697430928
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,8192,65536,0.19942720731099445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,8192,65536,0.3267061233520508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,12288,512,0.006887466708819072
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,65536,0.40244372685750324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,12288,256,0.00598826656738917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,12288,128,0.005644799768924713
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,65536,0.3192991892496745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,8192,16384,0.055335466066996256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,8192,16384,0.09909866650899252
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,16384,0.07606399854024251
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,16384,0.10608426729838054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,8192,12288,0.08280533154805501
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,8192,12288,0.0469215989112854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,12288,0.07375679810841879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,12288,0.05728000005086263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,8192,10240,0.03725866476694743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,8192,10240,0.07100053628285727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,10240,0.06178880135218302
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,10240,0.04953600168228149
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,8192,8192,0.06191146771113078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,8192,8192,0.03113066752751668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,8192,0.05026880105336508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,8192,0.04152853488922119
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,8192,7168,0.07667413552602133
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,8192,7168,0.0280074675877889
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,7168,0.04443200031916301
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,7168,0.03710613250732422
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,8192,6144,0.06937066713968912
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,8192,6144,0.025089067220687867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,6144,0.039450665314992264
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,10240,65536,0.25672426223754885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,6144,0.03176106611887614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,8192,5120,0.06164586544036865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,8192,5120,0.02142080068588257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,5120,0.03318933248519897
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,5120,0.02929813265800476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,8192,4096,0.05078293482462565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,8192,4096,0.017350399494171144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,4096,0.02750399907430013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,4096,0.02260800004005432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,8192,3584,0.04520426591237386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,3584,0.024858667453130086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,3584,0.02066133419672648
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,8192,3072,0.012643200159072877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,8192,3072,0.04085439840952555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,3072,0.01831573247909546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,3072,0.02472426692644755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,8192,2560,0.03761813243230184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,2560,0.017046399911244712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,2560,0.016268799702326454
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,10240,7168,0.03516800006230672
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,8192,2048,0.03383786678314209
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,8192,2048,0.010158933202425639
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,8192,1536,0.03086079955101013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,2048,0.010880000392595927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,2048,0.013895466923713684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,8192,1536,0.008534399668375652
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,1536,0.00921493371327718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,1536,0.01188373366991679
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,8192,1024,0.028727465867996217
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,8192,1024,0.007333333293596904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,1024,0.0071712002158164975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,1024,0.009525332848230999
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,8192,768,0.02682773272196452
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,768,0.0065311998128890995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,768,0.008648533622423809
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,8192,512,0.02540053327878316
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,512,0.005369600156943003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,512,0.0076896001895268755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,8192,256,0.023626667261123658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,8192,256,0.005679999788602194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,10240,2048,0.010430933038393656
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,256,0.0043722664316495265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,256,0.007029333213965099
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,8192,128,0.02318399945894877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,8192,128,0.0052138666311899815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,128,0.003756800045569738
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,8192,64,0.02341759999593099
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,128,0.0067114666104316715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,64,0.003542399903138479
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,64,0.007356800138950348
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,8192,32,0.023319466908772787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,32,0.0038463999827702843
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,32,0.0075082664688428235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,7168,65536,0.18342506090799968
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,65536,0.3029322624206543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,7168,65536,0.2977973302205404
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,65536,0.3517120043436686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,7168,16384,0.09315626621246338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,16384,0.07364586989084879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,16384,0.10623253186543782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,7168,12288,0.07864320278167725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,7168,12288,0.04080959955851237
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,12288,0.06531200011571249
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,12288,0.053958400090535485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,7168,10240,0.06747732957204183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,7168,10240,0.03431253433227539
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,10240,0.05600320100784302
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,10240,0.04693333307902019
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,7168,8192,0.05880853335062662
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,7168,8192,0.028964267174402876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,8192,0.044429866472880046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,8192,0.041510399182637533
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,7168,7168,0.07399786313374837
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,7168,0.04036266803741455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,7168,0.035531731446584065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,7168,6144,0.06729386647542318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,7168,6144,0.023562665780385336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,6144,0.03552533388137817
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,6144,0.032001066207885745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,7168,5120,0.05993280013402304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,7168,5120,0.018810667594273887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,5120,0.03039039969444275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,5120,0.028519467512766523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,7168,4096,0.04960213502248128
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,7168,4096,0.01647040049235026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,4096,0.025221333901087446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,4096,0.021937066316604616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,7168,3584,0.04508800109227498
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,7168,3584,0.015029333035151162
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,3584,0.022252800067265828
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,3584,0.019912532965342202
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,7168,3072,0.04037439823150635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,7168,3072,0.012625066439310708
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,3072,0.01914880077044169
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,3072,0.017617066701253258
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,8192,3584,0.014929067095120749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,7168,2560,0.012349866827329
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,7168,2560,0.03753279844919841
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,2560,0.01074026624361674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,2560,0.016105600198109946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,7168,2048,0.033454934755961105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,2048,0.008970666925112407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,2048,0.013140267133712769
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,7168,1536,0.031344000498453775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,7168,1536,0.008734933535257975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,8192,2560,0.01139520009358724
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,1536,0.0076896001895268755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,1536,0.011766399939854939
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,7168,1024,0.028492800394694012
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,7168,1024,0.00798399994770686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,1024,0.006178133189678192
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,1024,0.009126399954160053
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,7168,768,0.027553067604700728
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,7168,768,0.006840533514817555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,768,0.005585066477457682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,768,0.008524800340334576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,7168,512,0.02536853353182475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,512,0.004694400231043497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,512,0.007639466722806294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,7168,256,0.024038400252660116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,8192,768,0.006739200154940288
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,7168,256,0.005453866720199585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,256,0.004037333279848098
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,256,0.0071274667978286745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,8192,512,0.00617386649052302
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,7168,128,0.02294399936993917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,7168,128,0.00517439991235733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,128,0.0036960000793139136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,128,0.006657066444555919
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,7168,64,0.022696532805760703
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,64,0.003588266670703888
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,64,0.0074325333038965866
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,7168,32,0.022184532880783082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,32,0.003702399879693985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,32,0.007480533421039581
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,65536,0.2978314717610677
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,6144,65536,0.2785717328389486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,65536,0.30510079065958656
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,6144,16384,0.08716373443603516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,16384,0.08538133303324381
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,16384,0.06675093173980713
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,7168,16384,0.050641067822774256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,6144,12288,0.0405130664507548
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,6144,12288,0.07463680108388265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,12288,0.05695893367131551
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,12288,0.054340267181396486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,6144,10240,0.035309867064158125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,6144,10240,0.06392853260040283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,10240,0.04828266700108846
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,10240,0.04616533517837525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,6144,8192,0.05617280006408691
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,6144,8192,0.028691200415293376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,8192,0.039290666580200195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,8192,0.03810346523920695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,6144,7168,0.07113493283589681
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,7168,0.038865065574645995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,7168,0.03572160005569458
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,6144,6144,0.06461013158162435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,6144,6144,0.020062933365503945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,6144,0.03113600015640259
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,7168,7168,0.02581653396288554
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,6144,0.02902719974517822
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,6144,5120,0.0552512009938558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,6144,5120,0.01613759994506836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,5120,0.026097067197163898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,5120,0.027177600065867107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,6144,4096,0.04518826802571614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,6144,4096,0.013255467017491659
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,4096,0.02196906606356303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,4096,0.02136533260345459
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,6144,3584,0.04205333391825358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,6144,3584,0.012691199779510498
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,3584,0.019332265853881835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,3584,0.018923733631769815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,6144,3072,0.03919466733932495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,6144,3072,0.011249066392580668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,3072,0.011785599589347839
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,3072,0.01625599960486094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,6144,2560,0.010680533448855082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,6144,2560,0.03709333340326945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,2560,0.010436266660690308
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,2560,0.015213867028554281
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,6144,2048,0.03345813353856404
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,6144,2048,0.00906986693541209
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,2048,0.008636800448099773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,2048,0.012266666690508524
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,6144,1536,0.030958932638168336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,6144,1536,0.0076682666937510175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,1536,0.007635200023651123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,1536,0.010578133662541707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,6144,1024,0.02921066681543986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,6144,1024,0.006955733398596446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,1024,0.0059914668401082356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,1024,0.009048533439636231
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,6144,768,0.027177600065867107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,6144,768,0.006296533346176148
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,768,0.005456000069777171
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,768,0.008225066463152568
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,6144,512,0.02574399908383687
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,6144,512,0.005818666517734527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,512,0.004614399870236715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,512,0.0076223999261856076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,6144,256,0.023907200495402018
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,6144,256,0.005313066641489664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,256,0.004002133260170618
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,256,0.007107200225194295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,6144,128,0.022963200012842813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,6144,128,0.005027199784914652
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,128,0.0036138666172822317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,7168,2048,0.010284800330797832
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,128,0.006760533154010773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,6144,64,0.022626133759816487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,64,0.003499733408292135
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,64,0.007288533449172974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,6144,32,0.02229866584142049
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,32,0.003572266548871994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,32,0.00740479975938797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,5120,65536,0.17672319412231446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,5120,65536,0.2635829289754232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,65536,0.2440447966257731
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,5120,16384,0.08275413513183594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,65536,0.2953141212463379
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,7168,512,0.006320000191529592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,5120,16384,0.04347199996312459
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,16384,0.07118933200836182
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,16384,0.06535040140151978
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,5120,12288,0.07080106735229492
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,12288,0.05618133147557577
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,12288,0.05379946629206339
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,5120,10240,0.02794559995333354
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,5120,10240,0.060794667402903235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,10240,0.041427199045817056
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,10240,0.04238933324813843
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,5120,8192,0.024337067206700643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,5120,8192,0.0533461332321167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,8192,0.03339200019836426
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,8192,0.03601173162460327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,5120,7168,0.06933973630269369
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,5120,7168,0.022100265820821127
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,7168,0.029926399389902752
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,7168,0.032576000690460204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,5120,6144,0.06235520044962565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,5120,6144,0.01900906761487325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,6144,0.02693973382314046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,6144,65536,0.18234453201293946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,6144,0.027369600534439088
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,5120,5120,0.016774400075276693
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,5120,5120,0.05176213184992472
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,5120,0.02294720013936361
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,5120,0.023086933294932048
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,5120,4096,0.045450667540232345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,6144,16384,0.050611201922098795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,5120,4096,0.013954133788744608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,4096,0.018894932667414346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,4096,0.019425066312154134
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,5120,3584,0.04182933171590169
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,5120,3584,0.012778666615486146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,3584,0.011040000120798747
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,3584,0.018713599443435668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,5120,3072,0.03940266768137614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,5120,3072,0.01135040024916331
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,3072,0.009802666306495667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,3072,0.016167466839154564
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,5120,2560,0.036356266339619955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,5120,2560,0.01002346674601237
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,2560,0.008628267049789428
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,6144,7168,0.02578986684481303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,2560,0.014330666263898215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,5120,2048,0.032891732454299924
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,5120,2048,0.00860693355401357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,1536,0.010217600067456563
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,2048,0.007559466858704884
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,2048,0.01253973344961802
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,5120,1536,0.030576000610987347
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,1536,0.00653653343518575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,5120,1024,0.028257066011428834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,1024,0.005443199972311656
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,1024,0.008736000458399455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,5120,768,0.026529065767923993
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,5120,768,0.0062165334820747375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,768,0.00492799977461497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,768,0.008140799899895985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,5120,512,0.02505386670430501
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,5120,512,0.005910400052865346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,512,0.004244266450405121
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,512,0.0074325333038965866
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,5120,256,0.023611734310785927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,5120,256,0.005321600039800008
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,256,0.0038773333032925926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,5120,128,0.02289066712061564
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,256,0.007048533360163371
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,5120,128,0.004958933095137278
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,128,0.0035360001027584078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,128,0.006634666522343953
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,5120,64,0.022746666272481283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,64,0.003306666761636734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,64,0.00724480003118515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,5120,32,0.022105600436528525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,32,0.0034186666210492453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,32,0.007156266768773396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,4096,65536,0.15447360674540203
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,65536,0.1581173261006673
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,65536,0.1983850638071696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,4096,16384,0.07523307005564371
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,4096,16384,0.0400874654452006
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,4096,65536,0.2471029281616211
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,16384,0.041324798266092935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,16384,0.05569920142491659
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,4096,12288,0.06392319997151694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,12288,0.030913066864013673
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,12288,0.04439146518707275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,4096,10240,0.05532266696294149
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,4096,10240,0.026532266537348432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,10240,0.03869226773579915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,10240,0.028267733256022137
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,4096,8192,0.049652266502380374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,4096,8192,0.022132267554601036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,8192,0.03253226677576701
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,8192,0.023485867182413737
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,4096,7168,0.06537386576334635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,4096,7168,0.01920959949493408
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,7168,0.029658667246500653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,7168,0.02115839918454488
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,4096,6144,0.05809813340504964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,6144,0.018230400482813516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,6144,0.026650667190551758
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,4096,5120,0.049081599712371825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,4096,5120,0.014735999703407287
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,5120,0.01776319940884908
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,5120,0.01599360009034475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,4096,4096,0.04387199878692627
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,4096,0.010913067062695821
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,4096,0.013773866494496665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,4096,3584,0.04049386580785115
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,4096,3584,0.011784533659617107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,3584,0.008970666925112407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,3584,0.012863999605178833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,5120,12288,0.03165653347969055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,4096,3072,0.0379093329111735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,4096,3072,0.010331733028093974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,3072,0.008475733796755473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,3072,0.011416533589363098
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,4096,2560,0.03603733380635579
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,4096,2560,0.009026133020718892
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,2560,0.00740479975938797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,2560,0.010850133498509724
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,4096,2048,0.03205546736717224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,4096,2048,0.007834666470686594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,2048,0.0068917334079742435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,2048,0.009522133072217305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,4096,1536,0.029681066672007244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,4096,1536,0.007234133283297221
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,1536,0.0059797331690788266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,1536,0.008414933085441589
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,4096,1024,0.006239999830722809
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,4096,1024,0.028436267375946046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,1024,0.004862933357556661
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,1024,0.007514666517575581
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,4096,768,0.0268010675907135
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,4096,768,0.005926399926344554
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,768,0.00447680006424586
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,768,0.007101866602897644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,4096,512,0.024623999993006386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,512,0.004098133246103922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,512,0.006677333513895671
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,4096,256,0.023600000143051147
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,4096,256,0.005121066669623057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,256,0.003755733370780945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,256,0.006308266520500183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,4096,128,0.022797866662343343
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,4096,128,0.00490880012512207
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,128,0.0034122665723164878
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,128,0.006084266801675161
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,4096,64,0.023040000597635904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,64,0.0032469332218170166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,64,0.0064735998709996535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,4096,32,0.02377706567446391
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,32,0.003554133325815201
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,5120,1536,0.007547733187675476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,32,0.006497066716353099
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3584,65536,0.14284586906433105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3584,65536,0.23130133946736656
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,65536,0.18119786580403646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,65536,0.14371627171834309
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,5120,1024,0.006899199883143107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3584,16384,0.07305920124053955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3584,16384,0.038070400555928544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,16384,0.03979733387629191
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,16384,0.04989120165506999
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3584,12288,0.061341865857442224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3584,12288,0.02695573369661967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,12288,0.040480001767476397
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,12288,0.03033600052197774
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3584,10240,0.05393706560134888
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,10240,0.03486293156941732
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,10240,0.02534613410631816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3584,8192,0.04815786679585775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3584,8192,0.020444800456364952
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,8192,0.020962133010228475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,8192,0.029357866446177168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3584,7168,0.06325120131174723
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3584,7168,0.017806933323542277
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,7168,0.02672213315963745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,7168,0.0195360004901886
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3584,6144,0.0572106679280599
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3584,6144,0.016394666830698647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,6144,0.021356799205144248
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,6144,0.01671253244082133
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3584,5120,0.048019198576609294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3584,5120,0.014807466665903726
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,5120,0.011975466211636861
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,5120,0.015224533279736838
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3584,4096,0.04264853398005168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3584,4096,0.011925333738327026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,4096,0.00957546631495158
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,4096,0.013089066743850708
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3584,3584,0.040550398826599124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,3584,0.008865066369374593
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,3584,0.011858133474985759
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3584,3072,0.03744000196456909
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3584,3072,0.009713066617647807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,3072,0.00831573357184728
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,3072,0.010824533303578694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3584,2560,0.035257601737976076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3584,2560,0.008629332979520161
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,2560,0.007333333293596904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,2560,0.009970133503278095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,4096,12288,0.029315199454625445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3584,2048,0.03165546655654907
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3584,2048,0.007492266595363617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,2048,0.009164800246556599
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,2048,0.006562133133411407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3584,1536,0.006855466465155284
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3584,1536,0.030036266644795733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,1536,0.005957333246866862
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,1536,0.008201600114504496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3584,1024,0.028093866507212323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3584,1024,0.006097066899140676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,1024,0.004874666531880697
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,1024,0.007316266496976216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3584,768,0.0258026659488678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3584,768,0.005830400188763936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,768,0.004465066889921824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,4096,6144,0.016790399948755898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,768,0.0069578667481740315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3584,512,0.024699733654658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3584,512,0.005571199953556061
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,512,0.003975466638803482
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,512,0.006555733581384023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,4096,4096,0.012557866175969443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3584,256,0.023316266139348348
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3584,256,0.005189333359400431
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,256,0.0035648000737031303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,256,0.006206933160622915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3584,128,0.02254400054613749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3584,128,0.00494400014479955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,128,0.0034005333979924522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,128,0.005955199897289276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3584,64,0.02255679965019226
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,64,0.0032170665760835014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,64,0.0064085334539413456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3584,32,0.022805333137512207
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,32,0.003272533416748047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,32,0.006417066852251689
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3072,65536,0.1378378709157308
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,65536,0.15213972727457684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,65536,0.12959252993265788
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3072,16384,0.06969707012176514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3072,16384,0.03573226531346639
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3072,65536,0.22425386110941567
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,16384,0.043686401844024655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,16384,0.037828266620635986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3072,12288,0.059189331531524655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3072,12288,0.026758400599161784
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,12288,0.036166401704152425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,12288,0.02964266737302144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,4096,512,0.005694933235645294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3072,10240,0.05213226477305094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,10240,0.031148799260457355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,10240,0.024038400252660116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3072,8192,0.04671253363291423
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,8192,0.026349866390228273
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,8192,0.020364799102147422
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3072,7168,0.017091200749079386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3072,7168,0.06141120195388794
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,7168,0.01843520005544027
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,7168,0.02382826606432597
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3072,6144,0.055485868453979494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,6144,0.013190399607022604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,6144,0.01658986707528432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3072,5120,0.04813973506291707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,5120,0.011387733618418376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,5120,0.014456533392270408
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3072,4096,0.041832534472147624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3072,4096,0.011271466811498005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,4096,0.009486933549245197
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,4096,0.012449066837628682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3072,3584,0.03987199862798055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,3584,0.00892799993356069
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3584,10240,0.02418986757596334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,3584,0.011509333054224651
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3072,3072,0.03700266679128011
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3072,3072,0.009593600034713745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,3072,0.008281599978605907
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,3072,0.01053439974784851
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3072,2560,0.035094400246938065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3072,2560,0.00860800047715505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,2560,0.0073845331867535904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,2560,0.009739733735720317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3072,2048,0.03153066635131836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3072,2048,0.007928533355395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,2048,0.0064522668719291685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,2048,0.008903466661771138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3072,1536,0.029468800624211627
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3072,1536,0.006958933174610138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,1536,0.005815466741720835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,1536,0.007991466422875721
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3584,3584,0.011300266782442728
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3072,1024,0.027481599648793535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3072,1024,0.006126933296521505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,1024,0.0048426667849222815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,1024,0.007231999933719635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3072,768,0.02662613391876221
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,768,0.004452266792456309
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,768,0.0068906664848327635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3072,512,0.024522666136423746
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3072,512,0.0055871998270352686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,512,0.004058666775623957
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,512,0.006580266853173573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3072,256,0.023128533363342287
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3072,256,0.0051018665234247845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,256,0.0036415999134381616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,256,0.006266666452089946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3072,128,0.022549333175023396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3072,128,0.004889599978923798
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,128,0.0034346667428811393
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,128,0.005998933315277099
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3072,64,0.02239360014597575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,64,0.0032127998769283296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,64,0.006300800045331319
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,3072,32,0.022703999280929567
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,32,0.0033439998825391137
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,32,0.006400000055631001
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2560,65536,0.12499093214670817
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2560,65536,0.21864959398905434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,65536,0.1297770659128825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,65536,0.12027946313222249
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2560,16384,0.0670624017715454
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,16384,0.03773119846979777
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,16384,0.031245867411295574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2560,12288,0.05721813440322876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2560,12288,0.024987733364105223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,12288,0.030532266696294146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,12288,0.02505386670430501
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2560,10240,0.049415465195973715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2560,10240,0.020644267400105797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,10240,0.026998400688171387
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,10240,0.02086826761563619
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2560,8192,0.04328533411026001
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2560,8192,0.01817173361778259
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,8192,0.02170026699701945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,8192,0.018373332420984902
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2560,7168,0.059553066889444985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2560,7168,0.016198399662971496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3072,10240,0.022780799865722658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,7168,0.016269866625467935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,7168,0.017554134130477905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2560,6144,0.05300586620966593
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2560,6144,0.014907733599344889
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,6144,0.01660160024960836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,6144,0.015481600165367126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3072,8192,0.019107200702031455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2560,5120,0.046727466583251956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,5120,0.014688000082969666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,5120,0.013048533598581949
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3072,6144,0.0149536003669103
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2560,4096,0.0411135991414388
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,4096,0.01193173329035441
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3072,5120,0.013520000378290811
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,4096,0.011341866850852967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2560,3584,0.039297068119049074
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2560,3584,0.009998933474222819
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,3584,0.008573866883913676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,3584,0.010645332932472228
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2560,3072,0.03662186861038208
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2560,3072,0.00916373332341512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,3072,0.0099263995885849
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,3072,0.0099263995885849
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2560,2560,0.03444053332010905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2560,2560,0.008409600456555684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,2560,0.007326933244864146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,2560,0.009113599856694538
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2560,2048,0.031202133496602374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2560,2048,0.007439999779065449
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,2048,0.0065194666385650635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,2048,0.008407466610272725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2560,1536,0.029322665929794312
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2560,1536,0.006748799979686737
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,1536,0.005829333265622457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,1536,0.007709866762161255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2560,1024,0.027270400524139406
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2560,1024,0.006020266811052958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,1024,0.004840533435344696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3072,3584,0.010674132903416952
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,1024,0.007085866729418437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2560,768,0.02568639914194743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2560,768,0.005696000158786773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,768,0.004385066529115042
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,768,0.0067893331249554946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2560,512,0.024660267432530723
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2560,512,0.005468800167242686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,512,0.003980800012747447
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,512,0.00647573322057724
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2560,256,0.00505920002857844
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,256,0.006165333092212677
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2560,256,0.023409066597620647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,256,0.0036127999424934386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2560,128,0.022525866826375328
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2560,128,0.004820266862710317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,128,0.0032469332218170166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,128,0.0059445331494013464
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2560,64,0.02228800058364868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,64,0.003139200061559677
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,64,0.006233599781990051
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2560,32,0.02168853282928467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,32,0.0031989333530267083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,32,0.006211199859778086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3072,768,0.005838933090368906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,65536,0.10814932982126872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2048,65536,0.20104212760925294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,65536,0.09245866934458415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2048,16384,0.06580586830774943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2048,16384,0.03026240070660909
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,16384,0.03178453246752421
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,16384,0.026716800530751546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2048,12288,0.05585066477457682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,12288,0.026305067539215087
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,12288,0.02071040074030558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2048,10240,0.04843946695327759
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2048,10240,0.020317866404851278
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,10240,0.022377600272496544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,10240,0.018074667453765868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2048,8192,0.04154239892959595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2048,8192,0.01697280009587606
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,8192,0.014104533195495605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,8192,0.014450132846832275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2048,7168,0.0589738647143046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,7168,0.01681706706682841
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,7168,0.013864533106486002
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2048,6144,0.051038932800292966
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,6144,0.015650133291880287
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,6144,0.012294399738311767
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2048,5120,0.04569813410441081
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2048,5120,0.012087466319402058
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,5120,0.012116266290346782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,5120,0.011397332946459452
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2560,16384,0.034983468055725095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2048,4096,0.04091626803080241
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2048,4096,0.010571733117103577
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,4096,0.011105066537857056
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,4096,0.01025279959042867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2048,3584,0.03882666826248169
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2048,3584,0.009705600142478944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,3584,0.009744000434875489
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,3584,0.009799466530481974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2048,3072,0.036432000001271565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,3072,0.008699733018875121
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,3072,0.009092266360918682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2048,2560,0.03399680058161418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2048,2560,0.007886933286984761
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,2560,0.007729066908359528
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2560,5120,0.012642133235931396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,2560,0.00871573289235433
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2560,4096,0.010821333527565003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2048,2048,0.031125332911809283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2048,2048,0.007043200234572093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,2048,0.006963199873765309
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,2048,0.007990399996439617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2048,1536,0.029193600018819172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,1536,0.005662933488686879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,1536,0.007375999788443248
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2048,1024,0.027078400055567425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,1024,0.0047423998514811196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,1024,0.006917333106199901
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2048,768,0.02587839961051941
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2048,768,0.005623466769854228
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,768,0.004263466596603394
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,768,0.006758399804433187
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2048,512,0.024859732389450072
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2048,512,0.005498666564623515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,512,0.0038719999293486277
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,512,0.006403199831644694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2048,256,0.023415466149648033
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2048,256,0.005015466610590617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,256,0.003487999985615412
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,256,0.006138666470845541
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2048,128,0.022669865687688192
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2048,128,0.004755199948946635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,128,0.0032405334214369455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,128,0.0059445331494013464
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2048,64,0.022721066077550253
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,64,0.003054933249950409
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,64,0.006141866743564606
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,2048,32,0.022645332415898643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,32,0.0032490665713946023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,32,0.006093866626421611
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2048,65536,0.11262186368306477
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2048,12288,0.023603200912475586
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1536,65536,0.18986666997273763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,65536,0.08456532955169678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,65536,0.07462720076243082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1536,16384,0.06297173500061035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1536,16384,0.024493867158889772
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,16384,0.0214901328086853
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,16384,0.02622399926185608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1536,12288,0.05252480109532674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1536,12288,0.020351999998092653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2048,7168,0.015573333700497946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,12288,0.019127466281255088
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,12288,0.02002453406651815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1536,10240,0.046115199724833175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1536,10240,0.016370133558909098
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,10240,0.013389866550763449
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,10240,0.017543466885884602
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1536,8192,0.03916800022125244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1536,8192,0.014941866199175516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,8192,0.014522666732470194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,8192,0.01330560048421224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1536,7168,0.05608426729838053
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1536,7168,0.013372799754142762
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2048,6144,0.01369599997997284
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,7168,0.013753599921862283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,7168,0.014101333419481912
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,6144,0.01237226625283559
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1536,6144,0.050629333655039466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1536,6144,0.01183786690235138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,6144,0.012814933061599731
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1536,5120,0.0449343999226888
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1536,5120,0.010788266857465107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,5120,0.010436266660690308
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,5120,0.011873066425323486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1536,4096,0.040540798505147295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1536,4096,0.009171199798583985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,4096,0.008918399612108868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,4096,0.01030399998029073
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1536,3584,0.03863360087076823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,3584,0.008227199812730153
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,3584,0.009656533598899841
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1536,3072,0.035973334312438966
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1536,3072,0.008098133405049642
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2048,3072,0.00902186632156372
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,3072,0.0074976002176602675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,3072,0.00876693328221639
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1536,2560,0.03376746575037638
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1536,2560,0.007479466497898102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,2560,0.006849066913127899
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,2560,0.008337066570917765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1536,2048,0.03075733383496602
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,2048,0.0060127998391787205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,2048,0.007749333480993907
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1536,1536,0.02908266584078471
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1536,1536,0.0064735998709996535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,1536,0.00537066658337911
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,1536,0.007417599856853485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1536,1024,0.02725013295809428
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1536,1024,0.005901866654555003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,1024,0.004774400095144907
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,1024,0.00689386675755183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1536,768,0.02621440092722575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1536,768,0.0055167997876803074
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,768,0.006664533416430156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,768,0.0043818667531013485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1536,512,0.024292266368865965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1536,512,0.0054058666030565895
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,512,0.003908266623814901
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,512,0.006410666803518932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1536,256,0.022894932826360067
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1536,256,0.004955733319123586
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,256,0.0034634667138258614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,256,0.006141866743564606
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1536,128,0.022445867458979286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1536,128,0.004744533201058706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,128,0.0031829332311948144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,128,0.005925333499908448
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1536,64,0.022066134214401244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2048,1536,0.006413866579532623
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,64,0.0030752000709374744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,64,0.005981866518656413
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1536,32,0.021460266908009847
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,32,0.003151999910672506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,32,0.006037333110968272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2048,1024,0.005767466624577841
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,65536,0.06581546862920126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1024,65536,0.19365545908610027
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,65536,0.07866133054097493
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1024,16384,0.06043519973754883
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1024,16384,0.026537599166234334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,16384,0.01395626664161682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,16384,0.023788799842198692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1024,12288,0.051006933053334556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1024,12288,0.020713599522908528
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,12288,0.011158399780591329
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,12288,0.018780799706776936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1024,10240,0.04380160172780355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,10240,0.012528000275293985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,10240,0.01688106656074524
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1024,8192,0.03776106834411621
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1024,8192,0.015545599659283958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,8192,0.014461867014567056
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,8192,0.013134933511416116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1024,7168,0.05463146766026815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,7168,0.011911466717720032
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,7168,0.013335466384887695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1024,6144,0.012199466427167256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1024,6144,0.049303468068440756
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,6144,0.01111253301302592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,6144,0.011288533608118694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1024,5120,0.04511786699295044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1024,5120,0.01111466685930888
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,5120,0.009797333677609762
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,5120,0.010577066739400228
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1024,4096,0.039902933438618976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1024,4096,0.009405866265296936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,4096,0.008706133564313252
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,4096,0.009461333354314167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1024,3584,0.038446935017903645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1024,3584,0.00888320008913676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,3584,0.008028799792130788
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,3584,0.009033600489298504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1024,3072,0.03573760191599528
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1024,3072,0.008201600114504496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,3072,0.0073173334201176955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,3072,0.008624000350634257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1024,2560,0.03349546591440837
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1024,2560,0.007326933244864146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,2560,0.00647573322057724
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,2560,0.008156799773375193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1024,2048,0.030801065762837726
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1024,2048,0.006605866551399231
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,2048,0.00600853314002355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,2048,0.00769706666469574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1024,1536,0.029014400641123456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1024,1536,0.00624533345301946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1536,65536,0.09277013142903646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,1536,0.005031466484069824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,1536,0.00718506673971812
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1024,1024,0.027131734291712443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1024,1024,0.005596800148487091
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,1024,0.004356266558170318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,1024,0.006836266815662384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1024,768,0.025333333015441894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1024,768,0.005418666700522105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,768,0.004062933226426443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,768,0.006559999783833821
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1024,512,0.024693334102630617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1024,512,0.005264000097910563
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,512,0.0036490666369597114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,512,0.006302933394908905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1024,256,0.022875734170277915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1024,256,0.0048320000370343525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,256,0.003390933324893316
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,256,0.006067200005054474
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1024,128,0.02238933245340983
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1024,128,0.004729599754015604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,128,0.0031797334551811217
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,128,0.005807999769846598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1024,64,0.02254400054613749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,64,0.003009066730737686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,64,0.005901866654555003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,1024,32,0.022428800662358604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,32,0.0030229332546393077
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,32,0.005843199789524078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,768,65536,0.1787999947865804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,65536,0.05475093523661295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1536,3584,0.008724266290664673
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,65536,0.06904106934865316
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,768,16384,0.05880959828694662
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,768,16384,0.020938666661580403
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,16384,0.01686826745669047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,16384,0.019987199703852335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,768,12288,0.0488480011622111
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,768,12288,0.016556800405184428
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,12288,0.013691733280817667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,12288,0.01648533344268799
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,768,10240,0.04288746515909831
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,768,10240,0.014734933773676554
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,10240,0.01227839986483256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,10240,0.014794666568438211
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,768,8192,0.03643626769383748
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,768,8192,0.012607999642690024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,8192,0.010807466506958009
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,8192,0.012905599673589072
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,768,7168,0.011667199929555257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,768,7168,0.05387946764628092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,7168,0.010776533683141073
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1536,2048,0.006905599931875865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,7168,0.012768000364303589
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,768,6144,0.04927253325780233
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,6144,0.009430399537086487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,6144,0.011242666840553283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,768,5120,0.04450773398081462
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,768,5120,0.009666132926940917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,5120,0.00881706674893697
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,5120,0.010567466417948406
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,768,4096,0.039665067195892335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,4096,0.008124800026416778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,4096,0.009717333316802978
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,768,3584,0.03784853219985962
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,768,3584,0.008155733346939087
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,3584,0.0077450667818387345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,3584,0.009127466877301534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,768,3072,0.03594133456548055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,768,3072,0.007609599828720092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,3072,0.007415466507275899
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,3072,0.008631466825803121
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,768,2560,0.03333866596221924
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,768,2560,0.007119999825954437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,2560,0.006471466521422069
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,2560,0.008282666901747386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,768,2048,0.03068373401959737
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,768,2048,0.006629333396752675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,2048,0.005749333401521047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,2048,0.007701333363850911
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,768,1536,0.02898453275362651
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,768,1536,0.0062047998110453285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,1536,0.0050335998336474095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,1536,0.007285333176453908
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,768,1024,0.02695573369661967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,768,1024,0.005596800148487091
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1024,65536,0.0913535992304484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,1024,0.004407466451327006
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,1024,0.006810666620731353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,768,768,0.02540160020192464
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,768,768,0.005402666827042898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,768,0.004057600100835165
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,768,0.00659093310435613
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,768,512,0.024173865715662636
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,768,512,0.005129600067933401
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,512,0.003681066632270813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,512,0.006248533229033152
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,768,256,0.02281279961268107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,256,0.0034058667719364167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,256,0.0060810665289560955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,768,128,0.022342399756113688
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,768,128,0.004669866462548574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,128,0.0030879999200503034
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,128,0.005919999877611796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,768,64,0.021921066443125407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,64,0.0029663999875386557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,64,0.005873066683610281
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,768,32,0.02139520049095154
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,32,0.003019733230272929
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,32,0.005921066800753275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1024,10240,0.018012799819310508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,512,65536,0.07341333230336508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,65536,0.04231893221537272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,512,65536,0.18411839803059896
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1024,7168,0.013965866963068643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,65536,0.07002560297648111
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,512,16384,0.05787520011266073
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,512,16384,0.020731733242670695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,16384,0.012770133217175803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,16384,0.0207914670308431
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,512,12288,0.048051198323567704
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,512,12288,0.01657600005467733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,12288,0.010899200042088827
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,512,10240,0.01446293294429779
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,12288,0.01695466637611389
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,512,10240,0.040515200297037764
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,10240,0.010028800368309021
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,10240,0.01507306694984436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,512,8192,0.036073601245880126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,512,8192,0.012521599729855856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,8192,0.009091200431187947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,8192,0.013165866335233053
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,512,7168,0.053293867905934655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,7168,0.008706133564313252
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,7168,0.012246400117874146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,512,6144,0.049160532156626385
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,6144,0.008167466521263123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,6144,0.01130986710389455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,512,5120,0.044735999902089436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,512,5120,0.009523199995358785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,5120,0.0077461332082748415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,5120,0.010590933760007222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,512,4096,0.039640533924102786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,512,4096,0.008552533388137818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,4096,0.0072970668474833175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,4096,0.00939520001411438
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,512,3584,0.03796693483988444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,3584,0.007561600208282471
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,3584,0.008968533078829447
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,512,3072,0.03577599922815959
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,768,65536,0.08070933024088542
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,3072,0.007241599758466085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,3072,0.008522666494051616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,512,2560,0.03333333333333333
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,512,2560,0.006906666855017345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,2560,0.006426666676998138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,2560,0.008126933375994365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,512,2048,0.03070080081621806
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,512,2048,0.0064640000462532045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,2048,0.005723733206590017
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,2048,0.007594666878382365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,512,1536,0.028784000873565675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,512,1536,0.006042666733264923
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,1536,0.0050357331832249965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,1536,0.007218133409818013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,512,1024,0.026841600735982258
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,768,6144,0.01072106659412384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,1024,0.004374399781227112
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,1024,0.006790400048096975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,768,4096,0.008642133076985676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,512,768,0.025259733200073242
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,512,768,0.005343999962011973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,768,0.004060799876848856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,768,0.006385066608587901
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,512,512,0.024090667565663658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,512,512,0.005066666503747304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,512,0.0036533333361148832
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,512,0.0061941335598627726
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,512,256,0.022810665766398112
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,512,256,0.004791466891765595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,256,0.0033269333342711128
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,256,0.005972266693909963
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,512,128,0.004653866589069367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,512,128,0.022369066874186196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,128,0.003070933371782303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,128,0.005852800110975901
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,512,64,0.02250880002975464
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,64,0.0029525332152843474
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,64,0.0058783998092015585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,512,32,0.021240532398223877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,32,0.002948266764481862
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,32,0.005859200159708659
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,768,256,0.004849066833655039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,256,65536,0.17513173421223957
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,65536,0.02768320043881734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,65536,0.06679680347442626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,256,16384,0.054058667023976645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,256,16384,0.020797866582870483
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,16384,0.009306666254997254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,16384,0.019806933403015137
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,256,12288,0.04474666515986125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,256,12288,0.016670932372411094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,12288,0.008392533659934998
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,12288,0.016267733772595723
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,256,10240,0.03989439805348714
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,10240,0.007985066870848339
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,10240,0.014734933773676554
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,256,8192,0.03572266499201457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,512,7168,0.011621333161989848
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,8192,0.0075573335091273
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,8192,0.01285546620686849
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,512,6144,0.01060693363348643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,256,7168,0.05329813162485758
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,256,7168,0.011480533083279927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,7168,0.0073728000124295544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,7168,0.012058666348457337
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,256,6144,0.010557867089907328
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,256,6144,0.048724265893300374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,6144,0.007051733136177063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,6144,0.011190399527549744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,256,5120,0.04466346502304077
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,256,5120,0.009511466821034749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,512,3584,0.0081045334537824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,5120,0.007457066575686138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,5120,0.010442666212717692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,256,4096,0.03970880111058553
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,4096,0.007100800176461537
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,4096,0.00934933324654897
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,256,3584,0.037777066230773926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,512,3072,0.007593599955240886
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,3584,0.00740479975938797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,3584,0.008940800031026205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,256,3072,0.035784534613291424
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,256,3072,0.007534933090209961
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,3072,0.0070720002055168155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,3072,0.008377599716186523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,256,2560,0.033241599798202515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,256,2560,0.00692799985408783
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,2560,0.00625493327776591
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,2560,0.008110933502515157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,256,2048,0.030638933181762695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,256,2048,0.0064298664530118305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,2048,0.005592533449331919
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,2048,0.007495466868082683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,256,1536,0.028832000494003297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,256,1536,0.006017066538333893
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,1536,0.00489279975493749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,1536,0.007062399884064992
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,256,1024,0.026814933617909747
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,1024,0.004223999877770742
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,1024,0.0065653334061304735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,256,768,0.02537493308385213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,512,1024,0.005559466779232025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,768,0.0038965334494908653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,768,0.00645653357108434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,256,512,0.024038400252660116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,256,512,0.004977066814899444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,512,0.0035989334185918174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,512,0.00618453323841095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,256,256,0.022884267568588256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,256,256,0.004901333153247834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,256,0.003239466746648153
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,256,0.005940266450246175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,256,128,0.022312533855438233
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,256,65536,0.07301119963328043
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,128,0.003099733342727025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,256,8192,0.012453333536783854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,128,0.005761066575845083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,256,64,0.022060799598693847
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,64,0.0029578665892283124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,64,0.005838933090368906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,256,32,0.021249065796534218
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,32,0.0029205332199732465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,32,0.0058005332946777345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,256,10240,0.0145087997118632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,65536,0.01370133360226949
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,128,65536,0.07138773600260416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,128,65536,0.17481066385904948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,65536,0.06291840076446534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,128,16384,0.05311466852823893
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,128,16384,0.020734934012095134
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,16384,0.00735040009021759
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,16384,0.019690666596094766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,128,12288,0.04411626656850179
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,128,12288,0.016642133394877114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,12288,0.007303466896216075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,12288,0.016207999984423318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,128,10240,0.03985600074132283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,128,10240,0.014532267053922018
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,10240,0.014492799838383993
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,10240,0.0073066666722297665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,128,8192,0.012457600235939026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,128,8192,0.03582079807917277
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,8192,0.007299200197060903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,8192,0.012589866916338602
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,128,7168,0.053201067447662356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,7168,0.007102933526039123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,128,7168,0.01141439974308014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,7168,0.01184213360150655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,128,6144,0.04898560047149658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,6144,0.0071487997968991595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,6144,0.010957866907119751
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,128,5120,0.009504000345865887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,128,5120,0.044222935040791826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,5120,0.007325866818428039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,5120,0.010262399911880493
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,128,4096,0.039795200030008956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,128,4096,0.008504533767700195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,4096,0.009179733196894328
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,256,3584,0.00806826651096344
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,4096,0.006903466582298279
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,256,4096,0.008500267068545024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,128,3584,0.03791466554005941
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,3584,0.0073749333620071415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,3584,0.008683733145395915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,128,3072,0.03546666701634725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,128,3072,0.007457066575686138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,3072,0.0070613334576288865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,3072,0.008330666522185007
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,128,2560,0.03315946658452352
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,128,2560,0.00684799998998642
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,2560,0.00629013329744339
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,2560,0.007851733267307282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,128,2048,0.030454399188359578
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,2048,0.00556160012880961
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,2048,0.007347199817498524
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,128,1536,0.02926186720530192
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,128,1536,0.005929600199063619
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,1536,0.0048650667071342465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,1536,0.00701333334048589
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,128,1024,0.027458133300145467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,256,1024,0.0055871998270352686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,1024,0.0042912001411120095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,1024,0.006549333532651265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,256,768,0.005288533369700114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,128,768,0.025253333648045856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,768,0.003923200070858002
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,128,768,0.005201066533724466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,768,0.0063391998410224915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,128,512,0.02408000032107035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,128,512,0.005011199911435445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,512,0.0035914666950702667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,512,0.006083199878533682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,128,256,0.022689066330591836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,128,256,0.004713599880536398
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,256,0.0032149332265059153
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,256,0.005926399926344554
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,128,128,0.02214933236440023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,128,128,0.004484266539414724
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,128,0.0030645333230495454
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,128,0.005739733576774597
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,128,64,0.021718400716781616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,64,0.0029535998900731405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,64,0.0057429333527882894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,128,32,0.021016534169514975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,32,0.002937600016593933
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,32,0.005765333275000254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,65536,0.011621333161989848
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,64,65536,0.17492799758911132
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,65536,0.061926400661468504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,64,16384,0.05429866711298624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,16384,0.007150933146476746
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,256,128,0.004570666452248891
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,16384,0.019962666432062785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,64,12288,0.04443306525548299
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,12288,0.007062399884064992
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,64,10240,0.03894613186518352
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,12288,0.016025599837303162
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,10240,0.007325866818428039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,10240,0.014338133732477823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,64,8192,0.035692799091339114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,8192,0.0071263998746871945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,8192,0.012572800119717916
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,64,7168,0.05304213364919027
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,7168,0.007036800185839335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,7168,0.011734400192896526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,64,6144,0.048750933011372885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,6144,0.006872533261775971
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,6144,0.01090239981810252
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,64,5120,0.04398933251698812
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,5120,0.0071936001380284624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,5120,0.010137599706649781
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,64,4096,0.03922026554743449
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,4096,0.006867200136184692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,4096,0.009051733215649923
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,64,3584,0.037324798107147214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,3584,0.00730453332265218
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,64,3072,0.03516053358713786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,3072,0.006961066524187725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,3584,0.008646399776140849
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,3072,0.008357333143552144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,64,2560,0.0332479993502299
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,2560,0.006223999957243601
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,2560,0.007859200239181519
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,64,2048,0.030405332644780476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,2048,0.005584000051021576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,2048,0.00732479989528656
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,1536,0.004955733319123586
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,64,1536,0.028497066100438433
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,1536,0.006941866874694824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,1024,0.004208000004291534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,64,1024,0.026734934250513716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,1024,0.006841599941253662
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,64,768,0.025193599859873454
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,768,0.0038634667793909705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,768,0.006420266628265381
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,64,512,0.02397759954134623
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,512,0.0035487999518712364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,512,0.006144000093142191
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,64,256,0.022689066330591836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,256,0.005906133353710175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,256,0.0032650666932264962
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,64,128,0.022066134214401244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,128,0.0030229332546393077
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,128,0.0057429333527882894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,64,64,0.02175040046374003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,64,0.0029120000700155893
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,64,0.005747200051943461
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,32,0.005762133498986562
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,64,32,0.021022933721542358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,32,0.0028618666032950084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,65536,0.01134933332602183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,65536,0.06148053407669067
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,32,65536,0.1752138614654541
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,32,16384,0.052501332759857175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,16384,0.0070709332823753355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,16384,0.019497599204381308
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,32,12288,0.04378773371378581
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,12288,0.006921599805355072
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,10240,0.007110400001207988
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,12288,0.016109866897265117
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,32,10240,0.039289601643880206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,10240,0.01446613371372223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,32,8192,0.03563946485519409
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,8192,0.007048533360163371
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,8192,0.012549333771069846
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,32,7168,0.05306773185729981
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,7168,0.006975999971230824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,7168,0.011734400192896526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,32,6144,0.048582398891448976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,6144,0.006795733173688252
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,6144,0.010945066809654236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,32,5120,0.04388906558354695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,5120,0.007166933516661327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,5120,0.010193066795667012
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,32,4096,0.03929493427276611
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,4096,0.006730666756629944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,4096,0.009155199925104777
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,32,3584,0.037484800815582274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,3584,0.00697386662165324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,3584,0.008821333448092144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,3072,0.006701866785685222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,32,3072,0.03548373381296794
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,3072,0.008444799979527792
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,2560,0.006231466432412466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,32,2560,0.032986666758855185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,2560,0.007830399771531422
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,32,2048,0.03020159999529521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,2048,0.005526400109132131
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,2048,0.007505066692829132
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,128,6144,0.010593066612879436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,32,1536,0.028571732838948566
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,1536,0.004839466512203216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,1536,0.007153066496054332
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,32,1024,0.02664426763852437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,1024,0.004197333256403605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,768,0.003857066730658213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,32,768,0.025310933589935303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,1024,0.006485333542029063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,768,0.0062730665008227035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,32,512,0.023995733261108397
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,512,0.003525333354870478
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,512,0.006080000102519989
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,32,256,0.022693334023157756
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,256,0.0032032000521818793
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,256,0.00589333325624466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,32,128,0.022153600056966146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,128,0.0030048000315825146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,128,0.005712000032265982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,128,3584,0.008032000064849854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,32,64,0.021635200579961142
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,64,0.0028778667251269023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,96,32,32,0.020989867051442464
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,64,0.005714133381843567
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,32,0.0028138667345046996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,32,0.005723733206590017
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,128,2048,0.006374399860699971
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,65536,16384,0.6966933568318685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,65536,16384,0.833507219950358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,65536,12288,0.3627968152364095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,65536,16384,0.44527359008789064
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,65536,12288,0.5975711822509766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,65536,12288,0.5407061258951823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,128,1024,0.005460266768932342
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,65536,10240,0.23656853040059408
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,65536,10240,0.28315200805664065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,65536,10240,0.4369632085164388
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,65536,8192,0.1943893273671468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,65536,8192,0.23442773818969725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,65536,10240,0.5318133354187011
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,65536,8192,0.3871295928955078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,65536,8192,0.3747093200683594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,65536,7168,0.16822080612182616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,65536,7168,0.3456789334615072
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,65536,7168,0.3207082748413086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,65536,7168,0.22907093365987144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,65536,6144,0.1466815948486328
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,65536,6144,0.20280319849650064
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,65536,6144,0.2971487998962402
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,65536,6144,0.2735871950785319
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,65536,5120,0.12539626757303873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,65536,5120,0.18014933268229166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,65536,5120,0.2517141342163086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,65536,4096,0.10151573022206624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,65536,5120,0.2343381404876709
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,65536,4096,0.14916586875915527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,65536,4096,0.17533547083536785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,65536,4096,0.19741759300231934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,65536,3584,0.09203946590423584
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,65536,3584,0.13547840118408203
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,65536,3584,0.17995840708414715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,65536,3584,0.1618826707204183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,65536,3072,0.1547061284383138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,65536,2560,0.07037333647410074
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,65536,3072,0.12366613547007244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,65536,2560,0.11023146311442059
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,65536,3072,0.14040212631225585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,65536,2560,0.12966612974802655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,65536,2560,0.12258133093516033
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,65536,2048,0.05713599920272827
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,65536,2048,0.10461226304372151
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,65536,2048,0.09793386459350586
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,65536,1536,0.0456714669863383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,65536,2048,0.09332799911499023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,65536,1536,0.0841386636098226
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,65536,1536,0.08163306713104249
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,65536,1024,0.03342080116271973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,65536,1536,0.0699722687403361
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,65536,1024,0.05658026536305746
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,65536,1024,0.071234130859375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,65536,1024,0.04723200003306071
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,65536,768,0.06475199858347574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,65536,768,0.03587413231531779
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,65536,768,0.045821865399678544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,65536,512,0.05782399972279867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,65536,512,0.022359466552734374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,65536,512,0.03314773241678874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,65536,512,0.027808000644048054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,65536,256,0.04491733312606812
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,65536,256,0.012248532970746358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,65536,256,0.017464532454808553
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,65536,256,0.018720000982284546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,65536,128,0.039574400583903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,65536,128,0.007891199986139933
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,65536,16384,0.3724938710530599
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,65536,12288,0.282257080078125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,65536,128,0.016547200083732606
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,65536,64,0.039851733048756915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,65536,64,0.024075732628504435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,65536,32,0.007256533205509186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,65536,32,0.02574399908383687
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,65536,64,0.007107200225194295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,65536,32,0.03938773473103841
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,16384,16384,0.14693439801534017
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,16384,65536,0.4057738622029622
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,16384,65536,0.510208002726237
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,65536,0.7089375813802083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,65536,0.8127189636230469
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,16384,0.1909557342529297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,16384,0.17950933774312336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,16384,12288,0.11854080359141032
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,12288,0.1460693359375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,16384,10240,0.06614933411280313
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,12288,0.12933759689331054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,16384,10240,0.10280959606170655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,10240,0.12057279745737712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,10240,0.1111946662267049
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,16384,8192,0.05454399983088175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,16384,8192,0.08606826464335124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,8192,0.09459520181020101
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,8192,0.10302186806996663
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,16384,7168,0.09838293393452963
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,16384,7168,0.04837226470311483
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,7168,0.09118613402048746
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,7168,0.08032533327738443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,16384,6144,0.08826346397399902
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,16384,6144,0.04415359894434611
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,6144,0.0688693364461263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,6144,0.07515412966410319
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,16384,5120,0.07902293205261231
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,16384,5120,0.03691626787185669
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,5120,0.06316479841868082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,5120,0.05959893465042114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,16384,4096,0.06785919666290283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,16384,4096,0.03126186728477478
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,4096,0.045638398329416914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,4096,0.05422826608022054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,16384,3584,0.028347732623418172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,65536,3072,0.08594026565551757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,16384,3584,0.06308799982070923
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,3584,0.04218560059865316
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,3584,0.048097066084543866
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,16384,3072,0.05884693463643392
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,16384,3072,0.025198932488759356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,3072,0.03993279933929443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,3072,0.036714665095011395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,16384,2560,0.054024533430735266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,16384,2560,0.02115839918454488
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,2560,0.03572906653086345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,2560,0.03194453318913777
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,16384,2048,0.04521919886271159
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,16384,2048,0.014935466647148132
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,2048,0.029305599133173627
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,2048,0.02600533366203308
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,16384,1536,0.011985066533088683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,16384,1536,0.039698131879170734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,1536,0.0233514666557312
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,1536,0.020080000162124634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,16384,1024,0.00958079993724823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,16384,1024,0.03412266572316487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,1024,0.010003200173377991
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,1024,0.015006933609644571
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,16384,768,0.03125866651535034
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,768,0.008278400202592214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,768,0.01241386632124583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,16384,512,0.029022934039433797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,16384,512,0.007793066898981731
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,512,0.006358399987220764
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,65536,768,0.028252800305684406
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,512,0.010635733604431152
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,16384,256,0.026796799898147584
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,256,0.005052799979845682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,256,0.008896000186602275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,16384,128,0.025283199548721314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,16384,128,0.006031999985376993
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,128,0.004227200150489807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,128,0.008265600105126699
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,16384,64,0.025434666872024538
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,64,0.00396373321612676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,64,0.009839999675750732
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,16384,32,0.026039467255274458
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,32,0.004075733323891958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,32,0.010145066181818645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,65536,128,0.0100490669409434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,12288,65536,0.2956181208292643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,65536,0.5720810572306315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,12288,65536,0.43788585662841795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,12288,16384,0.12747626304626464
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,12288,16384,0.09169812997182211
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,65536,0.5417450586954753
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,16384,0.13341439565022786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,16384,0.154695463180542
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,12288,12288,0.06236480077107748
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,12288,12288,0.10033386548360188
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,12288,0.11580800215403239
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,12288,0.09589440027872721
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,16384,16384,0.10063786506652832
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,12288,10240,0.08854400316874186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,10240,0.09773120085398355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,10240,0.08187306722005208
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,12288,8192,0.05045013427734375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,12288,8192,0.07623253663380941
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,8192,0.0801472028096517
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,8192,0.06988053321838379
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,16384,12288,0.07867946624755859
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,12288,7168,0.08973973592122396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,7168,0.0695520003636678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,7168,0.06025600035985311
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,12288,6144,0.034815998872121175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,6144,0.06100373268127442
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,12288,6144,0.08031360308329263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,6144,0.052245334784189856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,12288,5120,0.07246507008870443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,5120,0.050596264998118076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,5120,0.04559893210728963
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,12288,4096,0.02515733242034912
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,12288,4096,0.06297813256581625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,4096,0.04208533366521199
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,4096,0.035837864875793456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,12288,3584,0.0231605331103007
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,12288,3584,0.058662398656209314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,3584,0.03684373299280803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,3584,0.03302186727523804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,12288,3072,0.05458026727040609
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,12288,3072,0.017604267597198485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,3072,0.03231146732966105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,3072,0.02802666624387105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,12288,2560,0.0473685344060262
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,12288,2560,0.013416533668835958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,2560,0.02768213351567586
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,2560,0.02502506573994954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,12288,2048,0.041383465131123856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,12288,2048,0.011971199512481689
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,2048,0.02344640096028646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,2048,0.020173867543538414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,12288,1536,0.036610134442647296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,12288,1536,0.009810133775075277
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,1536,0.010904533664385478
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,1536,0.016141866644223533
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,12288,1024,0.03288639982541402
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,1024,0.008141866823037466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,1024,0.01266986628373464
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,12288,768,0.031199999650319415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,12288,768,0.0077354664603869125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,768,0.007090133428573608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,768,0.01088533302148183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,12288,512,0.028880000114440918
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,12288,512,0.0069930667678515124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,512,0.005901866654555003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,512,0.009548800190289815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,12288,256,0.026470400889714557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,12288,256,0.006060799956321717
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,256,0.004660266637802124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,256,0.008075733482837678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,12288,128,0.025381332635879515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,12288,128,0.0056085333228111265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,128,0.004074666649103165
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,128,0.007429333527882893
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,12288,64,0.025399466355641682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,64,0.0036586667100588477
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,16384,768,0.008852266271909078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,64,0.008676266670227051
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,12288,32,0.02490880091985067
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,32,0.0038624001046021783
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,32,0.008849066495895386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,16384,256,0.0064640000462532045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,10240,65536,0.25646613438924154
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,65536,0.49187733332316086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,10240,16384,0.11838719844818116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,10240,65536,0.39126399358113606
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,65536,0.4681429227193196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,10240,16384,0.06957013607025146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,10240,12288,0.09494187037150065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,16384,0.11198613643646241
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,16384,0.1551487922668457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,12288,0.09202133019765218
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,12288,0.08357439835866293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,10240,10240,0.046937600771586103
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,10240,10240,0.08335253397623697
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,10240,0.07814506689707437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,10240,0.07193280061086019
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,10240,8192,0.037894399960835774
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,10240,8192,0.0721407969792684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,8192,0.06287786563237509
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,8192,0.060812799135843906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,10240,7168,0.08570453325907389
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,10240,7168,0.03493333260218302
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,7168,0.05525333484013876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,7168,0.05165013472239176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,10240,6144,0.07744853496551514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,10240,6144,0.030636799335479737
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,6144,0.049453866481781
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,6144,0.046581331888834634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,10240,5120,0.06935146649678549
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,5120,0.041099735101064044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,12288,10240,0.053098666667938235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,5120,0.040377600987752275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,10240,4096,0.060540799299875894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,10240,4096,0.023870933055877685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,4096,0.03397866487503052
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,4096,0.03171093265215556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,12288,7168,0.03931093215942383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,10240,3584,0.055333332220713294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,3584,0.031414399544397995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,3584,0.03426239887873332
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,10240,3072,0.05020906527837118
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,3072,0.027229867378870648
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,3072,0.026173865795135497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,10240,2560,0.046223998069763184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,10240,2560,0.01241919994354248
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,12288,5120,0.03022186756134033
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,2560,0.023379200696945192
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,2560,0.02188053329785665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,10240,2048,0.011133866508801778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,10240,2048,0.03902186552683513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,2048,0.017722666263580322
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,2048,0.017883733908335368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,10240,1536,0.03584853410720825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,1536,0.0093450665473938
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,1536,0.015126400192578635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,10240,1024,0.03333013256390889
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,1024,0.007018666466077168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,1024,0.012014933427174886
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,10240,768,0.03101759950319926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,10240,768,0.007449600100517273
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,768,0.006190933287143707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,768,0.010217600067456563
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,10240,512,0.028827732801437377
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,10240,512,0.0067221333583196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,512,0.00525546669960022
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,512,0.00879146655400594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,12288,1024,0.0086517333984375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,10240,256,0.02654506762822469
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,256,0.004370133578777313
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,256,0.008024533092975617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,10240,128,0.025544534126917522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,10240,128,0.005485866467158
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,128,0.004001066585381826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,128,0.007496533294518788
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,10240,64,0.025383466482162477
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,64,0.0037098666032155357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,64,0.008154666423797608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,10240,32,0.025820799668629962
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,32,0.0036981334288915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,32,0.008500267068545024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,8192,65536,0.308350944519043
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,8192,65536,0.23250346183776854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,10240,12288,0.054324265321095785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,65536,0.39799572626749674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,65536,0.3503360112508138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,8192,16384,0.09598399798075358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,8192,16384,0.06275413433710733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,16384,0.1104202667872111
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,16384,0.08677546977996826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,8192,12288,0.07788799603780111
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,12288,0.07255893548329671
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,12288,0.06459840138753256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,8192,10240,0.06861973603566487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,8192,10240,0.03717759847640991
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,10240,5120,0.02671146591504415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,10240,0.06089280049006144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,10240,0.0555733323097229
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,8192,8192,0.06007786591847738
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,8192,8192,0.031013333797454835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,10240,3584,0.016394666830698647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,8192,0.04983573357264201
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,8192,0.047042131423950195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,8192,7168,0.02780906756718953
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,7168,0.04394773244857788
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,10240,3072,0.014139733711878457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,8192,7168,0.07474133173624674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,7168,0.04142613410949707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,8192,6144,0.06776959896087646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,8192,6144,0.025013333559036253
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,6144,0.03902933200200399
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,6144,0.03654719988505046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,8192,5120,0.02246613303820292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,8192,5120,0.060362664858500156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,5120,0.033138134082158405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,5120,0.030382933219273885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,8192,4096,0.05034773349761963
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,10240,1536,0.00922986666361491
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,4096,0.02722346584002177
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,4096,0.025245867172876996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,8192,3584,0.04554133415222168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,8192,3584,0.015819733341534935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,3584,0.024295467138290405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,10240,1024,0.008016000191370647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,3584,0.023084799448649086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,8192,3072,0.013843199610710144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,8192,3072,0.04132159948348999
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,3072,0.021563732624053956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,3072,0.019950934251149497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,8192,2560,0.03739519913991292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,8192,2560,0.012405332922935487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,2560,0.016669867436091106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,2560,0.017798399925231932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,8192,2048,0.033802668253580734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,8192,2048,0.010875733693440755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,2048,0.010248532891273499
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,2048,0.014907733599344889
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,8192,1536,0.03141760031382243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,8192,1536,0.009083732962608337
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,1536,0.008550399541854858
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,1536,0.012610133488972983
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,8192,1024,0.029132799307505293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,10240,256,0.005820799867312113
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,1024,0.006804266571998596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,1024,0.010085333387056987
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,8192,768,0.02762773235638936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,8192,768,0.00698773314555486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,768,0.0060149331887563075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,768,0.00862613320350647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,8192,512,0.025513599316279095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,8192,512,0.006412800153096516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,512,0.005094400048255921
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,512,0.007732266684373219
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,8192,256,0.02360853354136149
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,256,0.004283733169237773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,256,0.007098666826883952
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,8192,128,0.02291733423868815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,8192,128,0.005295999844868978
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,128,0.0037258667250474296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,128,0.006705066561698914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,8192,64,0.022474666436513267
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,64,0.0035413332283496858
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,64,0.007451733450094859
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,8192,32,0.022833067178726196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,32,0.003622400015592575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,32,0.0074432000517845156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,65536,0.32922560373942056
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,65536,0.35320428212483723
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,7168,65536,0.28323094050089515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,7168,16384,0.09119359652201334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,16384,0.10677440166473388
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,16384,0.07335893313090006
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,7168,12288,0.04082026481628418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,7168,12288,0.0733354647954305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,12288,0.06390186548233032
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,12288,0.0640394647916158
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,7168,10240,0.06525866587956747
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,7168,10240,0.03593279918034871
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,10240,0.05430933237075806
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,10240,0.052621865272521974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,7168,8192,0.05723093350728353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,7168,8192,0.02943466703097026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,8192,12288,0.04341546694437663
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,8192,0.04381866852442424
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,8192,0.04387413263320923
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,7168,7168,0.025947733720143633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,7168,7168,0.07224319775899252
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,7168,0.03991146485010783
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,7168,0.040047999223073324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,7168,6144,0.06595413287480673
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,7168,6144,0.022954666614532472
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,6144,0.034647464752197266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,6144,0.035442133744557694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,7168,5120,0.058296533425649014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,7168,5120,0.021028266350428263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,5120,0.02934933304786682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,5120,0.029167999823888142
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,7168,4096,0.048478933175404866
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,7168,4096,0.01758613387743632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,4096,0.02472533384958903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,4096,0.022792534033457438
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,7168,3584,0.04338879982630412
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,7168,3584,0.015618133544921874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,3584,0.02217280069986979
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,3584,0.021819732586542764
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,7168,3072,0.04098453521728516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,7168,3072,0.014135467012723288
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,3072,0.018901334206263224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,3072,0.019377066691716512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,7168,2560,0.03703146775563558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,7168,2560,0.012410666545232136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,2560,0.010637866457303365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,2560,0.01780160069465637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,7168,2048,0.033879466851552326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,7168,2048,0.01097920040289561
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,2048,0.008712533116340637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,2048,0.014259200294812521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,7168,1536,0.03139413396517436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,7168,1536,0.008945066730181377
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,1536,0.007393066585063934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,1536,0.012388267119725545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,7168,1024,0.02940586606661479
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,8192,4096,0.017882666985193887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,7168,1024,0.007654400169849395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,1024,0.00603413333495458
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,1024,0.010025599598884582
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,7168,768,0.026791467269261675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,7168,768,0.0069354668259620665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,768,0.005364266534646353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,768,0.008637866377830506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,7168,512,0.025911466280619307
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,7168,512,0.006397866706053417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,512,0.004583466549714406
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,512,0.008026666442553202
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,7168,256,0.023653332392374674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,7168,256,0.005499733487764994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,256,0.003945599993069967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,256,0.00701333334048589
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,7168,128,0.022797866662343343
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,7168,128,0.0052928000688552855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,128,0.0036085332433382668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,128,0.006623999774456024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,7168,64,0.023106133937835692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,64,0.003588266670703888
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,8192,1024,0.007457066575686138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,64,0.0073525334397951765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,7168,32,0.02265920042991638
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,32,0.003689600030581156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,32,0.007426133255163829
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,6144,16384,0.08551999727884928
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,65536,0.29654080073038735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,65536,0.3298442522684733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,6144,65536,0.2599551995595296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,16384,0.08400853474934897
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,16384,0.07942612965901694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,8192,256,0.00565226674079895
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,6144,12288,0.04063146511713664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,6144,12288,0.0731391986211141
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,12288,0.05600853363672892
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,12288,0.057564798990885416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,6144,10240,0.061735467116038
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,6144,10240,0.03509653409322103
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,10240,0.04728960196177165
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,10240,0.04946879943211873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,6144,8192,0.05461013317108154
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,6144,8192,0.02887253363927205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,8192,0.03915946483612061
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,6144,7168,0.07044053077697754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,8192,0.041842134793599446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,6144,7168,0.024762666225433348
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,7168,0.038194131851196286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,7168,0.03601813316345215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,6144,6144,0.06308159828186036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,6144,0.03039146661758423
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,6144,0.03255253235499064
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,6144,5120,0.05513493220011393
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,7168,65536,0.227837864557902
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,6144,5120,0.01735573410987854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,5120,0.02561066746711731
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,5120,0.025476266940434773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,7168,16384,0.05243200063705444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,6144,4096,0.014099199573198953
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,6144,4096,0.04592853387196859
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,4096,0.021309866507848104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,4096,0.022177066405614218
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,3584,0.019266132513682047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,6144,3584,0.01309440036614736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,6144,3584,0.04251413345336914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,3584,0.019672532876332603
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,6144,3072,0.040133333206176756
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,3072,0.01154026687145233
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,3072,0.018572799364725747
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,2560,0.0099263995885849
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,6144,2560,0.036228267351786296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,2560,0.015962666273117064
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,6144,2048,0.033650132020314534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,6144,2048,0.00928106705347697
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,2048,0.008528000116348267
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,2048,0.01362986663977305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,6144,1536,0.008197333415349324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,6144,1536,0.030668799082438154
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,1536,0.007110400001207988
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,1536,0.01118293305238088
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,6144,1024,0.00714026689529419
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,6144,1024,0.028592000404993694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,1024,0.005904000004132589
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,1024,0.009522133072217305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,6144,768,0.026587732632954914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,6144,768,0.006470400094985962
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,768,0.00528106689453125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,768,0.008335999647776286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,6144,512,0.02513386607170105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,6144,512,0.005932799975077311
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,512,0.004527999957402547
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,512,0.007710933188597361
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,6144,256,0.02323306600252787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,6144,256,0.005339733262856802
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,256,0.0038880000511805216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,256,0.00699839989344279
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,6144,128,0.022871466477711995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,128,0.003572266548871994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,128,0.006621866424878438
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,6144,64,0.02230613430341085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,64,0.0033471999069054925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,64,0.007146666447321574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,6144,32,0.022514132658640544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,32,0.0034175999462604523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,32,0.007354666789372762
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,5120,65536,0.24583892822265624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,5120,65536,0.2056704044342041
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,65536,0.24523092905680338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,5120,16384,0.08052799701690674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,65536,0.2851306597391764
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,5120,16384,0.049396268526713055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,16384,0.07050027052561442
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,16384,0.06698880195617676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,5120,12288,0.06573653221130371
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,5120,12288,0.03842879931131999
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,12288,0.05565333366394043
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,12288,0.05408213138580322
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,5120,10240,0.05872426827748617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,10240,0.040660266081492105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,10240,0.048654933770497635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,6144,65536,0.18705280621846515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,6144,16384,0.05094186862309774
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,5120,8192,0.0516319990158081
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,5120,8192,0.026599466800689697
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,8192,0.033191466331481935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,8192,0.039502934614817305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,5120,7168,0.024551467100779215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,5120,7168,0.06754453182220459
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,7168,0.029418667157491047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,7168,0.03322880069414775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,5120,6144,0.059741866588592527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,5120,6144,0.020171733697255452
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,6144,0.026532266537348432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,6144,0.028973867495854694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,5120,5120,0.05077546834945679
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,5120,0.022284799814224245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,5120,0.026187733809153242
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,5120,4096,0.04401599963506063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,5120,4096,0.014877866705258688
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,4096,0.017413334051767985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,4096,0.02033066749572754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,5120,3584,0.041317331790924075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,5120,3584,0.013312000036239623
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,6144,6144,0.02077546715736389
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,3584,0.01064639985561371
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,3584,0.017552000284194947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,5120,3072,0.012551466623942057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,5120,3072,0.03932906786600749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,3072,0.010069333513577779
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,3072,0.016768000523249307
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,5120,2560,0.035844266414642334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,5120,2560,0.01062506635983785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,2560,0.008322133123874665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,2560,0.015280000368754067
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,5120,2048,0.03285013238588969
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,5120,2048,0.00913813312848409
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,2048,0.007228800157705943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,2048,0.012108799815177918
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,5120,1536,0.03056959907213847
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,1536,0.006305066744486491
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,5120,1536,0.007768533130486806
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,1536,0.010678399602572124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,5120,1024,0.028553599119186403
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,6144,3072,0.011811199784278869
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,1024,0.005335466563701629
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,1024,0.00885653297106425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,6144,2560,0.011125333110491435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,5120,768,0.02667413353919983
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,5120,768,0.00636053333679835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,768,0.004729599754015604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,768,0.008249600231647492
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,5120,512,0.025197867552439374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,5120,512,0.005996799965699514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,512,0.00420906643072764
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,512,0.007595733304818471
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,256,0.0037600000699361167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,5120,256,0.023579732577006022
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,5120,256,0.005306666592756907
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,256,0.007018666466077168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,5120,128,0.0233024001121521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,128,0.003483733286460241
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,128,0.006723199784755707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,5120,64,0.023281067609786987
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,64,0.0033312000334262846
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,64,0.0071285332242647815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,5120,32,0.02253226637840271
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,32,0.0033429334561030067
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,6144,128,0.005122133096059163
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,32,0.00713919997215271
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,4096,65536,0.16669440269470215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,4096,65536,0.23229653040568032
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,65536,0.20012799898783365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,65536,0.17045226097106933
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,4096,16384,0.07520853678385417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,4096,16384,0.04413866599400838
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,16384,0.054491734504699706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,16384,0.0453002651532491
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,4096,12288,0.060296531518300375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,4096,12288,0.0344160000483195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,12288,0.04378453493118286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,12288,0.03617386817932129
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,4096,10240,0.0547264019648234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,4096,10240,0.028750934203465778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,10240,0.03755199909210205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,10240,0.030989867448806763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,4096,8192,0.04912960131963094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,8192,0.03189226587613424
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,8192,0.025036799907684325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,4096,7168,0.06469653447469076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,4096,7168,0.020567466815312706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,7168,0.02231360077857971
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,7168,0.028955733776092528
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,4096,6144,0.05698133309682211
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,4096,6144,0.017772799730300902
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,6144,0.026191999514897663
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,6144,0.019704532623291016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,4096,5120,0.050037332375844325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,4096,5120,0.015745066603024802
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,5120,0.01542080044746399
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,5120,0.01732053359349569
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,4096,4096,0.043518932660420735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,4096,4096,0.013091199596722922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,4096,0.01055573324362437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,4096,0.014499200383822122
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,4096,3584,0.04121919870376587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,4096,3584,0.01213653286298116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,3584,0.009378133217493693
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,3584,0.013412266969680786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,4096,3072,0.011366400122642516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,4096,3072,0.03853653271993001
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,3072,0.00832533339659373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,3072,0.012604799866676331
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,5120,10240,0.031388799349466964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,4096,2560,0.034680533409118655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,4096,2560,0.009484799702962239
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,2560,0.007578666508197785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,2560,0.011027200023333232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,4096,2048,0.0320362667242686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,2048,0.009786666433016459
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,4096,2048,0.008237866560618083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,2048,0.0066549330949783325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,4096,1536,0.029734400908152263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,1536,0.005621333420276642
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,1536,0.008734933535257975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,4096,1024,0.027937066555023194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,1024,0.004906666775544485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,1024,0.007696000238259633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,5120,5120,0.017757866779963175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,4096,768,0.026753065983454387
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,4096,768,0.006179200112819671
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,768,0.004407466451327006
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,768,0.007208533088366191
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,4096,512,0.024711465835571288
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,512,0.004070399949947992
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,512,0.006733866532643636
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,4096,256,0.023602133989334105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,4096,256,0.005169066786766052
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,256,0.0036320000886917113
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,256,0.006306133170922597
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,4096,128,0.023075199127197264
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,4096,128,0.004915200173854828
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,128,0.0033610666791598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,128,0.006043733159701029
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,4096,64,0.0231221338113149
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,64,0.003189333279927572
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,64,0.0064064001043637585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,4096,32,0.022350933154424033
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,32,0.003382399926582972
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,32,0.006517333288987477
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,5120,1024,0.006915199756622315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3584,65536,0.15564053853352863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3584,65536,0.2125845273335775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,65536,0.15579946835835773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,65536,0.18093973795572918
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3584,16384,0.07256639798482259
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3584,16384,0.04281599919001262
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,16384,0.051310932636260985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,16384,0.04231893221537272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3584,12288,0.05875626802444458
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,12288,0.03971840143203735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,12288,0.03439679940541585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3584,10240,0.05350399812062582
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3584,10240,0.027188267310460406
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,10240,0.0346560001373291
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,10240,0.028361600637435914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3584,8192,0.047516798973083495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,5120,128,0.005086933573087057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3584,8192,0.021835732460021972
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,8192,0.02884053389231364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,8192,0.022900267442067464
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3584,7168,0.01963520050048828
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3584,7168,0.06328746477762857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,7168,0.022451200087865195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,7168,0.02132373253504435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3584,6144,0.05607253313064575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,6144,0.01768746574719747
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,6144,0.01857066750526428
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3584,5120,0.049531733989715575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3584,5120,0.014999467134475707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,5120,0.011705600221951802
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,5120,0.016289066274960837
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3584,4096,0.04306986729303996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3584,4096,0.012801067034403483
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,4096,0.0093450665473938
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,4096,0.014019200205802917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3584,3584,0.04102506637573242
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3584,3584,0.011918933192888895
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,3584,0.008738133311271667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,4096,8192,0.02337706685066223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,3584,0.013051733374595642
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3584,3072,0.03886186679204305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,3072,0.008257066706816356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,3072,0.011805867155392964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3584,2560,0.03446293274561564
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3584,2560,0.00927786628405253
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,2560,0.007189333438873291
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,2560,0.010571733117103577
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3584,2048,0.03178346753120422
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3584,2048,0.00804373323917389
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,2048,0.006332799792289734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,2048,0.009409067034721375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3584,1536,0.029600000381469725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3584,1536,0.007036800185839335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,1536,0.005662933488686879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,1536,0.008331733445326488
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3584,1024,0.027777065833409626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3584,1024,0.00624533345301946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,1024,0.004902400076389313
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,1024,0.007525333265463512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3584,768,0.026054400205612182
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3584,768,0.0059349333246548975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,768,0.00436160018046697
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,768,0.007052800059318543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3584,512,0.024606933196385704
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3584,512,0.005649066468079885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,512,0.003987200061480204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,512,0.0067071999112765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3584,256,0.023179733753204347
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3584,256,0.005227733155091604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,256,0.0035477332770824433
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,256,0.006248533229033152
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3584,128,0.022616533438364665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3584,128,0.004939733445644379
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,128,0.0033173332611719764
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,128,0.006020266811052958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3584,64,0.022056533892949422
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,64,0.003160533308982849
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,64,0.006353066861629486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,4096,1536,0.007338666419188182
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3584,32,0.023245867093404135
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,32,0.0032085334261258446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,32,0.006340266764163971
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,4096,1024,0.0063391998410224915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3072,65536,0.1499519983927409
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,65536,0.1523157278696696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,65536,0.1436949412027995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3072,65536,0.20888214111328124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3072,16384,0.06985173225402833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,16384,0.04629546801249186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,16384,0.03748906850814819
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,4096,512,0.005729066828886667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3072,12288,0.057240533828735354
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3072,12288,0.03096853295962016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,12288,0.03423466682434082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3072,10240,0.051590398947397864
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,12288,0.03141226569811503
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3072,10240,0.024875734249750772
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,10240,0.029639466603597002
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,10240,0.026555732885996504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3072,8192,0.04632746775945028
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,8192,0.02533973256746928
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,8192,0.022292266289393105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3072,7168,0.06185919841130575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3072,7168,0.018969599405924478
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,7168,0.019665066401163736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,7168,0.022524799903233847
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3072,6144,0.01670080025990804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3072,6144,0.054626135031382236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,6144,0.016196266810099284
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,6144,0.017802667617797852
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3072,5120,0.04852480093638102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3072,5120,0.014454399545987448
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,5120,0.0131221334139506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,5120,0.015876266360282897
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3072,4096,0.04185280005137126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3072,4096,0.012157866358757019
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,4096,0.0113045334815979
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,4096,0.013255467017491659
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3072,3584,0.039443198839823404
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3072,3584,0.011291733384132386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,3584,0.008669867118199667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,3584,0.012483200430870056
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3072,3072,0.03779413302739461
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3072,3072,0.010440533359845478
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3584,12288,0.032501333951950075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,3072,0.008005333443482716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,3072,0.011585066715876263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3072,2560,0.034202667077382405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,2560,0.00709440012772878
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,2560,0.01030399998029073
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3072,2048,0.03191039959589641
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3072,2048,0.008340266346931458
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,2048,0.006313600142796834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,2048,0.009195733070373534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3072,1536,0.029471999406814574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3072,1536,0.0070826664566993715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,1536,0.0056320001681645715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,1536,0.008187733093897502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3072,1024,0.02759786645571391
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,1024,0.004799999793370565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,1024,0.0074976002176602675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3072,768,0.026040534178415935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3072,768,0.005963733295599619
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,768,0.004374399781227112
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3584,6144,0.017735467354456583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,768,0.00703359991312027
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3072,512,0.024675200382868447
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3072,512,0.005598933498064677
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,512,0.003999999910593033
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,512,0.00666133314371109
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3072,256,0.023137066761652628
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3072,256,0.0050453335046768185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,256,0.0035264000296592714
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,256,0.006302933394908905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3072,128,0.02262079914410909
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3072,128,0.004884266853332519
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,128,0.003310933212439219
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,128,0.0059456000725428265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3072,64,0.02262293299039205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,64,0.0031871999303499854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,64,0.0062837332487106325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,3072,32,0.023081600666046143
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,32,0.003239466746648153
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,32,0.0063509335120519
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3584,3072,0.010935466488202412
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2560,65536,0.13787946701049805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2560,65536,0.19403732617696126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,65536,0.13092052936553955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2560,16384,0.06715733210245768
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,65536,0.12882986863454182
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2560,16384,0.036645332972208663
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,16384,0.03700799942016601
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,16384,0.03704959948857625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2560,12288,0.02770986755688985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2560,12288,0.05479146639506022
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,12288,0.030051199595133464
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,12288,0.028331732749938963
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2560,10240,0.048536535104115805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2560,10240,0.024218666553497314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,10240,0.026025599241256712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,10240,0.023720532655715942
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2560,8192,0.04326080083847046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,8192,0.02178986668586731
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,8192,0.020693333943684895
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2560,7168,0.05956586599349976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,7168,0.014485333363215128
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,7168,0.018212266763051353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2560,6144,0.01567146678765615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2560,6144,0.05228480100631714
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,6144,0.012452266613642375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,6144,0.016764799753824867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2560,5120,0.04691733519236247
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2560,5120,0.014196266730626425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,5120,0.011220266421635944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,5120,0.01487573285897573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2560,4096,0.041571199893951416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2560,4096,0.012101333340009053
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,4096,0.009705600142478944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,4096,0.012619733810424805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2560,3584,0.03935360113779704
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,3584,0.009004799524943034
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,3584,0.01130560040473938
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2560,3072,0.03731093406677246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2560,3072,0.00965333382288615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,3072,0.008225066463152568
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,3072,0.01067733367284139
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3072,16384,0.04057066837946574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2560,2560,0.0339957316716512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2560,2560,0.009243733684221904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,2560,0.0070720002055168155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,2560,0.009963732957839967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2560,2048,0.03181866606076558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2560,2048,0.007560533285140991
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,2048,0.006259199976921081
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,2048,0.008585600058237712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3072,8192,0.02091946601867676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2560,1536,0.029883732398351032
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,1536,0.005551999807357788
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,1536,0.007855999966462452
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2560,1024,0.027900799115498857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2560,1024,0.0060576001803080235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,1024,0.004682666560014089
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,1024,0.007157333195209503
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2560,768,0.026173865795135497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2560,768,0.0059125334024429325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,768,0.004321066538492838
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,768,0.006772266825040181
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2560,512,0.024774400393168132
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2560,512,0.005500799914201101
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,512,0.003917866696914038
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,512,0.006495999793211619
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2560,256,0.023364265759785972
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3072,2560,0.009293867150942485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,256,0.0035349334279696143
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,256,0.006138666470845541
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2560,128,0.0224671999613444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,128,0.00325546662012736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,128,0.0058794667323430385
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2560,64,0.022361600399017335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3072,1024,0.006168533364931742
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,64,0.0031338666876157125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,64,0.006202666461467743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2560,32,0.02211093306541443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,32,0.003206400076548258
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,32,0.0061749334136645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2048,65536,0.19358080228169758
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,65536,0.10697279771169026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,65536,0.10349866549173992
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2048,16384,0.06453866561253865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2560,8192,0.020431999365488687
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,16384,0.031089067459106445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2560,7168,0.018263467152913413
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,16384,0.03170666694641113
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2048,12288,0.053058131535847985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2048,12288,0.026707200209299724
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,12288,0.02523946762084961
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,12288,0.022855466604232787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2048,10240,0.04775253136952718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2048,10240,0.02241706649462382
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,10240,0.02186453342437744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,10240,0.019332265853881835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2048,8192,0.04252266486485799
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,8192,0.013026133179664612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,8192,0.018473599354426065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2048,7168,0.057922132809956874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2048,7168,0.017172267039616905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2560,3584,0.010722133517265319
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,7168,0.012166399757067363
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,7168,0.01574613352616628
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2048,6144,0.05224106709162394
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,6144,0.01125759979089101
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,6144,0.014906666676203408
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2048,5120,0.04707306623458862
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2048,5120,0.013250133395195008
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,5120,0.010307199756304423
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,5120,0.013260799646377563
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2048,4096,0.041458133856455484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2560,1536,0.006862933437029521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,4096,0.011575466394424439
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,4096,0.010901332894961039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2048,3584,0.038755198319753006
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2560,256,0.005132799843947092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,3584,0.010500267148017883
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,3584,0.010272000233332317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2560,128,0.005026133358478546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2048,3072,0.036873598893483475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2048,3072,0.009654399752616883
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,3072,0.009311999877293904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,3072,0.009677867094675701
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2048,2560,0.03376106818517049
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,2560,0.006730666756629944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,2560,0.008905599514643352
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2048,2048,0.03165439963340759
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2048,2048,0.007309866448243459
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,2048,0.005938133100668589
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,2048,0.008001066744327545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2048,1536,0.029666133721669513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2048,65536,0.12101226647694904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2048,1536,0.006728533407052357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,1536,0.005203199883302053
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,1536,0.007536000013351441
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2048,1024,0.02728533347447713
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2048,1024,0.006026666859785716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,1024,0.00448639988899231
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,1024,0.006960000097751618
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2048,768,0.025655466318130492
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2048,768,0.005776000022888183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,768,0.004161066561937332
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,768,0.006681600213050842
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2048,16384,0.03370026747385661
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2048,512,0.024551467100779215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2048,512,0.0055061335364977515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,512,0.003866666555404663
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,512,0.006417066852251689
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2048,256,0.022667733828226726
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2048,256,0.005051733553409576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,256,0.00347626656293869
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,256,0.006121600170930227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2048,128,0.022333866357803343
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2048,128,0.004775466521581014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,128,0.0032000000278155005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,128,0.005910400052865346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2048,64,0.02167466680208842
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,64,0.0030794667700926462
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,64,0.006044800082842508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,2048,32,0.022874667247136434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,32,0.003101866692304611
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,32,0.006102400024731954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1536,65536,0.09945706526438394
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1536,65536,0.1793941338857015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,65536,0.08340373039245605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,65536,0.08577280044555664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1536,16384,0.05921386480331421
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,16384,0.025754666328430174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,16384,0.021993599335352578
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1536,12288,0.04789546728134155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1536,12288,0.022232532501220703
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,12288,0.016687999169031777
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,12288,0.01791999936103821
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1536,10240,0.042792534828186034
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,10240,0.012572800119717916
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,10240,0.015889066457748412
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2048,8192,0.01908373236656189
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1536,8192,0.036659200986226395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1536,8192,0.01730239987373352
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,8192,0.010966400305430096
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,8192,0.0151829332113266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1536,7168,0.015711999932924905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1536,7168,0.05452479918797811
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,7168,0.010437333583831787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,7168,0.01299626628557841
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1536,6144,0.050222933292388916
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1536,6144,0.013651200135548911
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,6144,0.009732266267140705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2048,6144,0.015356799960136414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,6144,0.012001066406567892
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1536,5120,0.04510720173517863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,5120,0.009231999516487122
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,5120,0.012044800321261089
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1536,4096,0.03999360005060832
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1536,4096,0.010368000467618306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,4096,0.008293333152929943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,4096,0.010653866330782573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1536,3584,0.03811200062433879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1536,3584,0.009732266267140705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,3584,0.008105599880218506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,3584,0.009716266393661499
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2048,4096,0.01116373340288798
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1536,3072,0.035947732130686444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1536,3072,0.008986666798591614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,3072,0.00935040016969045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,3072,0.008824533224105835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1536,2560,0.033538134892781575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1536,2560,0.008055466910203297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,2560,0.00860693355401357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,2560,0.006692266464233399
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1536,2048,0.031048534313837688
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1536,2048,0.007076266904671986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,2048,0.005926399926344554
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,2048,0.00819413314263026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1536,1536,0.029028266668319702
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,1536,0.005178666611512502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2048,3584,0.010390399893124899
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,1536,0.007482666770617168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1536,1024,0.027165865898132323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1536,1024,0.005797333518664042
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,1024,0.004497066636880239
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,1024,0.00695253312587738
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1536,768,0.0255840003490448
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1536,768,0.0056320001681645715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,768,0.004304000238577525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,768,0.0066453332702318835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1536,512,0.02427840034166972
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1536,512,0.0053951998551686605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,512,0.0037930667400360107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,512,0.006426666676998138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1536,256,0.023036799828211465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1536,256,0.004887466629346212
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,256,0.0034485332667827605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,256,0.006107733150323232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1536,128,0.022392533222834268
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,128,0.003257599969704946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1536,128,0.0047872001926104225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,128,0.005925333499908448
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2048,2560,0.008433066805203756
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,64,0.006087466577688853
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1536,64,0.02173866629600525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,64,0.0030773334205150605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1536,32,0.02177706758181254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,32,0.0031178665657838186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,32,0.006027733286221823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1024,65536,0.16925759315490724
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,65536,0.0645525336265564
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,65536,0.08170346419016519
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1024,16384,0.057557332515716556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1024,16384,0.031259733438491824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,16384,0.012757333119710288
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,16384,0.024218666553497314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1024,12288,0.047622398535410566
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1024,12288,0.02369920015335083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,12288,0.010680533448855082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,12288,0.01923519968986511
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1024,10240,0.0404853343963623
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1024,10240,0.02025493383407593
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,10240,0.010069333513577779
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,10240,0.017526400089263917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1024,8192,0.036091732978820804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1024,8192,0.017282134294509886
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,8192,0.008875733613967896
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,8192,0.014999467134475707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1024,7168,0.015475199619928996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1024,7168,0.053415465354919436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,7168,0.00865280032157898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,7168,0.013598933815956116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1024,6144,0.04883840084075928
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1024,6144,0.01362879971663157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,6144,0.011648000280062357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,6144,0.01230506698290507
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1024,5120,0.044241066773732504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1024,5120,0.012302933136622111
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,5120,0.01015786627928416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,5120,0.011479467153549194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1024,4096,0.039537068208058676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1024,4096,0.01035520037015279
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,4096,0.008847999572753906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,4096,0.010055466492970785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1024,3584,0.03767253160476684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,3584,0.008045866588751475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,3584,0.009397332866986591
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1024,3072,0.03569706678390503
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1024,3072,0.008887466788291932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,3072,0.007416533430417378
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,3072,0.00897173285484314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1024,2560,0.033308800061543783
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1024,2560,0.007708799839019775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1536,16384,0.030577067534128827
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,2560,0.0066549330949783325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,2560,0.008477866649627686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1024,2048,0.030746666590372722
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1024,2048,0.007131733496983846
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,2048,0.0060917332768440245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,2048,0.00780266672372818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1024,1536,0.02895359992980957
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1024,1536,0.006469333171844482
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,1536,0.005191466708978017
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,1536,0.007326933244864146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1024,1024,0.026978133122126262
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1024,1024,0.005736533304055532
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,1024,0.0044938668608665465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,1024,0.006791466474533081
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1024,768,0.02551146745681763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1024,768,0.005723733206590017
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,768,0.004136533290147781
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,768,0.006506666541099548
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1024,512,0.024203733603159586
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1024,512,0.0053610667586326596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,512,0.0037759999434153237
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1536,10240,0.019657599925994872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,512,0.006296533346176148
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1024,256,0.022887466351191203
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,256,0.0033642667035261786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,256,0.006065066655476888
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1024,128,0.02250880002975464
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,128,0.003123199939727783
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,128,0.005838933090368906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1024,64,0.022089600563049316
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,64,0.0029535998900731405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,64,0.005883733431498209
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,1024,32,0.021986132860183714
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,32,0.0030410667260487873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,32,0.005922133227189382
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,768,65536,0.09162240028381348
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,768,65536,0.16117439270019532
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,65536,0.054797867933909096
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,65536,0.07367253303527832
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1536,5120,0.012409599622090657
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,768,16384,0.05533653497695923
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,768,16384,0.02632533311843872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,16384,0.01123306651910146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,16384,0.02218666672706604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,768,12288,0.04460266828536987
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,12288,0.009958400328954061
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,12288,0.018040533860524496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,768,10240,0.03902933200200399
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,768,10240,0.01735146641731262
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,10240,0.013317333658536276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,10240,0.016396799683570863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,768,8192,0.03538346687952677
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,768,8192,0.015553067127863566
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,8192,0.011715199549992878
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,8192,0.013960533340771995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,768,7168,0.05296106735865275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,768,7168,0.014146133263905843
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,7168,0.011142399907112122
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,7168,0.012786133090655008
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,768,6144,0.04845973253250122
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,768,6144,0.012873599926630655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,6144,0.010116266210873921
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1536,1536,0.006387199958165486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,6144,0.011817600329717
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,768,5120,0.011431466539700825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,768,5120,0.04465173482894898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,5120,0.009165866176287334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,5120,0.010539733370145162
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,768,4096,0.039241600036621097
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,768,4096,0.00909866690635681
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,4096,0.008293333152929943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,4096,0.009429333607355754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,768,3584,0.03724586566289266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,768,3584,0.008691199620564779
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,3584,0.007991466422875721
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,3584,0.009058133761088053
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,768,3072,0.03543359835942586
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,768,3072,0.008202666540940602
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,3072,0.00755626658598582
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,3072,0.00865280032157898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,768,2560,0.03317226568857829
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,768,2560,0.007080533107121785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,2560,0.006431999802589417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,2560,0.008229333162307739
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,768,2048,0.030746666590372722
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,768,2048,0.006558933357397716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,2048,0.005713066458702088
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,2048,0.007514666517575581
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,768,1536,0.02895466685295105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,768,1536,0.0062047998110453285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,1536,0.005052799979845682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,1536,0.0071263998746871945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,768,1024,0.02672426700592041
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,768,1024,0.005614933371543884
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,1024,0.0043722664316495265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,1024,0.006678399940331777
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,768,768,0.025646932919820148
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,768,768,0.005471999943256378
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,768,0.003973333289225897
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,768,0.006478933493296306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1024,65536,0.10160640080769856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,768,512,0.024210133155186973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,768,512,0.005287466446558634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,512,0.0036821333070596062
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,512,0.006202666461467743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,768,256,0.02290133237838745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,768,256,0.004867200056711833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,256,0.0033791999022165934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,256,0.0060138667623202005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,768,128,0.02246613303820292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,768,128,0.004681600133577982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,128,0.003138133386770884
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,128,0.0058378666639328
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,768,64,0.0216757337252299
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,64,0.002998399982849757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,64,0.005892266829808554
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,768,32,0.021806933482487998
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,32,0.003019733230272929
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,32,0.005884799857934316
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,512,65536,0.07302400271097818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,512,65536,0.15401387214660645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,65536,0.04068479935328166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,65536,0.06912853717803955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,512,16384,0.054484268029530845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,512,16384,0.020304000377655028
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,16384,0.010905599594116211
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,16384,0.02067199945449829
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,512,12288,0.04319146474202474
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,512,12288,0.016570666432380678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,12288,0.009664000074068705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,12288,0.017013333241144814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,512,10240,0.014429866274197897
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,512,10240,0.037912531693776445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,10240,0.010622933506965637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,10240,0.015566933155059814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1024,3584,0.009698133667310078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,512,8192,0.03494506676991781
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,8192,0.009343999624252319
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,8192,0.01322773297627767
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,512,7168,0.05213866631189982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,512,7168,0.011502933502197266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,7168,0.008973866701126099
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,7168,0.012357333302497863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,512,6144,0.047940266132354734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,512,6144,0.01048533320426941
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,6144,0.009352533022562663
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,6144,0.011507200201352437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,512,5120,0.04355839888254802
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,512,5120,0.00976639986038208
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,5120,0.008611200253168742
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,5120,0.010573866963386535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,4096,0.0079434668024381
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,512,4096,0.03906879822413127
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,512,4096,0.008597333232561748
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,4096,0.009388800462086995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,512,3584,0.037196799119313555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,512,3584,0.008067200084527333
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,3584,0.00906773308912913
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,3584,0.007730133334795634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,512,3072,0.007479466497898102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1024,256,0.00492799977461497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,512,3072,0.03518720070521037
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,3072,0.007359999914964039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,3072,0.008646399776140849
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,512,2560,0.03326506614685058
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,512,2560,0.006981333096822103
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,2560,0.006322133541107178
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,2560,0.008186666667461396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,512,2048,0.0306986669699351
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1024,128,0.004693333307902018
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,2048,0.005659733215967814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,2048,0.007659733295440674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,512,1536,0.02890239953994751
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,512,1536,0.006051200131575266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,1536,0.0049685334165891016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,1536,0.007142400244871776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,512,1024,0.026818132400512694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,512,1024,0.005490133166313171
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,1024,0.004306133091449738
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,1024,0.006718933085600535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,512,768,0.02545386751492818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,512,768,0.005353599786758423
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,768,0.003973333289225897
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,768,0.006533333162466686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,512,512,0.02406186660130819
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,512,0.0036320000886917113
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,512,0.006169599791367849
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,512,256,0.022882133722305298
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,256,0.0033098667860031127
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,256,0.005948799848556519
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,512,128,0.02244266668955485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,512,128,0.004553600152333578
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,128,0.0031008000175158186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,128,0.005776000022888183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,512,64,0.022286933660507203
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,64,0.002951466788848241
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,64,0.0058229332168896995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,512,32,0.022616533438364665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,32,0.0029781334102153777
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,32,0.005852800110975901
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,768,12288,0.020581332842508952
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,256,65536,0.07160320281982421
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,256,65536,0.152292267481486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,65536,0.026014933983484905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,65536,0.06523733139038086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,256,16384,0.05188266833623251
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,256,16384,0.020284799734751384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,16384,0.01043839951356252
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,16384,0.0199072003364563
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,256,12288,0.041124268372853594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,256,12288,0.016500266393025716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,12288,0.009046399593353271
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,12288,0.016425599654515587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,256,10240,0.03772906859715779
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,256,10240,0.014401066303253173
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,10240,0.008411733309427898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,10240,0.014786133170127868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,256,8192,0.0348309318224589
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,256,8192,0.012485333283742269
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,8192,0.007795199751853943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,256,7168,0.052425599098205565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,8192,0.012834133704503379
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,256,7168,0.011655466755231221
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,7168,0.0074080000321070354
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,7168,0.01200320025285085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,256,6144,0.04782933394114176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,256,6144,0.010426666339238484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,6144,0.007096533477306366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,6144,0.011281067132949829
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,256,5120,0.04367680152257283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,256,5120,0.009468799829483033
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,5120,0.0075103998184204105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,5120,0.010422399640083313
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,256,4096,0.03869333267211914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,256,4096,0.008502399921417237
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,4096,0.007062399884064992
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,4096,0.009326933821042379
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,256,3584,0.037444265683492024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,256,3584,0.008024533092975617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,3584,0.007382399837176005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,256,3072,0.035442133744557694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,3584,0.008921600381533305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,256,3072,0.007467733323574066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,3072,0.0070933332045873005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,3072,0.008583466211954754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,256,2560,0.03308586676915486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,256,2560,0.006920533378918965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,2560,0.006259199976921081
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,2560,0.008171733220418293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,256,2048,0.03073386748631795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,256,2048,0.006402133405208588
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,2048,0.005553066730499268
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,2048,0.007649066547552745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,256,1536,0.028778666257858278
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,256,1536,0.006041599810123444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,1536,0.004904533425966898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,1536,0.0070250665148099255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,256,1024,0.026866134007771807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,256,1024,0.005542399982611338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,1024,0.004266666869322458
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,1024,0.00661653329928716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,256,768,0.025285333395004272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,256,768,0.005314133564631144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,768,0.0038794666528701783
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,768,0.006434133152167003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,256,512,0.023965867360432942
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,256,512,0.005019733309745788
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,512,0.003570133447647095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,512,0.006190933287143707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,256,256,0.022797866662343343
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,256,0.0032416000962257386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,256,0.005874133110046387
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,256,128,0.022223999102910362
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,256,128,0.00452693353096644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,128,0.0030517332255840302
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,128,0.005713066458702088
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,256,64,0.021460266908009847
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,64,0.0029109333952267963
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,64,0.005799466868241628
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,256,32,0.021683200200398763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,32,0.0028959999481836954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,32,0.005787733197212219
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,128,65536,0.07029226620992025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,128,65536,0.15034987131754557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,65536,0.01376106639703115
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,65536,0.06190400123596192
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,128,16384,0.05043306748072306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,128,16384,0.020266666014989217
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,16384,0.007918933530648549
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,16384,0.01965013345082601
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,128,12288,0.04106560150782267
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,12288,0.007788800199826558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,512,8192,0.012461866935094197
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,12288,0.01604586640993754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,128,10240,0.03768746852874756
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,10240,0.007392000158627827
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,128,10240,0.014402133226394654
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,10240,0.014390400052070618
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,128,8192,0.034815998872121175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,128,8192,0.01250879963239034
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,8192,0.0072405333320299785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,8192,0.012558933099110922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,128,7168,0.05270080169041952
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,128,7168,0.011415466666221619
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,7168,0.007130666573842366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,7168,0.011733333269755047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,128,6144,0.04809173345565796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,128,6144,0.01037440001964569
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,6144,0.006898133456707001
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,6144,0.0108842670917511
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,128,5120,0.04379520018895467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,128,5120,0.009398399790128072
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,5120,0.00726506660381953
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,5120,0.010213333368301391
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,128,4096,0.03946773211161296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,4096,0.00689386675755183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,4096,0.009063466389973959
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,128,3584,0.037333333492279054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,128,3584,0.008010666569073994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,3584,0.007339733342329661
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,3584,0.008659199873606364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,128,3072,0.03528106609980265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,128,3072,0.007456000149250031
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,3072,0.007021866738796234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,3072,0.008293333152929943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,128,2560,0.033063467343648276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,128,2560,0.00688213308652242
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,2560,0.006223999957243601
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,2560,0.007867733140786488
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,128,2048,0.03038613398869832
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,128,2048,0.0064181332786877945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,2048,0.005584000051021576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,2048,0.007371733089288075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,512,2048,0.006517333288987477
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,128,1536,0.028676267464955645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,128,1536,0.006038400034109751
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,512,256,0.004752000172932943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,1536,0.0049216002225875854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,1536,0.0069461335738499955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,128,1024,0.02718293269475301
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,128,1024,0.005397333204746247
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,1024,0.004206933577855428
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,1024,0.006415999929110209
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,128,768,0.025668267409006757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,128,768,0.005178666611512502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,768,0.0038912000755469004
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,768,0.006278400123119354
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,512,512,0.0050687998533248905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,128,512,0.02454293370246887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,128,512,0.00492799977461497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,512,0.003575466573238373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,512,0.006009600063165029
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,128,256,0.02274880011876424
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,128,256,0.004753066599369049
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,256,0.0032287999987602235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,256,0.005810133119424184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,128,128,0.022193066279093423
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,128,0.0030410667260487873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,128,0.005748266478379568
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,128,64,0.02185279925664266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,64,0.002890666574239731
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,64,0.005738666653633118
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,128,32,0.02151040037473043
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,32,0.00296426663796107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,32,0.005719466507434845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,65536,0.011654399832089742
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,64,65536,0.15162986119588215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,65536,0.06180266539255778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,64,16384,0.049353599548339844
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,16384,0.007181866466999054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,16384,0.019615999857584634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,64,12288,0.04114346504211426
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,12288,0.007046400010585785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,12288,0.016080000003178916
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,64,10240,0.03784106572469075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,10240,0.007281066477298736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,10240,0.014371200402577718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,64,8192,0.03456533352533976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,8192,0.007136000196139018
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,8192,0.01251520017782847
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,7168,0.0069482664267222095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,64,7168,0.052716799577077235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,7168,0.011690666278203327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,64,6144,0.047446401913960774
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,6144,0.0068693334857622785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,6144,0.010856533050537109
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,64,5120,0.04330986738204956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,5120,0.007277866701285045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,5120,0.010076799988746643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,4096,0.006830933193365733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,64,4096,0.03903893232345581
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,4096,0.008993066350618998
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,64,3584,0.03773973385492961
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,3584,0.008730666836102803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,3584,0.007256533205509186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,64,3072,0.03579093217849731
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,3072,0.0069482664267222095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,3072,0.008293333152929943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,2560,0.006243200103441874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,64,2560,0.033336532115936277
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,2560,0.0078005333741505934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,2048,0.005599999924500784
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,64,2048,0.0302239994208018
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,2048,0.007325866818428039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,64,1536,0.0291701336701711
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,1536,0.004877866804599762
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,1536,0.006899199883143107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,1024,0.004187733431657155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,64,1024,0.02726400097211202
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,1024,0.006528000036875407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,64,768,0.02510400017102559
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,768,0.0038730666041374207
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,768,0.006287999947865804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,64,512,0.023908267418543495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,512,0.0035455999275048576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,512,0.006041599810123444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,64,256,0.02296533385912577
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,256,0.0032106667757034303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,256,0.005862399935722351
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,64,128,0.022009599208831786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,128,0.003044266750415166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,128,0.005729066828886667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,64,0.005851733187834421
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,64,64,0.02178986668586731
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,64,0.002926933268706004
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,64,32,0.0214464008808136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,32,0.002883200099070867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,32,0.005797333518664042
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,65536,0.011368532975514729
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,32,65536,0.151034673055013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,65536,0.06175359884897867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,32,16384,0.05047146479288737
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,16384,0.007117866476376851
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,16384,0.0194432000319163
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,32,12288,0.04128533204396566
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,12288,0.007086933155854543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,12288,0.015986133615175882
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,10240,0.014359466234842935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,32,10240,0.03750720024108887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,10240,0.007181866466999054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,256,256,0.004766933123270671
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,32,8192,0.03439573446909587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,8192,0.0070271998643875126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,8192,0.01237546702226003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,32,7168,0.05268266598383585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,7168,0.006867200136184692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,7168,0.011668266852696736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,32,6144,0.047637331485748294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,6144,0.006737066805362702
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,6144,0.010908800363540649
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,32,5120,0.043435732523600265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,5120,0.007130666573842366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,5120,0.010018133123715718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,32,4096,0.03889066775639852
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,4096,0.006818133095900218
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,4096,0.008969600001970928
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,32,3584,0.036925868193308516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,3584,0.007253333429495494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,32,3072,0.03488853375116984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,3584,0.008550399541854858
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,3072,0.006846933563550313
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,3072,0.008268799881140392
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,2560,0.007891199986139933
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,32,2560,0.03284800052642822
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,2560,0.006155733267466227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,2048,0.00747519979874293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,32,2048,0.03023146589597066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,2048,0.0055402666330337524
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,32,1536,0.02844160000483195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,1536,0.00487253318230311
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,1536,0.006900266806284587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,32,1024,0.026357332865397137
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,1024,0.00418453315893809
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,1024,0.006457599997520447
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,32,768,0.025044266382853193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,768,0.003835733234882355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,768,0.006243200103441874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,32,512,0.0239519993464152
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,512,0.0035264000296592714
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,512,0.006046933432420095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,32,256,0.02263039946556091
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,256,0.003218133250872294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,256,0.005899733304977417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,32,128,0.022016000747680665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,128,0.0029663999875386557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,128,0.0057536001006762184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,32,64,0.021771732966105142
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,64,0.002828799933195114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,64,0.0056970665852228795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,80,32,32,0.021465599536895752
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,32,0.0028394666810830434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,32,0.005654400090376536
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,128,12288,0.01644480029741923
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,65536,16384,0.39836479822794596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,65536,16384,0.3808970769246419
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,65536,16384,0.4262624104817708
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,65536,16384,0.8334378560384115
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,65536,12288,0.27851521174112953
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,65536,12288,0.2996618588765462
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,65536,12288,0.35075947443644206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,65536,12288,0.638745625813802
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,65536,10240,0.23283732732137047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,128,4096,0.00840106705824534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,65536,10240,0.24290453592936195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,65536,10240,0.27537387212117515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,65536,10240,0.49145600001017253
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,65536,8192,0.19136212666829427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,65536,8192,0.24470613797505697
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,65536,8192,0.1948352018992106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,65536,8192,0.3870880126953125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,65536,7168,0.17392533620198566
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,65536,7168,0.3443445205688477
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,65536,7168,0.22438613573710123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,65536,6144,0.14506667455037434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,65536,6144,0.19135360717773436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,65536,6144,0.1509386698404948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,65536,6144,0.29593706130981445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,65536,5120,0.12806719938913982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,65536,5120,0.1335466702779134
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,65536,5120,0.25052800178527834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,65536,5120,0.17197972933451336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,65536,4096,0.13621973991394043
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,65536,4096,0.21700480779012046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,65536,4096,0.1017568031946818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,65536,3584,0.08900159994761149
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,65536,3584,0.1252031962076823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,65536,3584,0.09017279942830404
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,65536,3584,0.17756160100301105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,65536,3072,0.07814506689707437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,65536,3072,0.11424319744110108
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,65536,3072,0.07884480158487955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,65536,3072,0.1531775951385498
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,65536,2560,0.06726613044738769
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,65536,2560,0.09946239789326985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,65536,2560,0.06716586748758951
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,65536,2560,0.12940906683603923
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,65536,2048,0.055396266778310145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,65536,2048,0.05556586583455404
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,65536,2048,0.0865226666132609
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,128,128,0.004485333462556204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,65536,2048,0.10312853654225666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,65536,1536,0.0735210657119751
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,65536,1536,0.0442901333173116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,65536,1536,0.0816490650177002
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,65536,1536,0.04415573279062907
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,65536,1024,0.06386133432388305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,65536,1024,0.05552213191986084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,65536,1024,0.033316266536712644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,65536,768,0.05830186605453491
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,65536,768,0.02613866726557414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,65536,512,0.05093439817428589
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,65536,768,0.04415359894434611
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,65536,768,0.028075732787450153
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,65536,512,0.019024000565210978
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,65536,512,0.031369600693384805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,65536,512,0.01918399930000305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,65536,256,0.036748798688252766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,65536,256,0.014535466829935709
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,65536,256,0.01246506671110789
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,65536,128,0.03581759929656982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,65536,128,0.008293333152929943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,65536,128,0.0072618668278058365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,65536,128,0.011623467008272808
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,65536,64,0.035920000076293944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,65536,64,0.00639573335647583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,65536,64,0.011890133221944172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,65536,32,0.03595626751581828
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,65536,32,0.006471466521422069
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,65536,32,0.01216426690419515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,16384,65536,0.3951189359029134
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,65536,0.8129301071166992
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,16384,65536,0.4763594627380371
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,65536,0.4327274640401204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,16384,16384,0.13415679931640626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,16384,16384,0.104258131980896
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,16384,0.11257920265197754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,16384,0.22079146703084312
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,16384,12288,0.11068267027537029
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,16384,12288,0.08023040294647217
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,12288,0.14923733075459797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,12288,0.08640320301055908
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,16384,10240,0.0640725334485372
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,16384,10240,0.10025386810302735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,10240,0.0730133295059204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,10240,0.1207477331161499
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,16384,8192,0.08113386631011962
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,8192,0.05813119808832804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,8192,0.11073386669158936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,16384,7168,0.09405226707458496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,16384,7168,0.04716693162918091
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,7168,0.05182826519012451
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,7168,0.09513920148213705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,65536,7168,0.1662453333536784
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,16384,6144,0.08427840073903402
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,16384,6144,0.04147626558939616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,6144,0.07954773108164469
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,6144,0.045466665426890054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,16384,5120,0.035547733306884766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,16384,5120,0.07493653297424316
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,5120,0.06965013345082602
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,5120,0.03892373243967692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,16384,4096,0.06575786670049032
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,16384,4096,0.02967146635055542
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,65536,4096,0.10120533307393391
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,4096,0.057068800926208495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,4096,0.03226026693979899
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,16384,3584,0.061026132106781004
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,3584,0.04581439892450968
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,3584,0.029288534323374433
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,16384,3072,0.056277334690093994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,16384,3072,0.024152533213297526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,3072,0.0396725336710612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,3072,0.02615999976793925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,16384,2560,0.052040533224741614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,16384,2560,0.021588265895843506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,2560,0.03627093235651652
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,2560,0.02358506719271342
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,16384,2048,0.010993066430091857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,16384,2048,0.04300053517023723
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,2048,0.02917226751645406
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,2048,0.013666133085886637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,16384,1536,0.03601599931716919
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,1536,0.02283733288447062
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,1536,0.011714133620262145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,65536,1024,0.032151466608047484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,16384,1024,0.03070080081621806
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,1024,0.0090037335952123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,1024,0.010024533669153849
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,16384,768,0.02877013285954793
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,16384,768,0.006984533369541168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,768,0.007354666789372762
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,768,0.009186133742332458
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,16384,512,0.02640639940897624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,16384,512,0.006405333181222279
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,512,0.006295466423034668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,512,0.008210133512814839
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,16384,256,0.02490133245786031
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,16384,256,0.0057429333527882894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,256,0.0046304002404212955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,256,0.007663999994595845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,16384,128,0.02480640014012655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,16384,128,0.005388799806435903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,128,0.004054400076468786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,128,0.007411199808120728
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,16384,64,0.0248416006565094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,64,0.0038090666135152185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,64,0.007499733567237854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,16384,32,0.024835199117660522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,32,0.007585066556930542
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,32,0.003899733225504557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,65536,256,0.010052266716957092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,12288,65536,0.28481601079305013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,65536,0.28967787424723307
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,65536,0.5810634613037109
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,12288,16384,0.08914986451466879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,12288,16384,0.11346133550008138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,12288,65536,0.39816532135009763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,16384,0.14076372782389324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,16384,0.07946879863739013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,12288,0.10749226411183674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,12288,12288,0.10042880376180012
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,12288,0.06240213314692179
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,12288,10240,0.08059199651082358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,12288,10240,0.05094826618830363
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,10240,0.053236265977223717
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,10240,0.09012479782104492
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,12288,8192,0.04272213379542033
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,12288,8192,0.07018880049387613
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,8192,0.07281706333160401
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,8192,0.04338239828745524
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,12288,7168,0.03744746843973796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,12288,7168,0.08335359891255698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,7168,0.06450986862182617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,7168,0.039051731427510575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,12288,6144,0.07585813204447428
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,12288,6144,0.03342080116271973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,6144,0.05655466715494791
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,6144,0.03441173235575358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,16384,8192,0.052697598934173584
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,12288,5120,0.06789440313975016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,12288,5120,0.028780800104141236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,5120,0.03004586696624756
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,5120,0.05496213436126709
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,12288,4096,0.024153600136439003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,12288,4096,0.05973333517710368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,4096,0.03869013388951619
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,4096,0.025011199712753295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,12288,3584,0.05588159958521525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,12288,3584,0.02176106572151184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,3584,0.022750933965047203
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,3584,0.039561601479848225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,12288,3072,0.05068480173746744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,12288,3072,0.015338666240374246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,3072,0.03025813301404317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,3072,0.014365866780281067
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,12288,2560,0.04293440183003743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,2560,0.025870933135350542
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,2560,0.012184533476829528
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,12288,2048,0.03750186761220296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,12288,2048,0.009316266576449076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,2048,0.021266132593154907
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,2048,0.010901332894961039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,12288,1536,0.03378560145696004
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,1536,0.009899733463923137
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,16384,3584,0.02701759934425354
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,1536,0.009618133306503296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,12288,1024,0.029667200644810994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,1024,0.00791786660750707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,1024,0.008251733581225077
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,12288,768,0.028435200452804565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,12288,768,0.006466133395830791
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,768,0.006669866542021434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,768,0.0077909335494041445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,12288,512,0.02680533329645793
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,12288,512,0.005832533538341522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,512,0.00551146666208903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,512,0.00728000005086263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,12288,256,0.024842667579650878
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,12288,256,0.005463466544946035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,16384,1536,0.00905386706193288
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,256,0.00432640016078949
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,256,0.00691840002934138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,12288,128,0.024213333924611412
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,12288,128,0.00514986664056778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,128,0.0037685332198937735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,128,0.006670933465162914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,16384,1024,0.007749333480993907
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,12288,64,0.02496853272120158
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,64,0.0035402665535608927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,64,0.006713599960009258
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,12288,32,0.024175999561945598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,32,0.003647999962170919
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,32,0.0067658667763074234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,10240,65536,0.23816533088684083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,65536,0.2504447937011719
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,65536,0.487337589263916
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,10240,65536,0.3583274523417155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,10240,16384,0.07472319602966308
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,10240,16384,0.10472640196482341
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,16384,0.12668906847635905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,16384,0.07911146481831868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,10240,12288,0.08471253712972006
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,10240,12288,0.05114346742630005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,12288,0.0893344004948934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,12288,0.05463253259658814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,10240,10240,0.043953065077463785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,10240,10240,0.07483306725819906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,10240,0.07522986729939779
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,10240,0.04652373393376668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,10240,8192,0.06527466773986816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,10240,8192,0.03627200126647949
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,8192,0.060842665036519375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,8192,0.03823146820068359
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,10240,7168,0.08006400267283122
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,7168,0.05358186562856039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,10240,7168,0.03257813254992167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,7168,0.03445546627044678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,10240,6144,0.07243200143178305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,10240,6144,0.028648533423741657
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,6144,0.04736640055974324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,6144,0.030781867106755574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,10240,5120,0.024975999196370443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,10240,5120,0.06466773351033529
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,5120,0.039574400583903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,5120,0.02688639958699544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,12288,12288,0.05973333517710368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,10240,4096,0.057416534423828124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,4096,0.020041600863138834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,4096,0.03431679805119832
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,10240,3584,0.052269868055979406
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,10240,3584,0.012088533242543538
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,3584,0.028937600056330365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,3584,0.014737066626548768
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,10240,3072,0.045339731375376384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,3072,0.025493333737055462
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,3072,0.012708266576131185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,10240,2560,0.040215468406677245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,10240,2560,0.00979200005531311
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,2560,0.02222933371861776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,2560,0.011496532956759136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,10240,2048,0.034694401423136394
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,10240,2048,0.008556800087292989
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,2048,0.01735573410987854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,2048,0.010452266534169514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,10240,1536,0.03180053234100342
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,1536,0.008909866213798523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,1536,0.009156266848246258
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,10240,1024,0.029309866825739543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,10240,1024,0.006757333377997081
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,1024,0.006920533378918965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,1024,0.008218666911125183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,10240,768,0.02765546639760335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,10240,768,0.0063381334145863845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,768,0.005985066791375478
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,768,0.007829333345095318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,10240,512,0.02606719930966695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,10240,512,0.006229333579540253
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,512,0.004983466863632202
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,512,0.007354666789372762
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,10240,256,0.024312533934911094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,10240,256,0.005436799923578898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,256,0.00409706657131513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,256,0.0068693334857622785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,10240,128,0.024130133787790935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,12288,2560,0.01011946698029836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,10240,128,0.005053866902987162
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,128,0.003626666714747747
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,128,0.006666666766007741
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,10240,64,0.02474453250567118
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,64,0.003399466723203659
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,12288,1536,0.007909333209196727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,64,0.00680213322242101
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,10240,32,0.024919466177622477
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,32,0.0034005333979924522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,12288,1024,0.006940799951553345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,32,0.006657066444555919
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,8192,65536,0.19373332659403483
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,65536,0.23138559659322105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,8192,16384,0.0887615998586019
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,8192,16384,0.05361920197804769
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,65536,0.38631680806477864
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,8192,65536,0.27102079391479494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,16384,0.06424640019734701
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,8192,12288,0.07354559898376464
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,16384,0.09977493286132813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,8192,12288,0.04748693307240804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,12288,0.050162132581075033
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,12288,0.08392319679260254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,8192,10240,0.06524693171183268
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,10240,0.06063359975814819
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,10240,0.04312106768290202
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,8192,8192,0.05705600182215372
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,8192,0.049719464778900144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,8192,8192,0.029958399136861165
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,8192,0.03510293165842692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,8192,7168,0.07240107059478759
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,8192,7168,0.02699306607246399
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,7168,0.04368533293406169
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,7168,0.0321343998114268
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,8192,6144,0.06497173309326172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,8192,6144,0.023908267418543495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,6144,0.038702932993570964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,6144,0.028394667307535808
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,8192,5120,0.0199455996354421
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,8192,5120,0.05834559996922811
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,5120,0.03745280106862386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,5120,0.024343466758728026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,8192,4096,0.045716265837351486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,8192,4096,0.01129919985930125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,4096,0.02714879910151164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,4096,0.014707199732462563
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,8192,3584,0.03925120035807292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,3584,0.024201599756876628
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,3584,0.013313066959381104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,8192,3072,0.03629333178202311
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,8192,3072,0.009760000308354696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,3072,0.021221333742141725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,3072,0.012542933225631714
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,8192,2560,0.033885868390401204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,8192,2560,0.00901759962240855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,2560,0.01660160024960836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,2560,0.011621333161989848
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,8192,2048,0.031218133370081586
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,8192,2048,0.008025600016117096
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,2048,0.009746133287747701
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,2048,0.010213333368301391
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,8192,1536,0.02943039933840434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,8192,1536,0.007354666789372762
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,1536,0.008094933132330577
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,1536,0.009301333626111349
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,8192,1024,0.02709226608276367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,8192,1024,0.006563200056552887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,1024,0.0064085334539413456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,10240,4096,0.019033600886662803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,1024,0.008205866813659668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,8192,768,0.02547840078671773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,8192,768,0.005963733295599619
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,768,0.005925333499908448
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,768,0.007762133578459422
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,10240,3072,0.010809600353240967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,8192,512,0.024359466632207234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,8192,512,0.0055861334005991616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,512,0.004739200075467428
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,512,0.007258666555086772
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,8192,256,0.022895999749501548
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,8192,256,0.005284266670544943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,256,0.003996799886226654
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,256,0.0068256000677744556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,8192,128,0.022498132785161336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,8192,128,0.004901333153247834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,128,0.003619199991226196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,10240,1536,0.007669333120187123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,128,0.006566399832566579
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,8192,64,0.022258132696151733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,64,0.003479466587305069
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,64,0.006609066824118296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,8192,32,0.022284799814224245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,32,0.0035829332967599234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,32,0.0066453332702318835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,7168,65536,0.1745845317840576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,7168,16384,0.04912213484446208
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,65536,0.17767893473307292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,7168,16384,0.08361706733703614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,16384,0.04970560073852539
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,7168,65536,0.25388479232788086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,16384,0.09075413544972738
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,65536,0.35183467864990237
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,7168,12288,0.06931413014729818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,12288,0.06511786778767904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,12288,0.04486293395360311
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,7168,10240,0.06190079847971598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,7168,10240,0.0372106671333313
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,10240,0.055579733848571775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,10240,0.034485332171122235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,8192,0.04351679881413777
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,8192,0.02887786626815796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,7168,8192,0.05504213174184164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,7168,8192,0.027562665939331054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,7168,7168,0.07036586602528891
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,7168,7168,0.025131734212239583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,7168,0.03943039973576863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,7168,0.026415999730428057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,7168,6144,0.06370346546173096
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,7168,6144,0.02224853237469991
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,6144,0.03459839820861817
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,6144,0.025964800516764325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,7168,5120,0.05661653280258179
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,7168,5120,0.014254933595657349
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,5120,0.029227733612060547
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,5120,0.013674666484196981
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,7168,4096,0.04440960089365641
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,7168,4096,0.010972799857457478
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,4096,0.025171200434366863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,4096,0.011427199840545655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,7168,3584,0.038990934689839676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,3584,0.021384533246358237
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,7168,3072,0.009633066256841023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,3584,0.010760533809661865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,7168,3072,0.036287999153137206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,3072,0.01795519987742106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,3072,0.009985066453615824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,7168,2560,0.03434880177179973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,7168,2560,0.008620799581209818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,2560,0.010291199882825215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,8192,10240,0.03628906806310018
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,2560,0.009307733178138733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,7168,2048,0.03144960006078084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,7168,2048,0.007648000121116638
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,2048,0.009532800316810608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,2048,0.008680533369382222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,7168,1536,0.02940373420715332
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,7168,1536,0.00699839989344279
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,1536,0.007478400071461995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,1536,0.007868800063927968
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,7168,1024,0.026986666520436603
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,7168,1024,0.006303999821345012
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,1024,0.006027733286221823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,1024,0.007146666447321574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,7168,768,0.026104533672332765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,7168,768,0.005829333265622457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,768,0.005233066777388254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,768,0.006821333368619282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,7168,512,0.02477759917577108
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,7168,512,0.0055402666330337524
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,512,0.004387199878692627
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,512,0.006503466765085857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,8192,3584,0.01053013304869334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,7168,256,0.02341866691907247
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,7168,256,0.00518506666024526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,7168,128,0.022460800409317017
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,256,0.003705599904060364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,256,0.006150400141874949
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,7168,128,0.004880000154177348
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,128,0.003605333218971888
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,128,0.00597866674264272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,7168,64,0.021661865711212158
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,64,0.003385599950949351
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,64,0.006073600053787232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,7168,32,0.02126400073369344
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,32,0.003455999990304311
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,32,0.006052266558011373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,6144,65536,0.15188159942626953
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,65536,0.1611562728881836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,6144,16384,0.07903040250142415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,6144,16384,0.043274664878845216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,6144,65536,0.2331530729929606
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,65536,0.29269011815389
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,16384,0.07921600341796875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,16384,0.04551253318786621
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,6144,12288,0.06554346481959025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,6144,12288,0.03826773166656494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,12288,0.036429866154988604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,12288,0.06347519954045613
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,6144,10240,0.02913813392321269
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,6144,10240,0.05844586690266927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,10240,0.048343467712402347
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,10240,0.031362134218215945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,6144,8192,0.051496533552805576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,6144,8192,0.024741333723068238
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,8192,0.040666667620340984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,8192,0.02664746642112732
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,6144,7168,0.06737279891967773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,6144,7168,0.022165334224700926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,7168,0.038753068447113036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,7168,0.023312000433603923
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,6144,6144,0.014254933595657349
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,6144,6144,0.061353600025177
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,6144,0.03109546701113383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,6144,0.014427733421325684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,6144,5120,0.05104426542917887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,6144,5120,0.01162559986114502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,5120,0.013300266861915589
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,5120,0.025973333915074663
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,6144,4096,0.040938667456309
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,6144,3584,0.03849173386891683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,4096,0.021342933177947998
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,4096,0.011136000355084736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,6144,3584,0.00938986639181773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,6144,3072,0.008939733107884724
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,3584,0.018837332725524902
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,3584,0.010614400108655293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,7168,12288,0.038957866032918294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,6144,3072,0.03688853184382121
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,3072,0.011927466591199238
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,3072,0.009816533327102661
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,6144,2560,0.03415040175120036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,2560,0.009249066313107807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,2560,0.009217066566149394
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,2048,0.008610133330027263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,6144,2048,0.031175466378529866
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,6144,2048,0.007316266496976216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,2048,0.008517332871754964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,6144,1536,0.029682133595148725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,1536,0.006770133475462596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,1536,0.007798400024573009
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,1024,0.0056618665655454
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,6144,1024,0.027159466346104937
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,6144,1024,0.006251733501752217
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,1024,0.0070602665344874065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,6144,768,0.025536000728607178
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,7168,3584,0.010261332988739014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,6144,768,0.00576853354771932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,768,0.004901333153247834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,768,0.006792533397674561
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,6144,512,0.0243231991926829
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,512,0.004343466460704803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,512,0.006445866823196411
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,6144,256,0.023000532388687135
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,256,0.003836799909671148
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,256,0.006137600044409434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,6144,128,0.022418133417765298
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,6144,128,0.004977066814899444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,128,0.0035989334185918174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,128,0.0059232001503308615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,6144,64,0.021643733978271483
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,64,0.0032437334458033243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,64,0.00595306654771169
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,6144,32,0.022294400135676067
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,32,0.0032661333680152893
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,32,0.006072533130645752
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,5120,65536,0.1290602684020996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,5120,65536,0.20810453097025552
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,65536,0.2428981304168701
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,65536,0.13113173643747966
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,5120,16384,0.07205866972605388
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,5120,16384,0.03813013235727946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,16384,0.06905066967010498
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,16384,0.038815999031066896
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,5120,12288,0.029795199632644653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,5120,12288,0.06038933197657267
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,12288,0.05182826519012451
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,12288,0.031133866310119628
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,5120,10240,0.025352533658345538
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,5120,10240,0.05442026853561401
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,10240,0.0403989315032959
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,10240,0.02712213397026062
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,5120,8192,0.048438398043314616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,5120,8192,0.021539199352264404
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,8192,0.03295253316561381
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,8192,0.019989333550135293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,5120,7168,0.0630741318066915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,7168,0.029165865977605183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,7168,0.01470080018043518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,5120,6144,0.05372373263041178
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,5120,6144,0.01241600016752879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,6144,0.025996800263722735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,6144,0.012429866194725036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,5120,5120,0.045304532845815024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,5120,5120,0.010935466488202412
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,5120,0.02220906615257263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,5120,0.011354666948318482
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,5120,4096,0.040532267093658446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,5120,4096,0.009620267152786254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,4096,0.01623146633307139
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,4096,0.010007466872533162
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,5120,3584,0.03738240003585815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,5120,3584,0.00909440020720164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,3584,0.010212266445159912
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,3584,0.009585066636403402
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,5120,3072,0.03527040084203084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,5120,3072,0.008540800213813782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,3072,0.009245866537094116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,3072,0.008964266379674275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,5120,2560,0.0338261326154073
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,5120,2560,0.007915733257929485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,2560,0.008195200065771738
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,2560,0.008494933446248371
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,5120,2048,0.007149866720040639
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,5120,2048,0.030732800563176472
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,2048,0.00705813318490982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,6144,4096,0.009934932986895243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,2048,0.00781333347161611
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,5120,1536,0.029423999786376952
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,5120,1536,0.0067104001839955645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,1536,0.006183466811974844
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,1536,0.007264000177383423
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,5120,1024,0.026918399333953857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,5120,1024,0.005994666616121928
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,1024,0.005165866514046987
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,1024,0.006786133348941803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,5120,768,0.02550506591796875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,5120,768,0.005676800012588501
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,768,0.004577066500981649
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,768,0.006571733454863231
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,5120,512,0.024605866273244223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,5120,512,0.005400533477465311
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,512,0.004105599969625473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,6144,2560,0.008130133152008057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,512,0.0062943999965985615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,5120,256,0.02298240065574646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,5120,256,0.005042133231957754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,256,0.0036490666369597114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,256,0.0060597335298856105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,6144,1536,0.006837333242098491
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,5120,128,0.022345600525538127
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,5120,128,0.004902400076389313
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,128,0.003409066547950109
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,128,0.005881600081920624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,5120,64,0.021588265895843506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,64,0.0032373333970705668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,64,0.006045866509278615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,5120,32,0.021062399943669638
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,32,0.0032629333436489105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,32,0.006027733286221823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,6144,512,0.005397333204746247
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,4096,16384,0.06699413458506266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,4096,16384,0.032398933172225954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,4096,65536,0.18313706715901693
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,65536,0.11207786401112874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,65536,0.19523839950561522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,6144,256,0.005065600077311197
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,16384,0.05801706711451212
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,16384,0.0339349349339803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,4096,12288,0.05614293416341146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,4096,12288,0.025231999158859254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,12288,0.042771200338999435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,12288,0.02736213405927022
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,4096,10240,0.050779732068379727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,4096,10240,0.02028800050417582
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,10240,0.036690131823221846
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,10240,0.02214399973551432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,8192,0.030643200874328612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,8192,0.013522133231163025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,4096,8192,0.043366400400797527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,4096,8192,0.013458133737246195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,4096,7168,0.01232426663239797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,4096,7168,0.05749760071436564
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,7168,0.027964800596237183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,7168,0.0118367999792099
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,4096,6144,0.0482367992401123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,4096,6144,0.01130986710389455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,6144,0.021730132897694907
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,6144,0.010973866780598958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,4096,5120,0.04380160172780355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,4096,5120,0.010345600048700969
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,5120,0.016820265849431356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,5120,0.010262399911880493
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,4096,4096,0.03885013262430827
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,4096,0.008970666925112407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,4096,3584,0.037564798196156816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,4096,0.009099733829498292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,4096,3584,0.008598400155703227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,3584,0.008362666765848795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,3584,0.008770133058230083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,4096,3072,0.03512639999389648
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,3072,0.00796693315108617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,3072,0.008400000135103862
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,4096,2560,0.03299626708030701
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,4096,2560,0.007445333401362102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,2560,0.007934933404127757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,2560,0.007931733131408691
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,4096,2048,0.030400000015894574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,2048,0.006939733525117238
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,5120,7168,0.014507733782132468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,2048,0.007467733323574066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,4096,1536,0.028603732585906982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,4096,1536,0.006378666559855144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,1536,0.005786666770776113
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,1536,0.007047466437021891
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,4096,1024,0.02683839996655782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,4096,1024,0.005894400179386139
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,1024,0.00487253318230311
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,1024,0.006602666775385539
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,4096,768,0.025305600961049397
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,4096,768,0.005565866827964783
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,768,0.004380799829959869
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,768,0.006423466900984447
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,4096,512,0.024025599161783852
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,4096,512,0.005294933418432872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,512,0.003987200061480204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,512,0.006237866481145223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,4096,256,0.022793600956598915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,4096,256,0.005015466610590617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,256,0.0035114665826161706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,256,0.006020266811052958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,4096,128,0.022273067633310953
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,128,0.0035135999321937563
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,128,0.005849599838256836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,4096,64,0.021553067366282146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,64,0.005905066430568695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,64,0.0031167998909950255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,32,0.0031295999884605407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,4096,32,0.021112533410390218
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,32,0.005997866888840993
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3584,65536,0.1730965296427409
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,65536,0.17161067326863605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,65536,0.10099946657816569
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3584,16384,0.06353066762288412
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3584,16384,0.029572266340255737
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,16384,0.050836265087127686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,16384,0.031057065725326537
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3584,12288,0.021527467171351115
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3584,12288,0.054472533861796055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,12288,0.04086933135986328
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,12288,0.02531840006510417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3584,10240,0.04890666802724202
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3584,10240,0.016089600324630738
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,10240,0.033779199918111166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,10240,0.016145066420237223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3584,8192,0.040677332878112794
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3584,8192,0.013081600268681845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,8192,0.029623466730117797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,8192,0.012644267082214356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3584,7168,0.01213653286298116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3584,7168,0.05355519851048788
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,7168,0.025522132714589436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,7168,0.011841066678365071
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3584,6144,0.048265600204467775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3584,6144,0.011096533139546711
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,6144,0.021897600094477335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,6144,0.010980266332626342
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3584,5120,0.04423679908116658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3584,5120,0.010497066378593444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,5120,0.011245866616566975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,5120,0.01051200032234192
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3584,4096,0.03924586772918701
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3584,4096,0.008803199728329976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,4096,65536,0.10668266614278157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,4096,0.009092266360918682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,4096,0.009196799993515015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3584,3584,0.03758506774902344
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3584,3584,0.008517332871754964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,3584,0.00832426647345225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,3584,0.008750933408737182
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3584,3072,0.03521813154220581
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,3072,0.007812266548474629
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,3072,0.008289066453774769
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3584,2560,0.03352746566136678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3584,2560,0.007325866818428039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,2560,0.007118933399518331
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,2560,0.007957333326339721
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3584,2048,0.03118720054626465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3584,2048,0.00680213322242101
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,2048,0.006161066889762879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,2048,0.007464533547560374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3584,1536,0.029198932647705077
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3584,1536,0.006284800171852112
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,1536,0.005633066594600678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,1536,0.0070271998643875126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3584,1024,0.02717120051383972
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3584,1024,0.005750399827957153
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,1024,0.00469760000705719
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,4096,3072,0.008160000046094257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,4096,4096,0.008918399612108868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,1024,0.006640000144640605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3584,768,0.0252074658870697
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3584,768,0.005475200215975444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,768,0.004268800218900045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,768,0.0064522668719291685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3584,512,0.024016000827153525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3584,512,0.005285333096981049
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,512,0.003918933371702829
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3584,256,0.004952533543109894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,512,0.006248533229033152
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3584,256,0.022581332921981813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,256,0.0035103999078273775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,256,0.006002133091290792
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3584,128,0.022201599677403767
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3584,128,0.004747733473777771
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,4096,2048,0.006797866523265838
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,128,0.003256533294916153
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,128,0.005832533538341522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3584,64,0.021434666713078816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,64,0.003110400090614955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,64,0.0058218667904535925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3584,32,0.021026132504145305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,32,0.0031786667803923286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,32,0.005870933334032694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3072,65536,0.16405547459920247
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,65536,0.09227519830067953
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3072,16384,0.06149653196334839
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,65536,0.15081386566162108
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3072,16384,0.026162133614222212
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,4096,128,0.004785066843032837
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,16384,0.042574934164683026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,16384,0.028318933645884198
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3072,12288,0.051785600185394284
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3072,12288,0.01867626706759135
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,12288,0.03489813407262166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,12288,0.018796799580256145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3072,10240,0.046691199143727616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3072,10240,0.015544533729553223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,10240,0.0294815997282664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,10240,0.015335466464360556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3072,8192,0.036975999673207596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3072,8192,0.012958932916323343
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,8192,0.012643200159072877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,8192,0.02563199996948242
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3072,7168,0.05186560153961182
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3072,7168,0.01204373339811961
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,7168,0.022613332668940226
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,7168,0.011804800232251484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3072,6144,0.04681813319524129
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,6144,0.013018666704495748
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3584,65536,0.09717866579691568
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,6144,0.01090666651725769
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3072,5120,0.042734932899475095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3072,5120,0.010294399658838908
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,5120,0.0110944002866745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,5120,0.010225066542625427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3072,4096,0.0387168010075887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3072,4096,0.008743466933568318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,4096,0.009040000041325887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,4096,0.009038933118184407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3072,3584,0.03671573400497437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3072,3584,0.008329600095748901
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,3584,0.008239999910195668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,3584,0.008729599912961324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3072,3072,0.03537813425064087
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,3072,0.00774186650911967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,3072,0.008329600095748901
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3072,2560,0.0328000009059906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3072,2560,0.007355733215808869
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,2560,0.007037866612275441
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,2560,0.007930666704972585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3072,2048,0.03038826584815979
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3072,2048,0.00673280010620753
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,2048,0.00626986672480901
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,2048,0.00739519993464152
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3072,1536,0.029176533222198486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3072,1536,0.006281599899133046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,1536,0.005622399846712748
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,1536,0.007017600039641063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3072,1024,0.026549333333969118
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3072,1024,0.005656533439954122
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,1024,0.004701866706212362
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,1024,0.006597333153088887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3072,768,0.02520853281021118
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3072,768,0.0054730668663978575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,768,0.004311466713746389
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,768,0.006372266511122386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3072,512,0.024422399202982583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3072,512,0.0052704001466433205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,512,0.00386559988061587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,512,0.006259199976921081
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3072,256,0.02290453314781189
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3072,256,0.0048437332113583885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,256,0.003487999985615412
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,256,0.005993600189685822
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3072,128,0.0221343994140625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3072,128,0.0046847999095916745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,128,0.0032138665517171226
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3584,3072,0.00802346666653951
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,128,0.0058783998092015585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3072,64,0.022081067164738975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,64,0.0031061333914597826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,64,0.005896533528963724
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,3072,32,0.020985599358876547
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,32,0.0031456001102924346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,32,0.0058112000425656635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2560,65536,0.0796288013458252
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2560,65536,0.15741759936014812
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,65536,0.12856427033742268
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,65536,0.08470933437347412
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2560,16384,0.0597055991490682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2560,16384,0.024074665705362954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,16384,0.036541867256164554
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,16384,0.025251199801762898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2560,12288,0.05009066661198934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2560,12288,0.017629865805308023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,12288,0.02916693290074666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,12288,0.015957333644231162
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2560,10240,0.04382400115331014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2560,10240,0.015473066767056783
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,10240,0.014389333128929139
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,10240,0.025169066588083905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2560,8192,0.034169598420461016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2560,8192,0.012972799936930337
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,8192,0.021079466740290324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,8192,0.012502400080362954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2560,7168,0.05276266733805338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2560,7168,0.011915733416875202
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,7168,0.013802666465441385
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,7168,0.011779200037320454
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2560,6144,0.048292267322540286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2560,6144,0.010925867160161336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,6144,0.011788800358772278
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,5120,0.01053546667098999
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,6144,0.01093226671218872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2560,5120,0.04306453466415405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2560,5120,0.00990613301595052
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,5120,0.01013866662979126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2560,4096,0.038439468542734785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2560,4096,0.008732799688975017
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,4096,0.009108266234397889
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,4096,0.009275733431180318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2560,3584,0.03742400010426839
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2560,3584,0.008248533308506011
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,3584,0.008334933718045553
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,3584,0.00869653324286143
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2560,3072,0.03519573211669922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2560,3072,0.007772799829641979
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,3072,0.0076906666159629825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3072,65536,0.08818666934967041
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,3072,0.008231466511885326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2560,2560,0.032630399862925215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2560,2560,0.0072522665063540145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,2560,0.006769066552321117
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,2560,0.007837866743405659
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2560,2048,0.030190932750701904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,2048,0.006113066772619883
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,2048,0.007380266487598419
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2560,1536,0.028613332907358808
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2560,1536,0.006205866734186808
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,1536,0.0054624001185099285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,1536,0.007051733136177063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2560,1024,0.026594134171803792
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2560,1024,0.005740800003210703
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,1024,0.004567466676235199
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,1024,0.006629333396752675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,768,0.006378666559855144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3072,6144,0.011004799604415893
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2560,768,0.025255467494328814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2560,768,0.005463466544946035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,768,0.004227200150489807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2560,512,0.02405866583188375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2560,512,0.00513919989267985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2560,256,0.004818133513132731
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,512,0.003869866579771042
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,512,0.006243200103441874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2560,256,0.023085866371790567
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,256,0.003458133339881897
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,256,0.005967999994754791
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,128,0.005997866888840993
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2560,128,0.022356265783309938
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,128,0.003193599979082743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2560,64,0.021412267287572225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,64,0.003036800026893616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,64,0.005875200033187866
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3072,3072,0.007766399780909221
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2560,32,0.022038400173187256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,32,0.0031114667654037476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,32,0.0058559998869895935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2048,65536,0.07370773156483969
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2048,65536,0.14929173787434896
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,65536,0.10451520284016927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,65536,0.0775274674097697
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2048,16384,0.05664106607437134
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2048,16384,0.020972800254821778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,16384,0.030395734310150146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,16384,0.020747733116149903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,12288,0.02469866673151652
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2048,12288,0.04702613353729248
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2048,12288,0.01639573375384013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,12288,0.01593493322531382
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2048,10240,0.04072639942169189
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2048,10240,0.014538666605949402
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2048,8192,0.03389866749445598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,10240,0.020796799659729005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,10240,0.014225066701571146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2048,8192,0.012427733341852824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,8192,0.012199466427167256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,8192,0.012392533818880717
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2048,7168,0.051335465908050534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2048,7168,0.011527466773986816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,7168,0.011506133278210958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,7168,0.011591466267903645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2048,6144,0.04719680150349935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2048,6144,0.010501333077748616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2048,5120,0.04279786745707194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,6144,0.011612799763679505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,6144,0.01076800028483073
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2048,4096,0.03863573471705119
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2048,5120,0.009648000200589497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,5120,0.009629866480827332
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,5120,0.009988266229629516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2048,4096,0.008520533641179402
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,4096,0.008855467041333515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,3584,0.008514133095741273
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,4096,0.008919466535250347
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2048,3584,0.036789333820343016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2048,3584,0.008036266764005024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,3584,0.007736533383528392
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2048,3072,0.035239466031392414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2048,3072,0.007526400188604991
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,3072,0.007223466535409291
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,3072,0.008154666423797608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2048,2560,0.032824534177780154
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2048,2560,0.007108266651630402
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,2560,0.006524799764156342
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,2560,0.007694933315118153
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2048,2048,0.030118399858474733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2048,2048,0.006545066833496094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,2048,0.005752533177534739
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,2048,0.007287466526031494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2048,1536,0.028787199656168622
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2048,1536,0.006041599810123444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,1536,0.005115733544031779
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,1536,0.006899199883143107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2048,1024,0.026628265778223675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,1024,0.004418133199214936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,1024,0.0064746667941411335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2048,768,0.024989867210388185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2048,768,0.005444266895453135
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,768,0.004056533426046371
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,768,0.00633493314186732
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2048,512,0.023875200748443605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,512,0.0038399999340375268
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,512,0.006074666480223338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2048,256,0.02312320073445638
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,256,0.003469866762558619
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,256,0.005895466605822245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2048,128,0.022627200682957968
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,128,0.0031456001102924346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,128,0.0057781333724657696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2048,64,0.021290665864944457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,64,0.0029834667841593427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,64,0.0057888001203536986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,2048,32,0.02204586664835612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,32,0.0031040000418821974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,32,0.005804799993832906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2560,2048,0.006657066444555919
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1536,65536,0.07193919817606607
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1536,65536,0.1432202657063802
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,65536,0.08220586776733399
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,65536,0.07173013687133789
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1536,16384,0.05327146848042806
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1536,16384,0.020150399208068846
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,16384,0.024972800413767496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,16384,0.01941759983698527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1536,12288,0.042713598410288496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1536,12288,0.01623040040334066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,12288,0.014975999792416891
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,12288,0.015897599856058757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1536,10240,0.014375467101732889
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1536,10240,0.03891093333562215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,10240,0.012221866846084594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,10240,0.014246400197347006
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1536,8192,0.03339840173721313
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1536,8192,0.012245333194732666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,8192,0.010594133536020916
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,8192,0.01239466667175293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1536,7168,0.05222506523132324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,7168,0.01011946698029836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,7168,0.011668266852696736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1536,6144,0.04751786788304647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1536,6144,0.010441600282986959
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,6144,0.00940053363641103
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,6144,0.010890666643778484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1536,5120,0.043442134062449136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1536,5120,0.009461333354314167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2560,128,0.004658133288224538
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,5120,0.009149866302808125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,5120,0.010075733065605164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1536,4096,0.03890560070673625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1536,4096,0.008469333251317341
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,4096,0.008326399823029835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,4096,0.009011200070381165
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1536,3584,0.037529599666595456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,3584,0.007885866860548655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,3584,0.008549333612124125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1536,3072,0.035010135173797606
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1536,3072,0.0074890668193499255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,3072,0.0071712002158164975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,3072,0.00819946676492691
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1536,2560,0.032979200283686325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,2560,0.006622933348019918
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,2560,0.007788800199826558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1536,2048,0.03095039923985799
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1536,2048,0.006516266862551372
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,2048,0.005814399818579356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,2048,0.007313066720962524
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1536,1536,0.029050666093826293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1536,1536,0.006064000229040781
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,1536,0.005130666494369507
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,1536,0.0069354668259620665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1536,1024,0.026525866985321046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1536,1024,0.005523199836413065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,1024,0.004385066529115042
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,1024,0.006548266609509785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1536,768,0.025620265801747637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1536,768,0.005302399893601735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,768,0.004049066702524821
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,768,0.006393600006898243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1536,512,0.024235733350118003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1536,512,0.0050240000089009605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,512,0.0037280000746250153
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,512,0.006120533247788747
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1536,256,0.022849067052205404
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1536,256,0.004783999919891357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,256,0.0033770665526390077
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,256,0.005925333499908448
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1536,128,0.02254826625188192
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1536,128,0.004542933404445648
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,128,0.0031093334158261614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,128,0.005791999896367391
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1536,64,0.02198293407758077
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,64,0.002993066608905792
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,64,0.005843199789524078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2048,1024,0.005684266487757364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1536,32,0.021978666385014854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,32,0.0030016000072161358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,32,0.005798399945100148
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1024,65536,0.07190399964650472
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1024,65536,0.1399168014526367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,65536,0.05893333355585734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,65536,0.06866772969563803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1024,16384,0.05044480164845785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2048,512,0.005121066669623057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1024,16384,0.020111999909083047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,16384,0.01234773298104604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,16384,0.019236266613006592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1024,12288,0.03995733261108399
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1024,12288,0.016177067160606386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,12288,0.010279466708501179
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,12288,0.015844266613324484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2048,256,0.0048437332113583885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1024,10240,0.03673386573791504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2048,128,0.004576000074545542
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1024,10240,0.014147200187047324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,10240,0.008930133779843648
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,10240,0.01423893372217814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1024,8192,0.03338239987691243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1024,8192,0.01220906674861908
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,8192,0.00801386684179306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,8192,0.012345600128173827
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1024,7168,0.051254399617513026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1024,7168,0.011273599664370219
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,7168,0.007778133451938629
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,7168,0.011595732967058818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1024,6144,0.04709226687749227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1024,6144,0.010418132940928141
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,6144,0.009331199526786804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,6144,0.010755200187365215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1024,5120,0.04256960153579712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1024,5120,0.00941973328590393
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,5120,0.008709333340326945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1024,4096,0.038551465670267744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,5120,0.010019200046857198
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1024,4096,0.00841813286145528
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,4096,0.008006399869918824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1024,3584,0.036467198530832926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,4096,0.008914132912953694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1024,3584,0.007969066500663757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,3584,0.007736533383528392
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,3584,0.008497066299120585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1024,3072,0.03429333368937175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1024,3072,0.007396266857783
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,3072,0.007326933244864146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,3072,0.008123733103275299
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1024,2560,0.032494932413101196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1024,2560,0.006894933183987935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,2560,0.006466133395830791
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1536,7168,0.01144426663716634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,2560,0.007736533383528392
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1024,2048,0.029909332593282063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1024,2048,0.006404266754786174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,2048,0.005770666897296906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,2048,0.007268266876538594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1024,1536,0.028485333919525145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1024,1536,0.006157866617043813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,1536,0.005020800232887268
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,1536,0.006878933310508728
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1024,1024,0.02662613391876221
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,1024,0.00432640016078949
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,1024,0.006467199822266896
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1024,768,0.02496853272120158
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1024,768,0.005286400020122528
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,768,0.0039893334110577905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,768,0.006258133550484974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1024,512,0.024462932348251344
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1024,512,0.005018666883309682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,512,0.0036703998843828833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,512,0.006055466830730438
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1536,3584,0.00802346666653951
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1024,256,0.02257386644681295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1024,256,0.0046858668327331545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,256,0.0033312000334262846
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,128,0.005767466624577841
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,256,0.00584853341182073
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1536,2560,0.0069482664267222095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1024,128,0.022073600689570108
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1024,128,0.004609066744645437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,128,0.003070933371782303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1024,64,0.021359999974568687
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,64,0.0031626666585604347
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,64,0.005748266478379568
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,768,65536,0.07114986578623453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,1024,32,0.020949333906173706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,32,0.003045333425203959
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,32,0.005765333275000254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,768,65536,0.13484800656636556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,65536,0.04801493485768636
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,65536,0.06735359827677409
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,768,16384,0.04827839930852254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,768,16384,0.01992959976196289
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,16384,0.012654933333396911
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,16384,0.01941546599070231
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,768,12288,0.039544534683227536
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,768,12288,0.01614720026652018
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,12288,0.010745599865913391
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,12288,0.015884799758593242
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,768,10240,0.036245334148406985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,768,10240,0.014169599612553915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,10240,0.010166399677594503
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,10240,0.014267733693122864
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,768,8192,0.03334186474482219
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,768,8192,0.012124799688657125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,8192,0.009027199943860371
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,8192,0.012389333049456278
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,768,7168,0.051361068089803064
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,768,7168,0.011313066879908244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,7168,0.008674133817354839
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,7168,0.011587199568748475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,768,6144,0.04692480166753133
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,768,6144,0.010283733407656353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,6144,0.008062933385372163
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,6144,0.010807466506958009
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,768,5120,0.04296213388442993
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,768,5120,0.00944640040397644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,5120,0.007623466849327088
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,5120,0.01002239982287089
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,768,4096,0.03849173386891683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,768,4096,0.008295466502507527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,4096,0.007277866701285045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,4096,0.009034666419029235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,768,3584,0.03662933508555095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,768,3584,0.007865599791208903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,3584,0.007468800246715546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,3584,0.00860053300857544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,768,3072,0.0348960002263387
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,768,3072,0.007393066585063934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,3072,0.007169066866238911
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,3072,0.008142933249473572
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,768,2560,0.032459733883539836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,768,2560,0.0068896000583966565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,2560,0.006379733482996623
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,2560,0.007793066898981731
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,768,2048,0.02996693253517151
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,768,2048,0.006381866832574208
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,2048,0.005662933488686879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,2048,0.007286400099595388
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,768,1536,0.028525867064793903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,768,1536,0.0060032000144322716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,1024,0.00429013321797053
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,1536,0.004985600213209788
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,1536,0.00690773328145345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,768,1024,0.026426666975021364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,768,1024,0.005459199845790863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,1024,0.006559999783833821
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,768,768,0.025085866451263428
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,768,768,0.005194666484991709
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,768,0.003940266619126002
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,768,0.00631573349237442
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,768,512,0.02404586672782898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,512,0.00363520011305809
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,512,0.006090666850407918
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,768,256,0.022925867636998495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,768,256,0.004700799783070883
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,256,0.003303466737270355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,128,0.0030602666238943735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,256,0.005914666752020518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,768,128,0.022127999862035116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,768,128,0.0045621335506439206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,128,0.005846400062243144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,768,64,0.021336533625920615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,64,0.002942933390537898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,64,0.005761066575845083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,768,32,0.021948800484339396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,512,65536,0.07154453595479329
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,32,0.0057781333724657696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,32,0.002948266764481862
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,512,65536,0.13269120057423908
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,65536,0.03595199982325236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,65536,0.06753173669179281
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,512,16384,0.046649599075317384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,512,16384,0.0200053334236145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,16384,0.009915733337402343
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,16384,0.01915093262990316
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,512,12288,0.03888533512751262
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,512,10240,0.03608640034993489
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,512,12288,0.016330666343371072
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,12288,0.008780800302823384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,12288,0.015828266739845276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,512,10240,0.01421333352724711
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,10240,0.00817920019229253
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,10240,0.014149333039919535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,512,8192,0.03330026666323344
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,512,8192,0.01221440037091573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,8192,0.007710933188597361
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,8192,0.01226026713848114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,512,7168,0.05105813344319662
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,7168,0.007402666906515758
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1024,1024,0.005460266768932342
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,7168,0.011592533191045125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,512,6144,0.04678613344828288
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,512,6144,0.010375466942787171
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,6144,0.007042133311430614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,6144,0.010784000158309937
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,512,5120,0.042771200338999435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,5120,0.007451733450094859
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,5120,0.009909333785374959
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,512,4096,0.038423466682434085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,512,4096,0.008326399823029835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,4096,0.007101866602897644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,4096,0.008938666184743245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,512,3584,0.03723520040512085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,512,3584,0.007875200112660725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,512,3072,0.03455679814020793
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,3584,0.0074442664782206226
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,3584,0.008570667107899983
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,512,3072,0.0073525334397951765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,512,2560,0.03293760021527608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,3072,0.007150933146476746
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,3072,0.008124800026416778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,512,2560,0.006846933563550313
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,2560,0.006230400005976359
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,2560,0.007709866762161255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,512,2048,0.030006400744120282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,512,2048,0.0063178668419520065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,2048,0.005637333293755849
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,2048,0.007188266515731812
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,512,1536,0.028706133365631104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,512,1536,0.005970133344332377
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,1536,0.004955733319123586
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,1536,0.006886399785677592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,512,1024,0.026624000072479247
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,512,1024,0.005452799797058106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,1024,0.004271999994913737
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,1024,0.006517333288987477
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,512,768,0.025167999664942424
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,512,768,0.005190399785836538
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,768,0.0039018665750821433
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,768,0.0063296000162760425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,512,0.0035914666950702667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,512,512,0.02390399972597758
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,512,512,0.0049674664934476215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,512,0.006072533130645752
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,512,256,0.022600533564885457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,512,256,0.004693333307902018
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,256,0.0032960000137488045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,256,0.005861333509286245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,512,128,0.022078933318456014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,128,0.0030826665461063385
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,128,0.0057888001203536986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,512,64,0.021305600802103676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,64,0.002946133414904277
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,64,0.005772800246874491
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,512,32,0.020853332678476968
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,32,0.002914133419593175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,32,0.005772800246874491
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,256,65536,0.13117012977600098
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,65536,0.02395306626955668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,65536,0.062376534938812254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,256,16384,0.045269334316253663
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,256,16384,0.019986132780710854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,16384,0.00785813331604004
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,16384,0.01914773384730021
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,256,12288,0.038786133130391434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,256,12288,0.01625599960486094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,12288,0.007303466896216075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,12288,0.015787733594576518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,256,10240,0.03608959913253784
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,256,10240,0.014263466993967692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,10240,0.007411199808120728
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,10240,0.01420266628265381
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,256,8192,0.033122134208679196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,256,8192,0.01227839986483256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,8192,0.007292800148328145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,8192,0.012325333555539449
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,256,7168,0.051083731651306155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,256,7168,0.011242666840553283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,7168,0.00718506673971812
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,7168,0.011552000045776367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,256,6144,0.047261865933736165
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,256,6144,0.01030293305714925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,6144,0.006970666845639547
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,6144,0.010722133517265319
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,256,5120,0.04270933469136556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,256,5120,0.009320533275604248
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,5120,0.007358933488527934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,5120,0.009958400328954061
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,256,4096,0.03837546507517497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,256,4096,0.008286933104197185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,4096,0.006955733398596446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,4096,0.008893866340319316
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,256,3584,0.037191466490427656
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,3584,0.007421866556008657
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,256,3584,0.007910400132338206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,3584,0.00844373305638631
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,256,3072,0.03466560045878093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,256,3072,0.007353599866231282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,3072,0.007092266778151194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,3072,0.00808426688114802
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,256,2560,0.006808533271153767
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,256,2560,0.03242133259773254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,2560,0.006267733375231425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,2560,0.00767573316891988
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,256,2048,0.030161066850026445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,2048,0.0055754666527112326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,2048,0.007246933380762736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,256,1536,0.028188800811767577
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,256,1536,0.005970133344332377
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,1536,0.0048991998036702475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,1536,0.006821333368619282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,256,1024,0.026411734024683636
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,768,512,0.0050687998533248905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,256,1024,0.005427200098832448
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,1024,0.0042346666256586705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,1024,0.0064522668719291685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,256,768,0.02501759926478068
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,256,768,0.005155199766159057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,768,0.003922133396069208
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,768,0.00626453310251236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,256,512,0.02381226619084676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,256,512,0.004941866795221964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,512,0.0060032000144322716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,512,0.0035573333501815797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,256,256,0.022662399212519328
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,256,256,0.004625066618124644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,256,0.0032405334214369455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,256,0.005873066683610281
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,256,128,0.022014933824539184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,256,128,0.004558933277924856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,128,0.003068800022204717
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,128,0.00576853354771932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,256,64,0.02134933272997538
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,64,0.0028885332246621448
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,64,0.005729066828886667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,32,0.0029088000456492106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,256,32,0.020760534207026164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,32,0.0057205334305763245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,128,65536,0.13012693723042806
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,128,65536,0.06859626770019531
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,65536,0.013326932986577352
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,65536,0.05983999967575073
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,128,16384,0.04461333354314168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,128,16384,0.01989333430926005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,16384,0.007144533097743988
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,16384,0.01913493275642395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,128,12288,0.03876373370488485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,128,12288,0.016139733791351318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,12288,0.007048533360163371
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,12288,0.015703466534614564
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,128,10240,0.036023465792338054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,128,10240,0.014176000157992044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,10240,0.007231999933719635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,10240,0.014123732844988504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,128,8192,0.03315200010935466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,512,5120,0.009406933188438415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,128,8192,0.01221440037091573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,8192,0.007192533214886982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,8192,0.012250666817029316
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,512,7168,0.011215999722480774
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,7168,0.0069472000002861025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,128,7168,0.05084799925486246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,7168,0.011464533209800721
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,128,6144,0.04748800198237101
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,128,6144,0.010240000486373902
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,6144,0.0068458666404088335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,6144,0.0106495996316274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,128,5120,0.042537601788838704
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,128,5120,0.009220266342163086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,5120,0.00720000018676122
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,5120,0.009922132889429728
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,128,4096,0.038321065902709964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,128,4096,0.008254933357238769
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,4096,0.006782933572928111
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,4096,0.00883840024471283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,128,3584,0.036968533198038736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,128,3584,0.00790826678276062
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,3584,0.007246933380762736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,3584,0.008429867029190064
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,128,3072,0.034884266058603924
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,128,3072,0.007277866701285045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,3072,0.008062933385372163
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,3072,0.006940799951553345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,128,2560,0.03244373401006063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,128,2560,0.006774400174617767
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,2560,0.00622080018122991
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,2560,0.007709866762161255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,128,2048,0.030168533325195312
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,128,2048,0.006302933394908905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,2048,0.00552106648683548
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,2048,0.007256533205509186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,128,1536,0.02834239999453227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,1536,0.004881066580613455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,1536,0.00689279983441035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,128,1024,0.02648640076319377
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,128,1024,0.005339733262856802
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,1024,0.004270933568477631
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,1024,0.006426666676998138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,128,768,0.024958932399749757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,512,128,0.0045855998992919925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,128,768,0.005122133096059163
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,768,0.0039018665750821433
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,768,0.006262399752934774
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,128,512,0.02387839953104655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,128,512,0.004855466882387797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,128,256,0.02269866665204366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,512,0.0035904000202814737
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,256,0.0031968000034491217
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,512,0.006045866509278615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,128,256,0.0046293333172798155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,256,0.005849599838256836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,128,128,0.021821866432825722
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,128,128,0.004297600189844767
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,128,0.005734399954477946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,128,0.003068800022204717
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,128,64,0.02130240003267924
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,64,0.00288426677385966
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,64,0.005725866556167603
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,128,32,0.02094506621360779
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,32,0.002869333326816559
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,32,0.005672533313433329
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,64,65536,0.13009493350982665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,65536,0.011585066715876263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,65536,0.05991573333740234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,64,16384,0.04466346502304077
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,16384,0.006968533496061961
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,16384,0.019242666165033975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,64,12288,0.038737066586812335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,12288,0.006876799960931141
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,12288,0.015715199708938598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,64,10240,0.03614720106124878
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,10240,0.007102933526039123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,10240,0.014089600245157877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,64,8192,0.03327359954516093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,8192,0.006909866631031036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,8192,0.012258133292198181
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,64,7168,0.05096746683120727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,7168,0.006816000243028005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,7168,0.011508267124493916
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,64,6144,0.04697813193003337
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,6144,0.006676266590754191
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,6144,0.0107424000898997
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,64,5120,0.04298666715621948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,5120,0.007028266787528992
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,256,65536,0.06980266571044921
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,5120,0.009940266609191895
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,64,4096,0.0385098656018575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,4096,0.006651733318964641
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,4096,0.008929066856702169
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,3584,0.008514133095741273
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,64,3584,0.03696533441543579
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,3584,0.007075199981530507
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,3072,0.007030400137106578
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,64,3072,0.035027201970418295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,3072,0.008107733229796093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,64,2560,0.03251519997914632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,2560,0.006159999966621399
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,2560,0.0077674667040507
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,2048,0.007235200206438701
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,64,2048,0.03006826639175415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,64,1536,0.028888533512751263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,2048,0.0055178667108217875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,1536,0.004806399842103322
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,1536,0.006846933563550313
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,64,1024,0.026855466763178508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,1024,0.004214400053024292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,1024,0.0064522668719291685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,64,768,0.024951465924580894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,768,0.00384853333234787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,768,0.006257066627343495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,64,512,0.023947733640670776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,512,0.0035434665779272715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,512,0.006109866499900818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,64,256,0.022722133000691733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,256,0.003173333406448364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,256,0.005807999769846598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,64,128,0.02257173260052999
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,128,0.0030570665995279947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,128,0.005694933235645294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,64,64,0.02104960083961487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,64,0.00290133332212766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,64,0.005758933226267497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,64,32,0.02084266742070516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,32,0.0028704000016053517
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,32,0.0056639999151229855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,65536,0.011471999684969585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,32,65536,0.13074986934661864
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,256,2048,0.00631573349237442
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,65536,0.05990720192591349
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,32,16384,0.045201067129770914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,16384,0.0070165331164995836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,16384,0.019146666924158732
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,32,12288,0.03886293172836304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,12288,0.00689386675755183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,12288,0.015625600020090738
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,32,10240,0.03591253360112508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,10240,0.007043200234572093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,10240,0.014916266997655234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,7168,0.006773333251476288
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,8192,0.006920533378918965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,32,8192,0.03317760030428569
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,8192,0.012264532844225566
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,32,7168,0.05097493330637613
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,7168,0.01156160036722819
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,32,5120,0.042897065480550126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,32,6144,0.04675626754760742
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,6144,0.0066101332505544026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,32,4096,0.038414935270945236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,6144,0.010761599739392598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,5120,0.007004799942175548
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,32,3584,0.03744213183720906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,5120,0.00993173321088155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,4096,0.006570666531721751
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,4096,0.008829866846402485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,3584,0.008447999755541485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,3584,0.007118933399518331
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,32,3072,0.03456533352533976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,3072,0.006632533172766368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,3072,0.008016000191370647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,32,2560,0.0326090673605601
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,2560,0.006145066519578298
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,2560,0.007729066908359528
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,32,2048,0.030135466655095415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,2048,0.005500799914201101
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,2048,0.00718506673971812
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,32,1536,0.028328533967336016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,1536,0.00483840008576711
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,1536,0.0068256000677744556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,32,1024,0.02650986711184184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,1024,0.004165333261092504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,1024,0.006392533580462138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,32,768,0.024963200092315674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,768,0.003807999938726425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,768,0.006300800045331319
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,32,512,0.02386346658070882
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,512,0.0034954667091369627
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,512,0.005980800092220307
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,32,256,0.022651733954747517
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,256,0.0031829332311948144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,256,0.00580266664425532
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,32,128,0.022030933698018392
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,128,0.0029653333127498626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,128,0.005662933488686879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,32,64,0.02215786576271057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,64,0.0028746667007605235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,64,0.00565226674079895
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,64,32,32,0.020696532726287842
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,32,0.0028586665789286296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,32,0.005640533566474914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,128,7168,0.011250133315722149
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,65536,16384,0.3758474667867025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,65536,16384,0.39551467895507814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,65536,12288,0.2761770566304525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,65536,16384,0.42033920288085935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,65536,12288,0.35249598821004235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,65536,16384,0.808463986714681
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,65536,12288,0.30265706380208335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,65536,10240,0.24783786137898764
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,65536,12288,0.6118879954020182
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,65536,10240,0.25318400065104163
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,65536,10240,0.29638398488362633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,65536,10240,0.48358720143636064
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,65536,8192,0.20115092595418296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,65536,8192,0.1920650641123454
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,65536,8192,0.23338987032572428
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,65536,8192,0.3806549390157064
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,65536,7168,0.1743061383565267
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,128,1536,0.00586346685886383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,65536,7168,0.17651839256286622
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,65536,7168,0.23520533243815103
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,65536,7168,0.33621333440144857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,65536,6144,0.14329066276550292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,65536,6144,0.18698347409566243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,65536,6144,0.28898986180623376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,65536,6144,0.15017280578613282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,65536,5120,0.12788693110148114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,65536,5120,0.24119466145833335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,65536,5120,0.168067200978597
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,65536,4096,0.09945706526438394
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,65536,4096,0.1349013328552246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,65536,4096,0.1957365353902181
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,65536,4096,0.1047818660736084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,65536,3584,0.08941866556803385
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,65536,3584,0.12400320370992024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,65536,3584,0.17094292640686035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,65536,3072,0.07716480096181234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,65536,3072,0.1126581350962321
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,65536,3072,0.07791039943695069
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,65536,3072,0.14876373608907062
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,65536,2560,0.12413866519927978
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,65536,2560,0.06615573167800903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,65536,2560,0.1000501314798991
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,65536,2048,0.054294399420420324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,65536,2048,0.08629759947458902
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,65536,2048,0.054723199208577475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,65536,2048,0.1008992036183675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,65536,1536,0.04272319873174031
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,65536,1536,0.07292906443277994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,65536,1536,0.077402663230896
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,65536,1536,0.04331946770350138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,65536,1024,0.054604800542195644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,65536,1024,0.06330986817677817
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,65536,1024,0.03125759959220886
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,65536,1024,0.03240426580111186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,65536,768,0.05721706549326579
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,65536,768,0.025603200991948443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,65536,768,0.042395734786987306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,65536,768,0.027371732393900554
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,65536,512,0.0482965350151062
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,65536,512,0.03388906717300415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,65536,512,0.017574399709701538
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,65536,256,0.010428800185521444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,65536,256,0.03551679849624634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,65536,256,0.012777599692344665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,65536,128,0.03667200009028117
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,65536,128,0.007959466675917308
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,65536,128,0.006619733572006225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,65536,128,0.011595732967058818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,65536,64,0.03529599905014038
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,65536,64,0.005751466751098633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,65536,64,0.011851732929547627
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,65536,32,0.0383072018623352
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,65536,32,0.005841066439946493
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,65536,32,0.012079999844233195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,16384,65536,0.460646406809489
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,65536,0.81604372660319
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,16384,16384,0.1297312021255493
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,16384,16384,0.10252799987792968
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,65536,0.43364906311035156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,16384,0.11143360137939454
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,16384,0.2004085381825765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,16384,12288,0.10803733666737873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,16384,12288,0.07937280337015787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,12288,0.1508394718170166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,12288,0.08707626660664877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,16384,10240,0.09469226996103922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,16384,10240,0.06668053468068441
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,10240,0.1317951997121175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,65536,5120,0.12111252943674725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,10240,0.07009387016296387
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,16384,8192,0.07895893255869547
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,16384,8192,0.05230400164922079
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,65536,3584,0.092576003074646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,8192,0.05678079922993978
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,8192,0.09552746613820394
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,16384,7168,0.09199466705322265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,16384,7168,0.04677439928054809
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,7168,0.08406399885813395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,7168,0.051423998673756924
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,16384,6144,0.08232426643371582
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,16384,6144,0.04127466678619385
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,65536,2560,0.07171733379364013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,6144,0.04481386741002401
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,6144,0.07330986658732096
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,16384,5120,0.07396907011667887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,5120,0.06179733276367187
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,5120,0.03926506837209066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,16384,4096,0.0294922669728597
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,16384,4096,0.06405866543451945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,4096,0.05061333179473877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,4096,0.03240533272425334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,16384,3584,0.06024320125579834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,3584,0.044540798664093016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,16384,3584,0.0267850657304128
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,3584,0.029234133164087933
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,16384,3072,0.023880533377329507
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,16384,3072,0.055947732925415036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,3072,0.03944106499354045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,3072,0.026405332485834758
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,16384,2560,0.05106026728947958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,16384,2560,0.02298133373260498
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,2560,0.03500586748123169
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,2560,0.01904639999071757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,65536,512,0.014794666568438211
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,65536,256,0.00988159974416097
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,16384,2048,0.04115840196609497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,16384,2048,0.01146346628665924
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,2048,0.028596266110738115
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,2048,0.01400213340918223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,16384,1536,0.034278400739034015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,16384,1536,0.009713066617647807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,1536,0.02474773327509562
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,1536,0.012337066729863485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,16384,1024,0.030833067496617635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,16384,1024,0.008137600123882293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,1024,0.008543999989827473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,1024,0.010153599580128988
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,16384,768,0.028756266832351683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,16384,768,0.007439999779065449
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,768,0.007326933244864146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,768,0.009341866771380106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,16384,512,0.026652799050013228
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,16384,512,0.00651093324025472
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,512,0.005916800101598104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,512,0.008289066453774769
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,16384,256,0.024737066030502318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,16384,256,0.005734399954477946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,128,0.004001066585381826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,256,0.004671999812126159
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,16384,65536,0.36494506200154625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,256,0.0077002664407094315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,16384,128,0.024734934171040855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,16384,128,0.0053951998551686605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,128,0.007412266731262207
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,16384,64,0.02448426683743795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,64,0.003798400113979975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,64,0.007495466868082683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,16384,32,0.02566186587015788
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,32,0.0038090666135152185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,32,0.007398400207360585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,12288,16384,0.0772266705830892
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,12288,65536,0.28227198918660484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,12288,16384,0.10862293243408203
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,65536,0.29754454294840493
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,16384,0.07926719983418783
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,16384,0.1506773312886556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,12288,12288,0.059290667374928795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,12288,12288,0.08891839981079101
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,12288,65536,0.3916757265726725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,65536,0.5769045511881511
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,12288,0.06212586561838786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,12288,10240,0.0503658652305603
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,12288,0.12362773418426513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,12288,10240,0.07876266638437907
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,10240,0.09500479698181152
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,10240,0.059537065029144284
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,12288,8192,0.07330239613850911
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,8192,0.04343786636988322
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,8192,0.07423146565755209
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,12288,7168,0.08220799763997397
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,12288,7168,0.042132266362508136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,7168,0.06742293039957682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,7168,0.03856000105539958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,12288,6144,0.07404053211212158
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,12288,6144,0.03291093309720357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,6144,0.05835946798324585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,6144,0.03430720170338948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,12288,5120,0.06663039922714234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,12288,5120,0.028380799293518066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,5120,0.04770666758219401
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,5120,0.029794132709503172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,12288,4096,0.059528533617655435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,12288,4096,0.023692800601323446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,4096,0.0408021330833435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,4096,0.024898133675257363
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,12288,3584,0.05508906841278076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,12288,3584,0.024194133281707764
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,3584,0.035947732130686444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,3584,0.02534613410631816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,12288,3072,0.049218134085337324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,12288,3072,0.013651200135548911
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,3072,0.03127893408139547
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,3072,0.01478506624698639
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,12288,2560,0.04301333427429199
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,2560,0.027154133717219038
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,2560,0.012524799505869547
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,12288,2048,0.0368618647257487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,2048,0.022267733017603555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,16384,5120,0.035306668281555174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,2048,0.01123199959595998
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,12288,1536,0.03310186664263408
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,12288,1536,0.008743466933568318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,1536,0.010338133573532105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,1536,0.01002239982287089
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,12288,1024,0.03022293249766032
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,1024,0.008754133184750875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,12288,1024,0.0073749333620071415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,12288,768,0.02826346755027771
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,1024,0.0076885332663853955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,12288,768,0.00697920024394989
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,768,0.006524799764156342
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,768,0.007923200229803721
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,12288,512,0.026336000363032026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,12288,512,0.006131199995676676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,512,0.005421866476535797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,12288,256,0.024530132611592613
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,512,0.007462400197982788
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,256,0.004243200023969015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,256,0.006926933427651723
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,12288,128,0.024076799551645912
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,128,0.0038922667503356934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,128,0.00662720004717509
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,12288,64,0.024549333254496257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,64,0.0035146666069825493
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,64,0.0066549330949783325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,12288,32,0.024682666858037314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,32,0.0035946667194366455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,32,0.0067456002036730455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,10240,65536,0.2356053352355957
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,10240,65536,0.3385973294576009
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,65536,0.48354345957438155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,65536,0.2527338663736979
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,10240,16384,0.07505813439687094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,10240,16384,0.10062932968139648
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,16384,0.12180373668670655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,16384,0.0689141352971395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,10240,12288,0.05103040138880412
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,10240,12288,0.08621973196665446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,12288,0.08982933362325032
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,12288,0.05439786513646444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,10240,10240,0.043611733118693034
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,10240,10240,0.07308800220489502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,10240,0.07790719668070475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,10240,0.04633173147837321
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,10240,8192,0.03573760191599528
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,10240,8192,0.06382506688435872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,8192,0.06152213414510092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,8192,0.03825706640879313
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,10240,7168,0.032315733035405476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,10240,7168,0.07908159891764323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,7168,0.055682134628295896
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,7168,0.034245332082112626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,10240,6144,0.07122986316680908
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,10240,6144,0.02829013268152873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,6144,0.0469653328259786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,6144,0.030524800221125286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,10240,5120,0.024791467189788818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,10240,5120,0.06424853404362997
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,5120,0.04034133354822795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,5120,0.026562132438023883
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,10240,4096,0.05718506574630737
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,4096,0.034378667672475174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,4096,0.020651733875274657
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,10240,3584,0.014019200205802917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,10240,3584,0.04951893488566081
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,12288,8192,0.04169066747029622
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,3584,0.030026666323343915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,3584,0.015244799852371215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,10240,3072,0.04285866816838582
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,10240,3072,0.011894399921099346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,3072,0.02581546703974406
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,3072,0.013146666685740152
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,10240,2560,0.03969599803288777
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,2560,0.022750933965047203
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,2560,0.011923199892044068
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,10240,2048,0.035684267679850265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,10240,2048,0.009360000491142273
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,2048,0.014500266313552857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,2048,0.0107232004404068
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,10240,1536,0.03282879988352458
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,10240,1536,0.008496000369389852
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,1536,0.00842133363087972
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,1536,0.009591466188430786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,10240,1024,0.03030186692873637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,10240,1024,0.007337599992752075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,1024,0.00660693347454071
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,1024,0.008583466211954754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,12288,2560,0.011155200004577637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,10240,768,0.028534400463104247
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,10240,768,0.006698666512966156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,12288,2048,0.009938133756319682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,768,0.005747200051943461
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,768,0.007901866734027863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,10240,512,0.02603093385696411
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,10240,512,0.006297599772612255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,512,0.004913066824277243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,512,0.007388799885908763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,10240,256,0.024613332748413087
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,10240,256,0.005497600138187409
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,256,0.00405973345041275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,256,0.006955733398596446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,10240,128,0.02433919906616211
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,128,0.0037258667250474296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,128,0.006690133114655812
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,10240,64,0.024881066878636678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,64,0.003453866640726725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,64,0.006760533154010773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,10240,32,0.024758400519688924
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,32,0.0035146666069825493
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,32,0.006673066814740498
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,8192,65536,0.18954133987426758
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,12288,256,0.005530666808287303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,65536,0.3822997411092123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,8192,65536,0.2699978510538737
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,12288,128,0.005363200108210246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,8192,16384,0.05312853256861368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,65536,0.2351914723714193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,8192,16384,0.08492159843444824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,16384,0.06380159854888916
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,16384,0.10912319819132488
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,8192,12288,0.04187306563059489
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,8192,12288,0.07044586340586344
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,12288,0.0717631975809733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,12288,0.04970453182856242
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,8192,10240,0.03561600049336751
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,8192,10240,0.06315626700719199
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,10240,0.06054826577504476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,8192,8192,0.029717334111531574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,10240,0.042574934164683026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,8192,0.03527573347091675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,8192,8192,0.055411199728647866
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,8192,7168,0.07047680219014486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,8192,0.04912319978078206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,7168,0.043661868572235106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,7168,0.03200746575991313
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,8192,6144,0.0636735995610555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,8192,6144,0.023817600806554158
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,6144,0.038230399290720626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,6144,0.028381866216659547
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,8192,5120,0.020475733280181884
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,8192,5120,0.05766719977060954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,5120,0.032331732908884685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,5120,0.022581332921981813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,8192,4096,0.04535786708196004
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,8192,4096,0.01251413325468699
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,4096,0.026837333043416338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,4096,0.014617600043614707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,8192,3584,0.041894400119781496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,8192,3584,0.011859200398127238
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,3584,0.023912533124287923
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,3584,0.013869866728782654
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,8192,3072,0.03826773166656494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,8192,3072,0.010609066486358643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,3072,0.021167999505996703
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,3072,0.012745599945386252
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,8192,2560,0.03535466591517131
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,8192,2560,0.009851732850074768
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,2560,0.016270933548609416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,2560,0.011703466375668842
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,8192,2048,0.03254826664924622
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,8192,2048,0.008521599570910136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,2048,0.00904746651649475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,2048,0.010407466689745586
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,8192,1536,0.02986026604970296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,8192,1536,0.007704533139864604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,10240,4096,0.020348799228668214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,1536,0.007899733384450276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,1536,0.00937493344148
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,8192,1024,0.02797226707140605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,8192,1024,0.00661653329928716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,1024,0.006155733267466227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,1024,0.008392533659934998
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,8192,768,0.025619200865427655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,8192,768,0.006431999802589417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,10240,2560,0.010664533575375874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,768,0.005594666798909505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,768,0.007922133306662242
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,8192,512,0.024233599503835045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,8192,512,0.005762133498986562
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,512,0.004713599880536398
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,512,0.007361066838105519
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,8192,256,0.02286613384882609
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,8192,256,0.0053258667389551794
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,256,0.003974399964014689
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,256,0.006849066913127899
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,8192,128,0.022295467058817544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,128,0.0036170666416486105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,128,0.006597333153088887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,8192,64,0.021921066443125407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,64,0.0034624000390370687
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,64,0.006705066561698914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,8192,32,0.023476266860961915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,32,0.00348693331082662
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,32,0.006681600213050842
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,7168,65536,0.17389334042867025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,65536,0.1796533266703288
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,7168,65536,0.24388906160990395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,65536,0.3601450602213542
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,7168,16384,0.08093226750691732
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,7168,16384,0.05287253459294637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,10240,128,0.005081599950790406
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,16384,0.0555295983950297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,16384,0.10255146821339925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,7168,12288,0.06732693513234457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,7168,12288,0.038661332925160725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,12288,0.04002773364384969
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,12288,0.06341653267542521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,7168,10240,0.05982079903284708
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,7168,10240,0.03302506605784099
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,10240,0.05333226521809896
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,10240,0.034721068541208905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,7168,8192,0.053041064739227296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,8192,0.04431573152542114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,8192,0.028410667181015016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,7168,7168,0.06895679632822672
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,7168,7168,0.024973867336908977
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,7168,0.025755733251571655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,7168,0.03945706685384114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,7168,6144,0.06235413153966268
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,7168,6144,0.021682133277257286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,6144,0.0344490647315979
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,7168,5120,0.014598400394121806
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,6144,0.02318613330523173
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,7168,5120,0.051208531856536864
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,5120,0.029355732599894206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,5120,0.014622933665911355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,7168,4096,0.04081919987996419
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,7168,4096,0.012328533331553142
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,4096,0.02426240046819051
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,4096,0.011988266309102377
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,7168,3584,0.039288532733917234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,7168,3584,0.011628799637158712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,3584,0.021628799041112264
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,3584,0.011435733238855997
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,7168,3072,0.036415998140970865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,3072,0.01720213294029236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,3072,0.010718933741251628
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,7168,2560,0.03495253324508667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,7168,2560,0.009428266684214275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,2560,0.009802666306495667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,8192,7168,0.026666667064030963
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,2560,0.009834667046864826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,7168,2048,0.03243946631749471
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,7168,2048,0.008413867155710856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,2048,0.008171733220418293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,2048,0.00893440047899882
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,1536,0.008148266871770223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,7168,1536,0.029848533868789672
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,7168,1536,0.0075103998184204105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,1536,0.007146666447321574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,7168,1024,0.02731093366940816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,7168,1024,0.006550399959087372
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,1024,0.0056320001681645715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,1024,0.007381333410739899
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,7168,768,0.025896533330281572
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,768,0.005015466610590617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,768,0.0069472000002861025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,7168,512,0.024718934297561647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,7168,512,0.005612800021966299
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,512,0.004337066908677419
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,512,0.006619733572006225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,7168,256,0.02334293325742086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,256,0.0062826668222745255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,256,0.003804799914360046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,7168,128,0.022921599944432578
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,7168,128,0.005055999755859375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,128,0.003502933432658514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,128,0.0060149331887563075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,7168,64,0.021572266022364298
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,64,0.006109866499900818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,64,0.0033546666304270422
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,7168,32,0.02200746734937032
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,32,0.003357866654793421
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,32,0.006065066655476888
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,6144,65536,0.2293013254801432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,65536,0.29385814666748045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,8192,128,0.005070933202902476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,6144,16384,0.07503039836883545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,65536,0.16775573094685872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,6144,16384,0.043032534917195636
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,16384,0.08279253641764323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,16384,0.045474131902058915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,6144,12288,0.06284906864166259
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,6144,12288,0.03394560019175212
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,12288,0.036246399084726974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,12288,0.0583349347114563
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,6144,10240,0.05602773427963257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,6144,10240,0.02895359992980957
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,10240,0.03076053261756897
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,10240,0.046945067246754964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,6144,8192,0.05070720116297404
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,8192,0.026027733087539674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,8192,0.04038933515548706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,6144,7168,0.06740372975667318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,6144,7168,0.021744000911712646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,7168,0.03532266616821289
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,7168,0.02376853426297506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,6144,6144,0.015131733814875283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,6144,6144,0.05757759809494019
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,7168,8192,0.027635200818379717
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,6144,0.031325866778691605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,6144,0.015518933534622192
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,6144,5120,0.04854613145192464
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,6144,5120,0.01332586705684662
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,5120,0.0267850657304128
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,5120,0.013727999726931252
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,6144,4096,0.04194773435592651
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,6144,4096,0.011285332838694255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,4096,0.011801600456237793
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,4096,0.022377600272496544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,6144,3584,0.03932906786600749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,6144,3584,0.010898133118947346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,3584,0.019966934124628702
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,3584,0.011040000120798747
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,6144,3072,0.03619946638743083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,3072,0.011095466216405232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,3072,0.010296533505121868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,6144,2560,0.03453760147094727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,6144,2560,0.009046399593353271
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,2560,0.009141332904497783
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,2560,0.009531733393669129
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,7168,3072,0.010605866710344952
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,6144,2048,0.03195733428001404
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,6144,2048,0.008134399851163227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,2048,0.008008533219496409
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,2048,0.008644266923268636
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,6144,1536,0.029411200682322187
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,1536,0.006913066903750102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,1536,0.008037333190441132
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,6144,1024,0.027296000719070436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,6144,1024,0.006386133531729381
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,1024,0.005468800167242686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,1024,0.007204266885916392
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,6144,768,0.025999999046325682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,6144,768,0.006178133189678192
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,768,0.004858666658401489
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,768,0.006830933193365733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,6144,512,0.024657066663106283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,6144,512,0.00555626650651296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,512,0.0042698666453361515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,512,0.006493866443634033
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,6144,256,0.023118933041890465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,6144,256,0.005160533388455709
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,256,0.003681066632270813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,256,0.006155733267466227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,6144,128,0.0228000005086263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,6144,128,0.004972800115744273
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,128,0.003419733295838038
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,128,0.005910400052865346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,6144,64,0.02214826742808024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,64,0.0032405334214369455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,64,0.005990399916966756
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,7168,768,0.006217599908510844
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,6144,32,0.021574399868647256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,32,0.005936000247796377
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,32,0.0032810665667057036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,7168,256,0.0050911997755368555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,65536,0.2359605312347412
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,65536,0.13149759769439698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,5120,65536,0.19678293863932292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,5120,16384,0.03813973267873128
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,5120,16384,0.06851413249969482
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,16384,0.06698559919993083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,16384,0.038888533910115555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,5120,12288,0.05794453223546346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,5120,12288,0.02969493269920349
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,12288,0.04674559831619263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,12288,0.031022934118906657
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,5120,10240,0.052183465162913
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,10240,0.040712531407674155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,10240,0.02723306616147359
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,5120,8192,0.04714346726735433
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,5120,8192,0.021255467335383096
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,8192,0.03321066697438558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,8192,0.022292266289393105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,6144,65536,0.14777493476867676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,5120,7168,0.06022400061289469
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,5120,7168,0.017303466796875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,7168,0.029369600613911945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,7168,0.014881066481272378
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,5120,6144,0.051826135317484534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,6144,0.026011733214060466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,6144,0.01323306659857432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,5120,5120,0.04557013511657715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,5120,5120,0.012758400042851767
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,5120,0.022206934293111165
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,5120,0.012160000205039979
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,5120,4096,0.040378665924072264
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,5120,4096,0.010802132884661357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,4096,0.014698666334152222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,4096,0.01060479978720347
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,5120,3584,0.03915093342463176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,5120,3584,0.010361599922180175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,3584,0.0100000003973643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,3584,0.010012800494829815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,6144,8192,0.024254933993021647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,5120,3072,0.03534613450368245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,5120,3072,0.009544533491134644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,3072,0.008849066495895386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,3072,0.009398399790128072
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,5120,2560,0.034227200349171955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,5120,2560,0.008914132912953694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,2560,0.007901866734027863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,2560,0.008900266885757447
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,5120,2048,0.030918399492899578
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,5120,2048,0.007843199868996937
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,2048,0.006917333106199901
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,2048,0.008072533210118612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,5120,1536,0.028962133328119914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,5120,1536,0.007110400001207988
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,1536,0.006046933432420095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,1536,0.00749120016892751
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,5120,1024,0.027413332462310792
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,5120,1024,0.006276266773541768
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,1024,0.004969599843025208
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,1024,0.006963199873765309
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,5120,768,0.025521065791447955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,5120,768,0.0060149331887563075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,768,0.004557866851488749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,768,0.0066538666685422255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,5120,512,0.024074665705362954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,512,0.004063999901215235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,5120,512,0.005543466905752818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,512,0.006557866434256236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,6144,3072,0.010030933221181234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,5120,256,0.02285226583480835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,5120,256,0.005006933212280273
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,256,0.003603200117746989
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,256,0.006111999849478403
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,5120,128,0.022132267554601036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,5120,128,0.004805333415667216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,5120,64,0.02136960029602051
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,6144,1536,0.0074538667996724445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,64,0.00591893345117569
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,128,0.003382399926582972
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,128,0.005932799975077311
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,64,0.003239466746648153
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,5120,32,0.021691733598709108
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,32,0.003257599969704946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,32,0.005929600199063619
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,4096,65536,0.11461973190307617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,4096,65536,0.16961386998494465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,65536,0.11021119753519695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,4096,16384,0.06297279993693033
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,4096,16384,0.031883732477823896
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,65536,0.19277653694152833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,16384,0.03317866722742717
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,16384,0.05304213364919027
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,4096,12288,0.02513706684112549
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,4096,12288,0.053641601403554284
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,12288,0.027294933795928955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,4096,10240,0.021577600638071695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,12288,0.04185493389765422
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,4096,10240,0.048826666673024495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,10240,0.03596906661987305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,10240,0.023360000054041544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,4096,8192,0.04012586673100789
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,4096,8192,0.016028799613316855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,8192,0.03004053235054016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,8192,0.013359999656677246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,4096,7168,0.053812265396118164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,4096,7168,0.014139733711878457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,7168,0.02723520000775655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,7168,0.011962667107582092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,5120,65536,0.13114453156789144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,4096,6144,0.04748266537984212
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,4096,6144,0.013058132926623025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,6144,0.02177493373552958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,6144,0.011069867014884948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,4096,5120,0.04270293315251668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,5120,0.018170666694641114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,5120,0.010316800077756245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,4096,0.008506666620572407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,4096,4096,0.03903466860453288
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,5120,10240,0.025106134017308553
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,4096,0.00913279950618744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,4096,3584,0.036746664841969805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,4096,3584,0.009761066238085429
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,3584,0.00811520020167033
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,3584,0.008799999952316284
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,4096,3072,0.03474453290303548
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,4096,3072,0.0090421328941981
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,3072,0.007485866546630859
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,3072,0.008339200417200725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,4096,2560,0.032714666922887166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,4096,2560,0.008208000163237254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,2560,0.00680320014556249
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,2560,0.007894399762153625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,4096,2048,0.030471465984980267
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,4096,2048,0.007394133508205414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,2048,0.0058773333827654515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,2048,0.007506133119265239
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,5120,6144,0.01497706671555837
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,4096,1536,0.028727465867996217
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,4096,1536,0.006955733398596446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,1536,0.005318399767080943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,1536,0.007097599903742473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,4096,1024,0.026579199234644572
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,4096,1024,0.006109866499900818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,1024,0.004769066472848257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,1024,0.006657066444555919
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,4096,768,0.02534719904263814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,4096,768,0.005981866518656413
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,768,0.004270933568477631
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,768,0.006424533327420552
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,4096,512,0.023966934283574423
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,4096,512,0.005454933146635691
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,512,0.0038922667503356934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,512,0.006268799801667531
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,4096,256,0.02300800085067749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,256,0.003575466573238373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,256,0.005966933568318685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,4096,128,0.022267733017603555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,4096,128,0.0048991998036702475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,128,0.0032320000231266023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,128,0.005838933090368906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,4096,64,0.02137706677118937
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,64,0.003099733342727025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,64,0.005865600208441416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,4096,32,0.02254400054613749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,32,0.0032586666444937387
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,32,0.00591893345117569
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3584,65536,0.1017738660176595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3584,16384,0.060013866424560545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,65536,0.10064000288645428
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,65536,0.1753738721211751
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,16384,0.047490131855010984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3584,65536,0.1598570664723714
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,16384,0.030754133065541582
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3584,12288,0.02365013360977173
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3584,12288,0.05173226594924927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,12288,0.02503040035565694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,12288,0.03762666781743367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3584,10240,0.046529066562652585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3584,10240,0.017780266205469766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,10240,0.0160480002562205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,10240,0.03231893380482991
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3584,8192,0.034942932923634845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3584,8192,0.015366400281588236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,8192,0.027342933416366576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,8192,0.01259519954522451
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3584,7168,0.05217386484146118
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3584,7168,0.014006400108337402
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,7168,0.024951465924580894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,7168,0.011815466483434041
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3584,6144,0.046809601783752444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3584,6144,0.01309866706530253
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,6144,0.0196288009484609
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,6144,0.01104853351910909
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3584,5120,0.04239146709442139
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3584,5120,0.011379200220108032
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,5120,0.010930132865905762
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,5120,0.010238933563232421
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3584,4096,0.038303999106089275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3584,4096,0.009966933727264404
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,4096,0.008949333429336548
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,4096,0.009101866682370504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3584,3584,0.036397866408030194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3584,3584,0.009469866752624512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,3584,0.008297599852085114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,3584,0.008714666962623597
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3584,3072,0.034533333778381345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3584,3072,0.008714666962623597
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,3072,0.00767680009206136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,3072,0.008376533786455791
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3584,2560,0.03269120057423909
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3584,2560,0.00805866668621699
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,2560,0.006875733534495036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,2560,0.007952000200748443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3584,2048,0.030434133609135945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3584,2048,0.007175466914971669
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,2048,0.0060138667623202005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,2048,0.007480533421039581
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3584,1536,0.02866026759147644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3584,1536,0.006743466854095459
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,1536,0.005343999962011973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,1536,0.0070613334576288865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3584,1024,0.026600533723831178
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,1024,0.004642133414745331
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,1024,0.006684799989064534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3584,768,0.025321600834528606
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3584,768,0.005819733440876007
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,4096,5120,0.011921067039171855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,768,0.0041685332854588825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,768,0.006483200192451477
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,4096,4096,0.010114133358001709
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,512,0.0038890667259693147
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3584,512,0.023858133951822916
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,512,0.006247466802597046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3584,256,0.022784000635147093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,256,0.0034645333886146545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,256,0.00602346658706665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3584,128,0.02183893322944641
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3584,128,0.004818133513132731
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,128,0.0031925333042939507
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,128,0.005898666878541311
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3584,64,0.02113386591275533
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,64,0.0030965333183606463
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,64,0.00584853341182073
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3584,32,0.021555199225743612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,32,0.0030805334448814393
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,32,0.005904000004132589
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3072,65536,0.09498666922251384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3072,65536,0.1510250727335612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,65536,0.14724373817443848
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,65536,0.09169387022654216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3072,16384,0.02681386669476827
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3072,16384,0.057468799750010166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,16384,0.044088534514109295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,16384,0.028578132390975952
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3072,12288,0.020372267564137778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3072,12288,0.04930560191472371
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,12288,0.03421440124511719
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,12288,0.018630399306615194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3072,10240,0.04139519929885864
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3072,10240,0.015454933047294617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,10240,0.015178666512171427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,10240,0.02927253246307373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3072,8192,0.03304106593132019
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3072,8192,0.01474240024884542
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,8192,0.025473066171010333
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,8192,0.0125109334786733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3072,7168,0.05074453353881836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,7168,0.02209279934565226
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,7168,0.011767466862996418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,4096,256,0.005098666747411092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3072,6144,0.04694399833679199
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,6144,0.010964266459147136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,6144,0.014050133029619851
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3072,5120,0.043084800243377686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3072,5120,0.011249066392580668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,5120,0.012258133292198181
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,5120,0.010124799609184266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3072,4096,0.03846293290456136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3072,4096,0.009658666451772054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,4096,0.010542933146158855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,4096,0.009037866195042928
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3072,3584,0.03735466798146565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3072,3584,0.009339732925097148
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,3584,0.008265600105126699
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,3584,0.008743466933568318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3072,3072,0.03521173397699992
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3072,3072,0.008646399776140849
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,3072,0.007577600081761678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,3072,0.008357333143552144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3072,2560,0.03338346481323242
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3072,2560,0.00786240001519521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3584,16384,0.02903466622034709
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,2560,0.00674773355325063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,2560,0.007890133559703827
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3072,2048,0.03041706681251526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3072,2048,0.006996266543865204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,2048,0.007540266712506611
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,2048,0.005975466469923655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3072,1536,0.029306666056315107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,1536,0.005336533486843109
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,1536,0.0070592001080513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3072,1024,0.026375466585159303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3072,1024,0.00582826683918635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,1024,0.0046165332198143
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,1024,0.00662720004717509
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3072,768,0.024784000714619954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3072,768,0.0056981335083643595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,768,0.0041696002086003625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,768,0.006424533327420552
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3072,512,0.023838933308919272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3072,512,0.00525546669960022
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,512,0.0037920000652472176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,512,0.00617386649052302
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3072,256,0.022746666272481283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3072,256,0.004982399940490723
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,256,0.0034304000437259674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,256,0.005913599828879039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3072,128,0.02211946646372477
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3072,128,0.004678399860858917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,128,0.003239466746648153
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,128,0.005827199916044871
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3072,64,0.021269333362579346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,64,0.0030303999781608583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,64,0.005854933460553487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,3072,32,0.02142613331476847
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,32,0.0030741333961486817
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,32,0.005818666517734527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2560,65536,0.14302719434102376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2560,65536,0.08406293392181396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,65536,0.124836270014445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3584,1024,0.005861333509286245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,65536,0.08397226333618164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2560,16384,0.05569599866867066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2560,16384,0.023989333709081014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,16384,0.03566399812698364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,16384,0.024728532632191977
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3584,512,0.005358933409055074
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2560,12288,0.047356800238291426
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2560,12288,0.018029866615931193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,12288,0.01596799989541372
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,12288,0.028705066442489623
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2560,10240,0.038176000118255615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2560,10240,0.015384533007939658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,10240,0.025523199637730913
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2560,8192,0.01344319979349772
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,10240,0.014327466487884521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,8192,0.01952106753985087
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3584,256,0.004904533425966898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2560,8192,0.03302293419837952
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,8192,0.012503467003504434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2560,7168,0.050374400615692136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2560,7168,0.011930666367212931
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,7168,0.01275200049082438
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,7168,0.01179200013478597
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2560,6144,0.04614293177922567
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2560,6144,0.011754666765530903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,6144,0.011782399813334147
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,6144,0.01090239981810252
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2560,5120,0.04275840123494466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,5120,0.010548266768455505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2560,4096,0.038072534402211505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,5120,0.010124799609184266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2560,4096,0.009408000111579894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,4096,0.009170132875442504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2560,3584,0.03790293137232463
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,4096,0.009055999914805095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2560,3584,0.008878933389981587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,3584,0.008796800176302593
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,3584,0.008585600058237712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2560,3072,0.0343178669611613
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,3072,0.007738666733105977
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,3072,0.008251733581225077
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2560,2560,0.03286186655362447
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2560,2560,0.0077450667818387345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,2560,0.006670933465162914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,2560,0.007875200112660725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2560,2048,0.03017599980036418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3072,6144,0.012360533078511555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2560,2048,0.006775466601053874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,2048,0.005884799857934316
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3072,7168,0.01369493305683136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,2048,0.007388799885908763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2560,1536,0.02837120095888774
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2560,1536,0.006308266520500183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,1536,0.005301333467165629
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,1536,0.00705813318490982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,1024,0.006572799881299336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2560,1024,0.02640639940897624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,1024,0.004492799937725067
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2560,768,0.025198932488759356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,768,0.004134399940570196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,768,0.006449066599210103
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2560,512,0.02392746607462565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2560,512,0.005217066903909048
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,512,0.0038101332883040107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,512,0.006208000083764395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2560,256,0.022679466009140014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2560,256,0.004835199813048045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,256,0.0034677334129810332
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,256,0.005940266450246175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2560,128,0.022638932863871256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2560,128,0.004669866462548574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,128,0.0031989333530267083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,128,0.005832533538341522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2560,64,0.021215999126434328
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,64,0.0030261332790056865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,64,0.00581333339214325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3072,1536,0.006678399940331777
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2560,32,0.02147946755091349
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,32,0.0030741333961486817
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,32,0.005835733314355215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2048,65536,0.08231680393218994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,65536,0.10267840226491291
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2048,65536,0.13568426767985026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,65536,0.07583466370900473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2048,16384,0.05327359835306803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2048,16384,0.023203200101852416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,16384,0.030065067609151203
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,16384,0.02091946601867676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2048,12288,0.04238506555557251
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2048,12288,0.018003199497858682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,12288,0.024259199698766075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,12288,0.01576533317565918
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2048,10240,0.03567146857579549
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2048,10240,0.01552959978580475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,10240,0.019713066021601357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,10240,0.014199466506640116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2048,8192,0.03234986662864685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2048,8192,0.01336853305498759
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,8192,0.011832533280054729
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,8192,0.012388267119725545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2048,7168,0.05049706697463989
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,7168,0.012682666381200155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,7168,0.011544533570607503
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2048,6144,0.046401067574818926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,6144,0.010196266571680705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,6144,0.010826667149861652
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2048,5120,0.042181332906087235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2048,5120,0.010403199990590414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,5120,0.009379200140635173
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,5120,0.010027733445167542
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2048,4096,0.038099201520284016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2048,4096,0.00915839970111847
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,4096,0.008417066931724549
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,4096,0.008945066730181377
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2048,3584,0.03661866585413615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2048,3584,0.008671999971071879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,3584,0.008145066599051159
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,3584,0.008515200018882752
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2048,3072,0.03445546627044678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2048,3072,0.00803306649128596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,3072,0.008269866804281871
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,3072,0.008101333181063335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2048,2560,0.03243520061175029
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2048,2560,0.007595733304818471
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,2560,0.006523733337720235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,2560,0.0077567999561627705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2048,2048,0.030029867092768354
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2048,2048,0.006858666737874349
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,2048,0.005843199789524078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,2048,0.007241599758466085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2048,1536,0.028420267502466838
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2048,1536,0.006364800035953522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,1536,0.005138133466243744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,1536,0.0068810666600863145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2048,1024,0.02653653422991435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2048,1024,0.005774933099746704
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,1024,0.004461866617202759
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,1024,0.006534400085608165
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2048,768,0.025067732731501265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2048,768,0.005585066477457682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,768,0.004057600100835165
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,768,0.006302933394908905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2560,5120,0.010384000341097514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2048,512,0.023865600426991783
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2048,512,0.005166933437188466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,512,0.0036992001036802924
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,512,0.0061247999469439185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2048,256,0.02295680046081543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2048,256,0.004866133133570353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,256,0.003389866650104523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,256,0.005906133353710175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2048,128,0.02195626695950826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2048,128,0.004834133386611939
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,128,0.0031541332602500914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,128,0.005688533186912537
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2048,64,0.021206400791803994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2560,3072,0.00823466678460439
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,64,0.003018666555484136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,64,0.005712000032265982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,2048,32,0.021460266908009847
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,32,0.002974933385848999
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,32,0.00576853354771932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1536,65536,0.12968426545461018
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,65536,0.08060159683227539
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,65536,0.07118079662322999
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1536,16384,0.04964053233464559
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1536,16384,0.020284799734751384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,16384,0.024423466126124064
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,16384,0.019383466243743895
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1536,12288,0.040055465698242185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1536,12288,0.016238933801651
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,12288,0.0139957328637441
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2560,1024,0.005678933362166087
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,12288,0.01584106683731079
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1536,10240,0.03549013137817383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1536,10240,0.014255999525388082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,10240,0.01241919994354248
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,10240,0.014254933595657349
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2560,768,0.005577600002288819
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1536,8192,0.0324181338151296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1536,8192,0.012356266379356384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,8192,0.01030293305714925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,8192,0.012400000294049581
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1536,7168,0.011412266890207927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1536,7168,0.05025813182195028
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,7168,0.009920000036557516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,7168,0.011582932869593303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1536,6144,0.046494932969411214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,6144,0.009173333644866943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,6144,0.01074666678905487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1536,5120,0.04267946481704712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,5120,0.00862506628036499
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,5120,0.010081066687901815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1536,4096,0.03806399901707967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1536,4096,0.00846506655216217
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,4096,0.008568533261617025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,4096,0.008958933750788371
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1536,3584,0.036611199378967285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,3584,0.007859200239181519
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,3584,0.008513066172599792
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1536,3072,0.03445333242416382
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1536,3072,0.007492266595363617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,3072,0.0071946665644645694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,3072,0.0081535999973615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1536,2560,0.03232746720314026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1536,2560,0.007134933272997539
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,2560,0.006483200192451477
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,2560,0.007751466830571492
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1536,2048,0.030096000432968138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1536,2048,0.006614399949709575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,2048,0.005786666770776113
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,2048,0.007268266876538594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1536,1536,0.028305067618687944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1536,1536,0.006150400141874949
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,1536,0.005072000126043955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,1536,0.006896000107129414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2048,7168,0.012556800246238708
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1536,1024,0.026205867528915405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1536,1024,0.005640533566474914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,1024,0.0044159998496373495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2048,6144,0.011394133170445759
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,1024,0.006472533444563548
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1536,768,0.025473066171010333
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1536,768,0.005486933390299479
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,768,0.004099199920892716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,768,0.006287999947865804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1536,512,0.024131200710932412
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1536,512,0.005134933193524678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,512,0.0037610667447249093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,512,0.006084266801675161
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1536,256,0.02289386590321859
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1536,256,0.004824533561865489
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,256,0.003320533285538355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,256,0.005873066683610281
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1536,128,0.022637865940729775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1536,128,0.004664533336957296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,128,0.0030975999931494398
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,128,0.005747200051943461
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1536,64,0.022371200720469157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,64,0.003009066730737686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,64,0.005756799876689911
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1536,32,0.021401600042978922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,32,0.003020799905061722
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,32,0.0058218667904535925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1024,65536,0.07018453280131022
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1024,65536,0.12469120025634765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,65536,0.05768746534983317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,65536,0.0677290678024292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1024,16384,0.045721598466237384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1024,16384,0.01988053321838379
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,16384,0.011918933192888895
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,16384,0.01919999917348226
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1024,12288,0.03787200053532918
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1024,12288,0.016082132856051125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,12288,0.010269866387049357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,12288,0.015802666544914246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1024,10240,0.034728535016377765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,10240,0.014180266857147216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,10240,0.009499733646710713
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1024,8192,0.03213333288828532
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1024,8192,0.01223360002040863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,8192,0.008409600456555684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,8192,0.012319999933242797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1024,7168,0.05035946766535441
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1024,7168,0.011385599772135418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,7168,0.00804906686147054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,7168,0.011548800269762675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1024,6144,0.04660906791687012
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,6144,0.007469866673151653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,6144,0.010769066214561463
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1024,5120,0.042182401816050215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1024,5120,0.009284266829490661
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,5120,0.010011733571688334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,5120,0.009909333785374959
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1024,4096,0.03860053221384684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1024,4096,0.008381866415341695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,4096,0.008355200290679932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,4096,0.008929066856702169
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1024,3584,0.036924799283345536
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1024,3584,0.007978666822115581
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,3584,0.008098133405049642
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,3584,0.008533333738644917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1024,3072,0.03504960139592488
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1024,3072,0.007437866429487865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,3072,0.007906133433183034
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,3072,0.008088533580303193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1024,2560,0.033164799213409424
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1024,2560,0.006853333115577698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,2560,0.006431999802589417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,2560,0.007751466830571492
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1024,2048,0.030477867523829145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1536,65536,0.0729365348815918
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1024,2048,0.006431999802589417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,2048,0.0058112000425656635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,2048,0.00727040022611618
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1024,1536,0.028413865963617963
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1024,1536,0.006064000229040781
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,1536,0.005106133222579956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,1536,0.006868266562620799
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1024,1024,0.026265599330266315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,1024,0.004420266548792521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,1024,0.0064522668719291685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1024,768,0.024804266293843587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1536,6144,0.010857599973678588
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,768,0.00401706670721372
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,768,0.006281599899133046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1536,5120,0.009506133198738099
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1024,512,0.0237610658009847
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,512,0.0036864000062147772
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,512,0.006066133578618368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1536,3584,0.008055466910203297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1024,256,0.022217599550882976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,256,0.003319466610749563
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,256,0.0058783998092015585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1024,128,0.021861332654953002
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1024,128,0.004603733122348785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,128,0.003092266619205475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,128,0.005748266478379568
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1024,64,0.021193599700927733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,64,0.00296426663796107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,64,0.005750399827957153
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,1024,32,0.02140053311983744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,32,0.002994133283694585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,32,0.005734399954477946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,768,65536,0.06992320219675699
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,768,65536,0.12230719725290935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,65536,0.046397864818573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,65536,0.06703253587086996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,768,16384,0.04401493469874064
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,16384,0.010542933146158855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,16384,0.019256534179051717
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,768,12288,0.03723413149515788
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,768,12288,0.01608746647834778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,12288,0.00947093367576599
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,12288,0.015871999661127727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,768,10240,0.03472426732381185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,768,10240,0.014057599504788718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,10240,0.008721066514650981
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,10240,0.014245333274205527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,768,8192,0.03237119913101196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,768,8192,0.012124799688657125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,8192,0.008059733112653096
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,8192,0.012359467148780823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,768,7168,0.050596264998118076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,768,7168,0.011221333344777425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,7168,0.009975467125574749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,7168,0.011635200182596842
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,768,6144,0.04604159990946452
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,768,6144,0.01030613382657369
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,6144,0.00928000013033549
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,6144,0.010807466506958009
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,768,5120,0.04235946734746297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,768,5120,0.00925439993540446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,5120,0.008636800448099773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,5120,0.010044800241788228
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,768,4096,0.03814613421758016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,768,4096,0.008327466746171314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,4096,0.007880533238252004
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1024,10240,0.01397119959195455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,4096,0.008930133779843648
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,768,3584,0.0360213319460551
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,768,3584,0.007852800190448761
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,3584,0.0075989335775375364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,3584,0.008514133095741273
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,768,3072,0.03413013219833374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,768,3072,0.00739519993464152
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,3072,0.007153066496054332
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,3072,0.008171733220418293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1024,6144,0.010272000233332317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,768,2560,0.032330665985743204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,768,2560,0.006876799960931141
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,2560,0.007738666733105977
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,2560,0.006325333317120869
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,768,2048,0.029908267656962077
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,768,2048,0.006390400230884552
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,2048,0.005619200070699056
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,2048,0.007293866574764251
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,768,1536,0.028321067492167156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,1536,0.004985600213209788
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,1536,0.006955733398596446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,768,1024,0.026190932591756182
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,768,1024,0.005523199836413065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,1024,0.004309333364168803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,1024,0.006492800017197926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,768,768,0.025094399849573772
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,768,768,0.005297066768010458
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,768,0.003990400085846583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,768,0.006365866462389629
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,768,512,0.02380160093307495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,768,512,0.005009066561857859
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,512,0.00365226666132609
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,512,0.0060810665289560955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,768,256,0.02234986623128255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,768,256,0.004677333434422811
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,256,0.0033557333052158357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,256,0.0059125334024429325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,768,128,0.021946666638056438
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,768,128,0.004517333209514618
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,128,0.0030975999931494398
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,128,0.0057888001203536986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,768,64,0.02121386726697286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,64,0.002930133293072383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,64,0.005780266722043356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,768,32,0.02144213318824768
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,32,0.002979200085004171
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,32,0.005769599974155426
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,512,65536,0.11991999944051106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,512,65536,0.06981759866078695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1024,1024,0.005530666808287303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,65536,0.03479359944661458
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,65536,0.06514666477839151
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,512,16384,0.04363733530044556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,512,16384,0.019783467054367065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,16384,0.008628267049789428
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,16384,0.019269333283106486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1024,768,0.005369600156943003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,512,12288,0.01614400049050649
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,512,12288,0.03734079996744792
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,12288,0.010548266768455505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,512,10240,0.034485332171122235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,12288,0.01593066652615865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,512,10240,0.014078933000564575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,10240,0.009699199597040813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,10240,0.014106667041778565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,512,8192,0.03226666649182637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,512,8192,0.012172800302505494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,8192,0.00886400043964386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,8192,0.01229866643746694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,7168,0.008413867155710856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1024,512,0.005019733309745788
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,512,7168,0.0505184014638265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,512,7168,0.01125866671403249
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1024,256,0.004855466882387797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,7168,0.011507200201352437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,512,6144,0.0465173323949178
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,512,5120,0.04216853380203247
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,512,6144,0.010281599561373393
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,6144,0.007968000074227651
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,6144,0.010759466886520385
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,512,5120,0.009237333138783773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,5120,0.007480533421039581
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,4096,0.008831999699274699
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,5120,0.00997440020243327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,512,4096,0.038098132610321044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,512,4096,0.008322133123874665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,4096,0.007100800176461537
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,512,3584,0.037793068091074626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,512,3584,0.007851733267307282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,3584,0.00747626672188441
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,3584,0.008534399668375652
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,512,3072,0.03414933284123738
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,512,3072,0.007349333167076111
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,3072,0.007089066505432129
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,2560,0.007771733403205872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,3072,0.00803306649128596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,512,2560,0.03248959978421529
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,512,2560,0.006826666494210561
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,2560,0.006321066617965698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,512,2048,0.030689066648483275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,512,2048,0.006409599880377452
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,2048,0.0055754666527112326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,2048,0.0072543998559316
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,768,16384,0.019684267044067384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,512,1536,0.028064000606536865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,1536,0.004970666766166687
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,1024,0.0064735998709996535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,1536,0.006922666728496551
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,512,1024,0.026268800099690754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,512,1024,0.005386666456858317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,1024,0.004244266450405121
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,512,768,0.02541866699854533
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,512,768,0.005211733281612396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,768,0.003934933245182038
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,768,0.006233599781990051
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,512,512,0.02382826606432597
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,512,512,0.004931200047334035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,512,0.0035978667438030243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,512,0.006051200131575266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,512,256,0.0226197342077891
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,512,256,0.0046858668327331545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,256,0.0032597333192825317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,256,0.005845333139101664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,512,128,0.022613332668940226
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,512,128,0.0044170667727788295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,64,0.005760000149408976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,128,0.0030559999247392017
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,128,0.005710933109124502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,512,64,0.021192532777786256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,64,0.0028757333755493166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,512,32,0.02239146629969279
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,32,0.0029002666473388673
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,32,0.0056874667604764305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,256,65536,0.11765013535817463
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,256,65536,0.0681600014368693
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,65536,0.020536533991495767
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,65536,0.0622101346651713
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,256,16384,0.042087467511494954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,16384,0.00909546713034312
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,16384,0.019243733088175455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,256,12288,0.03719360033671061
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,256,12288,0.016158933440844216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,12288,0.008133333424727123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,12288,0.015736533204714458
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,256,10240,0.034652801354726155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,10240,0.007732266684373219
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,10240,0.014151466886202493
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,256,8192,0.032280532519022624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,8192,0.0073173334201176955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,8192,0.01230400005976359
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,256,7168,0.05064746538798014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,256,7168,0.01120853324731191
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,7168,0.0071370666225751235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,768,1536,0.00619946668545405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,7168,0.011480533083279927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,256,6144,0.04633920192718506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,256,6144,0.010193066795667012
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,6144,0.006925866504510244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,6144,0.01074133316675822
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,256,5120,0.042184531688690186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,256,5120,0.00928000013033549
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,5120,0.009921066959698995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,5120,0.00734506646792094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,256,4096,0.037828266620635986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,4096,0.006950399776299794
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,4096,0.008850133419036866
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,256,3584,0.007805866499741872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,256,3584,0.0371018648147583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,3584,0.00745066652695338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,3584,0.008507733543713886
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,256,3072,0.03422826528549194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,256,3072,0.007357866565386455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,3072,0.007106133302052816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,3072,0.008099199831485748
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,256,2560,0.006841599941253662
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,256,2560,0.032230399052302044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,2560,0.0062943999965985615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,2560,0.00765119989713033
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,256,2048,0.02977919975916545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,256,2048,0.006366933385531108
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,2048,0.007257600128650665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,2048,0.005566933254400889
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,256,1536,0.028568534056345622
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,256,1536,0.005940266450246175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,1536,0.004875733455022176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,256,1024,0.0260917325814565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,1536,0.006888533135255177
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,256,1024,0.005474133292833964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,1024,0.0042303999265035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,1024,0.006438399851322174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,256,768,0.0249781330426534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,256,768,0.0052597333987553915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,768,0.0038602667550245917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,768,0.006257066627343495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,256,512,0.023564799626668295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,256,512,0.00498879998922348
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,512,0.0035648000737031303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,512,0.006037333110968272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,256,256,0.02294399936993917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,256,256,0.004634666442871094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,256,0.003201066702604294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,256,0.006017066538333893
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,256,128,0.022027732928593953
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,256,128,0.004413866500059763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,128,0.003047466774781545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,128,0.00572266678015391
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,256,64,0.02105600039164225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,64,0.0028768000503381092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,64,0.0056991999348004665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,256,32,0.021246933937072755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,32,0.002890666574239731
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,32,0.0056981335083643595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,128,65536,0.11635626951853435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,128,65536,0.0682207981745402
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,65536,0.013408000270525614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,65536,0.06029226779937744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,128,16384,0.04215146700541179
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,128,16384,0.019744000832239785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,16384,0.019126399358113607
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,16384,0.007161599894364675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,128,12288,0.037095467249552413
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,128,12288,0.016221867005030314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,12288,0.007097599903742473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,12288,0.015717333555221556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,128,10240,0.03461759885152181
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,128,10240,0.014043733477592468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,10240,0.007247999807198842
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,10240,0.014094932874043783
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,128,8192,0.03224853277206421
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,128,8192,0.012162133057912191
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,8192,0.00716480016708374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,8192,0.012277332941691081
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,128,7168,0.05041386683781942
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,128,7168,0.011152000228563944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,7168,0.007077333331108093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,7168,0.011529599626859028
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,128,6144,0.04577066500981648
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,6144,0.0068917334079742435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,6144,0.010645332932472228
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,128,5120,0.042420268058776855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,128,5120,0.009178666273752849
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,5120,0.007239466905593872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,5120,0.009869866569836934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,128,4096,0.03785813252131144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,128,4096,0.008227199812730153
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,4096,0.006836266815662384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,4096,0.008771199981371562
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,128,3584,0.03605973323186239
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,128,3584,0.0077237332860628765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,3584,0.0072970668474833175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,3584,0.008398933211962382
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,128,3072,0.03413973251978557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,128,3072,0.007245866457621257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,3072,0.007003733515739441
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,512,1536,0.006006399790445963
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,3072,0.008098133405049642
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,128,2560,0.03212160070737203
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,128,2560,0.006729599833488464
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,2560,0.006265600025653839
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,2560,0.007672533392906189
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,128,2048,0.00631573349237442
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,2048,0.005576533575852713
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,128,2048,0.030580266316731768
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,2048,0.007236266632874806
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,128,1536,0.028164267539978027
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,128,1536,0.005955199897289276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,1536,0.004890666902065277
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,1536,0.006829866766929626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,128,1024,0.026590933402379353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,1024,0.004261333247025808
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,1024,0.006438399851322174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,128,768,0.024921600023905435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,128,768,0.005186133086681366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,768,0.0038954667747020722
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,768,0.006323199967543285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,128,512,0.02363626758257548
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,128,512,0.004917333523432413
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,512,0.0036085332433382668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,512,0.006074666480223338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,128,256,0.02248426675796509
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,128,256,0.004690133531888326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,256,0.0032255999743938447
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,256,0.005987200140953064
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,128,128,0.0220085342725118
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,128,0.003048533449570338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,128,0.005727999905745188
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,128,64,0.02111146648724874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,64,0.0029290666182835894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,64,0.005682133138179779
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,128,32,0.02132479945818583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,32,0.0029472000896930695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,32,0.005673600236574808
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,64,65536,0.11622933546702068
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,65536,0.01164479951063792
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,65536,0.059869865576426186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,64,16384,0.042231468359629314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,16384,0.00703359991312027
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,16384,0.01908479928970337
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,256,10240,0.014039466778437296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,64,12288,0.037215999762217206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,12288,0.006919466455777486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,12288,0.01564800043900808
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,64,10240,0.03461440006891887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,10240,0.0070783997575442
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,10240,0.014108799894650779
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,64,8192,0.03230293393135071
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,8192,0.0070592001080513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,8192,0.012268799543380737
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,256,16384,0.01996586720148722
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,64,7168,0.05045760075251261
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,7168,0.006913066903750102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,7168,0.011499733726183573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,6144,0.006797866523265838
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,64,6144,0.04623039960861206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,6144,0.010689066847165425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,64,5120,0.04240106741587321
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,5120,0.007111466427644093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,5120,0.009985066453615824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,64,4096,0.03786666790644328
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,4096,0.006666666766007741
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,4096,0.00885973374048869
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,256,4096,0.008239999910195668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,64,3584,0.03611093362172445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,3584,0.007096533477306366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,3584,0.008508800466855367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,64,3072,0.034192001819610594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,3072,0.006941866874694824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,3072,0.008070399860541026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,64,2560,0.03235413432121277
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,2560,0.0061749334136645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,256,8192,0.012078932921091716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,2560,0.007712000111738841
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,64,2048,0.030123732487360638
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,2048,0.005568000177542368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,2048,0.007169066866238911
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,64,1536,0.028381866216659547
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,1536,0.004816000163555145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,1536,0.006828799843788147
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,64,1024,0.026121600468953447
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,1024,0.004145066688458125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,1024,0.006400000055631001
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,64,768,0.024986666440963746
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,768,0.0038463999827702843
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,768,0.0062282666563987735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,64,512,0.023676800727844238
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,512,0.0035605333745479585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,512,0.0059349333246548975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,64,256,0.022551467021306358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,256,0.0031968000034491217
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,256,0.0058037335673968
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,64,128,0.02187946637471517
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,128,0.0029994666576385496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,128,0.005705599983533224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,64,64,0.02111146648724874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,64,0.0028490667541821797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,32,0.0029343999922275543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,64,0.005686399837334951
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,64,32,0.021230934063593547
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,65536,0.011451733112335206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,32,0.005726933479309082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,32,65536,0.11620373725891113
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,65536,0.05988479852676391
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,32,16384,0.04213333527247111
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,16384,0.006951466699441274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,16384,0.019045333067576088
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,32,12288,0.037256534894307455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,12288,0.006849066913127899
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,12288,0.015624533096949259
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,32,10240,0.034687999884287515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,10240,0.00698773314555486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,10240,0.014083199699719749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,32,8192,0.03229440053304036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,8192,0.00693333347638448
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,8192,0.012170666456222534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,32,7168,0.05035093228022257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,7168,0.006826666494210561
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,7168,0.011528533697128297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,32,6144,0.04591786861419678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,6144,0.006727466483910878
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,6144,0.010644267002741497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,32,5120,0.04244266748428345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,5120,0.0071263998746871945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,5120,0.009821866949399311
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,32,4096,0.037972267468770346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,4096,0.006579199930032094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,4096,0.008828799923261006
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,32,3584,0.036187732219696046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,3584,0.0070826664566993715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,3584,0.008433066805203756
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,32,3072,0.03444053332010905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,3072,0.006729599833488464
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,3072,0.008060800035794576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,32,2560,0.03231679995854696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,2560,0.006098133325576782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,2560,0.007692799965540569
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,32,2048,0.030114134152730305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,2048,0.005480533341566721
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,1536,0.006810666620731353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,2048,0.007201066613197327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,32,1536,0.028246400753657024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,1536,0.004825599988301595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,32,1024,0.026126933097839356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,1024,0.004142933338880539
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,1024,0.00644053320089976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,32,768,0.024895999828974405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,768,0.0038272000849246977
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,768,0.006277333199977875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,32,512,0.023613866170247397
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,512,0.0035018667578697203
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,512,0.005942399799823761
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,32,256,0.022554665803909302
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,256,0.00314026673634847
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,256,0.005763199925422668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,32,128,0.021977599461873373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,64,0.0027978666126728057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,128,0.0029834667841593427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,32,32,0.021242666244506835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,128,0.0056650668382644655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,48,32,64,0.02102400064468384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,64,0.005656533439954122
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,32,0.002809600035349528
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,32,0.005648000041643778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,128,6144,0.01018986701965332
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,65536,16384,0.3934816042582194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,65536,16384,0.38313493728637693
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,65536,16384,0.42928212483723954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,65536,12288,0.36031678517659504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,65536,16384,0.8095338821411133
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,65536,12288,0.27624212900797523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,65536,12288,0.6109141031901042
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,65536,12288,0.30581226348876955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,65536,10240,0.22907946904500326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,65536,10240,0.3053824106852213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,65536,10240,0.4759850819905599
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,65536,10240,0.24735040664672853
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,65536,8192,0.18764373461405437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,65536,8192,0.23241066932678223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,65536,8192,0.1960576057434082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,65536,7168,0.1685717264811198
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,65536,8192,0.38145599365234373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,65536,7168,0.2385418732961019
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,65536,7168,0.17343039512634278
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,65536,6144,0.14205652872721355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,65536,7168,0.36371841430664065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,65536,6144,0.19723092714945475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,65536,6144,0.1504650592803955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,65536,6144,0.2875594774881999
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,128,1024,0.005366399884223938
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,65536,5120,0.1701248009999593
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,65536,5120,0.12682346502939862
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,65536,5120,0.23909440040588378
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,65536,4096,0.14486613273620605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,65536,4096,0.1928991953531901
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,65536,4096,0.10294613043467205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,65536,3584,0.12392000357309978
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,65536,3584,0.08782080014546713
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,128,128,0.004459733267625173
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,65536,3584,0.17007360458374024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,65536,3584,0.09184106985727945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,65536,3072,0.07637226581573486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,65536,3072,0.11115626494089763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,65536,3072,0.07922133604685465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,65536,3072,0.14770240783691407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,65536,2560,0.0689909299214681
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,65536,2560,0.12308586438496907
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,65536,2560,0.09724160035451254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,65536,2048,0.08513813018798828
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,65536,2048,0.05461333195368449
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,65536,2048,0.09891520341237386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,65536,2048,0.056796801090240476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,65536,1536,0.07283626397450765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,65536,1536,0.07684586842854818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,65536,1536,0.04477226734161377
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,65536,1024,0.030562132596969604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,65536,1024,0.05874346494674683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,65536,1024,0.053496531645456945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,65536,1024,0.033160533507665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,65536,768,0.054295468330383304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,65536,768,0.024771199623743693
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,65536,768,0.041547731558481855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,65536,768,0.028544000784556073
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,65536,512,0.012885333100954691
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,65536,512,0.043095465501149496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,65536,512,0.029356799523035687
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,65536,256,0.033717334270477295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,65536,512,0.017849600315093993
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,65536,256,0.009578667084376017
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,65536,256,0.008880000313123066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,65536,256,0.013377066453297934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,65536,128,0.007895466685295106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,65536,128,0.031114667654037476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,65536,128,0.00622080018122991
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,65536,128,0.011596799890200297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,65536,64,0.03109546701113383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,65536,64,0.005220266679922739
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,65536,64,0.011588266491889954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,65536,32,0.03115839958190918
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,65536,32,0.0053269331653912864
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,65536,32,0.012133333086967468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,65536,0.4194026629130046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,65536,0.8042218526204428
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,16384,65536,0.495850658416748
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,16384,16384,0.1439242680867513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,16384,16384,0.10285546779632568
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,16384,0.11807893117268879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,16384,0.2030794620513916
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,16384,12288,0.11589866479237873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,12288,0.15413759549458822
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,12288,0.09493652979532877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,16384,10240,0.06309866507848104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,16384,10240,0.09990932941436767
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,10240,0.11816319624582927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,10240,0.07370879650115966
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,16384,8192,0.054679465293884275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,16384,8192,0.08274346987406413
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,8192,0.10281600157419842
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,8192,0.05861546595891317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,16384,7168,0.046114134788513186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,16384,7168,0.09813546339670817
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,7168,0.08418666521708171
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,7168,0.05341226657231649
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,16384,6144,0.08561813036600749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,16384,6144,0.04254719813664754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,6144,0.07357227007548015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,6144,0.0475658655166626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,16384,5120,0.07636053562164306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,16384,5120,0.034959999720255534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,65536,5120,0.12238293488820393
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,5120,0.061509335041046144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,5120,0.04131306807200114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,16384,4096,0.06656320095062256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,16384,4096,0.03039039969444275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,65536,4096,0.10149973233540852
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,4096,0.05028266509373983
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,4096,0.03339946667353312
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,16384,3584,0.026570665836334228
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,16384,3584,0.06226453383763632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,3584,0.029334400097529096
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,3584,0.047193598747253415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,16384,3072,0.057630932331085204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,65536,2560,0.07026879787445069
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,3072,0.03888426621754964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,3072,0.026661332448323565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,16384,2560,0.019703465700149535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,16384,2560,0.05082986752192179
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,2560,0.02090559999148051
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,2560,0.03290133277575175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,65536,1536,0.0428330659866333
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,16384,2048,0.03935253222783407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,16384,2048,0.012007466952006022
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,2048,0.02703253428141276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,2048,0.014469333489735923
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,16384,1536,0.03451519807179769
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,16384,1536,0.00995199978351593
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,1536,0.021397332350413002
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,1536,0.01290773351987203
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,16384,1024,0.031274666388829545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,16384,1024,0.008177066842714947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,1024,0.007995733122030894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,1024,0.010297600428263347
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,16384,768,0.02935466567675273
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,16384,768,0.007367466886838277
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,768,0.00684799998998642
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,768,0.009327999750773112
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,16384,512,0.027458133300145467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,16384,512,0.006607999900976817
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,512,0.005499733487764994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,512,0.008431999882062276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,16384,256,0.025766400496164958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,16384,256,0.005703466633955637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,256,0.0044170667727788295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,256,0.007621333499749501
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,16384,128,0.0250271995862325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,16384,128,0.005239466826121012
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,128,0.0038250667353471124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,128,0.007383466760317485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,16384,64,0.02382826606432597
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,64,0.003538133452335993
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,64,0.007399466633796692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,16384,32,0.023622399568557738
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,32,0.0037119999527931214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,32,0.007402666906515758
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,16384,65536,0.36523520151774086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,12288,65536,0.2806709289550781
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,12288,16384,0.10741546948750813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,65536,0.30448640187581383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,12288,16384,0.07650667031606039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,16384,0.08006186485290527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,12288,65536,0.3938048044840495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,16384,0.14805439313252766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,65536,0.5716341018676758
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,12288,12288,0.0591754674911499
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,12288,12288,0.09486613273620606
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,12288,0.11441067059834797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,12288,0.07092906634012858
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,12288,10240,0.07772053082784017
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,12288,10240,0.057188268502553305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,12288,8192,0.04198720057805379
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,10240,0.09552000363667806
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,10240,0.05398720105489095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,12288,8192,0.06751573085784912
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,16384,12288,0.07477866808573405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,8192,0.07799893220265706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,8192,0.043910400072733564
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,12288,7168,0.08292693297068278
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,12288,7168,0.03668373425801595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,7168,0.06819307009379069
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,7168,0.03925653298695882
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,12288,6144,0.07389439741770426
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,6144,0.057615999380747476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,6144,0.03497386773427327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,12288,5120,0.06574399868647257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,5120,0.04963839848836263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,5120,0.030808534224828082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,12288,4096,0.023532799879709878
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,12288,4096,0.057825068632761635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,12288,3584,0.05326293309529623
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,4096,0.04084479808807373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,4096,0.025040000677108765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,12288,3584,0.02165973385175069
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,3584,0.0347818652788798
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,3584,0.023042132457097374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,12288,3072,0.041493332386016844
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,3072,0.031118933359781904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,3072,0.015631999572118124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,12288,2560,0.037251198291778566
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,2560,0.02683839996655782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,2560,0.013533866405487061
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,12288,2048,0.033868801593780515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,2048,0.021887999773025513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,2048,0.011731200416882833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,12288,1536,0.031198932727177935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,12288,1536,0.009633066256841023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,1536,0.00988159974416097
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,1536,0.010426666339238484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,12288,1024,0.02839360038439433
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,12288,1024,0.007776000102361043
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,16384,3072,0.023661865790685018
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,1024,0.007344000041484833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,1024,0.00890880028406779
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,12288,768,0.026779733101526898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,12288,768,0.007006933291753133
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,768,0.006287999947865804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,768,0.008226133386294047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,12288,512,0.025294933716456098
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,12288,512,0.006231466432412466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,512,0.005190399785836538
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,512,0.0074442664782206226
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,12288,256,0.02377600073814392
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,12288,256,0.005426133175690969
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,256,0.004208000004291534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,256,0.006961066524187725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,12288,128,0.022911999622980753
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,12288,128,0.005010133484999338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,128,0.003722666700681051
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,128,0.0066336000959078475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,12288,64,0.02234026590983073
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,64,0.0033941333492596946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,64,0.006723199784755707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,12288,32,0.021398399273554483
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,32,0.0034634667138258614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,32,0.0068256000677744556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,10240,65536,0.23356159528096518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,65536,0.47968320846557616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,10240,65536,0.33235626220703124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,65536,0.2624373277028402
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,10240,16384,0.09295573234558105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,10240,16384,0.07332159678141276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,16384,0.0782975991566976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,16384,0.11942933400472004
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,10240,12288,0.07914346853892008
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,10240,12288,0.05050133466720581
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,12288,0.09092053572336832
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,12288,0.05481173197428385
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,10240,10240,0.06885226567586264
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,10240,10240,0.04278186559677124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,10240,0.0787712017695109
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,10240,0.047133866945902506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,10240,8192,0.035358933607737224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,10240,8192,0.06137173175811768
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,8192,0.06181439956029257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,8192,0.038729600111643475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,10240,7168,0.031854933500289916
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,10240,7168,0.07575146357218424
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,7168,0.0564138650894165
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,7168,0.034835199515024826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,10240,6144,0.02805546720822652
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,10240,6144,0.06872639656066895
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,6144,0.048338135083516434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,6144,0.03168320059776306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,10240,5120,0.024477867285410564
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,10240,5120,0.06144533157348633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,5120,0.04169600009918213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,5120,0.02709973255793254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,10240,4096,0.020950400829315187
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,10240,4096,0.052858666578928626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,4096,0.03291626572608948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,4096,0.018931200106938682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,10240,3584,0.044072532653808595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,10240,3584,0.014436266819636025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,3584,0.01591146687666575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,3584,0.03027733365694682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,10240,3072,0.03968213399251302
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,10240,3072,0.01241919994354248
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,3072,0.02654293378194173
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,12288,6144,0.032707200447718306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,10240,2560,0.01141973336537679
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,3072,0.013918933272361756
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,10240,2560,0.03607999881108602
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,2560,0.012433066964149475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,2560,0.02454293370246887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,10240,2048,0.03308266599973043
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,10240,2048,0.009965866804122925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,12288,5120,0.027966932455698652
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,2048,0.015158399939537048
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,2048,0.011158399780591329
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,10240,1536,0.030849067370096843
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,10240,1536,0.008680533369382222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,1536,0.007990399996439617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,1536,0.009870933492978413
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,10240,1024,0.02814720074335734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,10240,1024,0.007358933488527934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,1024,0.006296533346176148
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,768,0.005480533341566721
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,12288,2560,0.013292800386746725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,12288,3072,0.015864533185958863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,1024,0.008709333340326945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,10240,768,0.026702932516733807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,768,0.00793280005455017
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,10240,512,0.024754132827123007
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,10240,512,0.00628053347269694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,512,0.004771199822425842
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,512,0.007438933352629344
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,10240,256,0.02319999933242798
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,12288,2048,0.011226666967074077
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,10240,256,0.005505066613356272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,256,0.003885866701602936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,128,0.0036330667634805044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,256,0.00702400008837382
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,10240,128,0.022779732942581177
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,10240,128,0.004999466737111409
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,128,0.00664106657107671
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,10240,64,0.02161173423131307
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,64,0.0033333333830038703
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,64,0.006785066425800323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,10240,32,0.02231360077857971
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,32,0.0033226666351159418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,32,0.006794666747252147
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,8192,16384,0.05260800123214722
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,8192,16384,0.08289706707000732
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,65536,0.23528639475504556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,8192,65536,0.25214400291442873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,65536,0.3781237284342448
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,16384,0.06413866678873698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,16384,0.09364906946818033
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,8192,12288,0.07068266868591308
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,12288,0.05012266635894776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,12288,0.07895680268605551
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,8192,10240,0.035913598537445066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,8192,10240,0.06186666488647461
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,10240,0.06041066646575928
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,10240,0.04391359885533651
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,8192,8192,0.05519040028254191
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,8192,0.049370666344960526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,8192,0.033904000123341875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,7168,0.04331733385721843
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,8192,7168,0.07130133310953776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,7168,0.032630399862925215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,8192,6144,0.06422079801559448
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,8192,6144,0.023388799031575522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,6144,0.03828800121943156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,6144,0.028178133567174274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,8192,5120,0.019636267423629762
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,8192,5120,0.05723093350728353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,5120,0.032340266307195026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,5120,0.023005867004394533
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,8192,4096,0.013416533668835958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,8192,4096,0.04214293162027995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,4096,0.0268394668896993
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,4096,0.0160480002562205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,8192,3584,0.04038293361663818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,8192,3584,0.012665599584579468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,3584,0.02367573380470276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,3584,0.014915200074513755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,8192,3072,0.03623893260955811
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,8192,3072,0.011345066626866658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,3072,0.020946133136749267
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,3072,0.013470932841300964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,8192,2560,0.033852799733479814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,8192,2560,0.010283733407656353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,2560,0.016846932967503867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,2560,0.012299733360608418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,8192,2048,0.031948800881703696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,8192,2048,0.009086933732032777
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,2048,0.008565333485603333
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,2048,0.010935466488202412
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,8192,1536,0.029531733194986982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,8192,1536,0.007904000083605449
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,1536,0.007565866907437642
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,1536,0.009692800045013428
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,8192,1024,0.027038933833440144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,1024,0.005705599983533224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,8192,1024,0.00689279983441035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,1024,0.008378666639328004
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,8192,768,0.02577173312505086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,8192,768,0.00628053347269694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,768,0.007830399771531422
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,768,0.005179733534653981
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,8192,512,0.024737066030502318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,512,0.004289066791534424
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,512,0.007289599875609081
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,8192,256,0.022973867257436116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,8192,256,0.005288533369700114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,256,0.003833599885304769
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,256,0.006916266679763794
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,8192,128,0.02188160022099813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,8192,128,0.00496319979429245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,128,0.003483733286460241
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,128,0.0065184002121289565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,8192,64,0.02141973376274109
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,64,0.003252266595760981
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,64,0.0065760001540184024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,8192,32,0.022243199745814006
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,32,0.0033610666791598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,32,0.006733866532643636
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,7168,65536,0.23659733136494956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,7168,65536,0.17745386759440104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,10240,768,0.006796800096829732
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,65536,0.18374613126118977
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,65536,0.3483935991923014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,7168,16384,0.07554240226745605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,7168,16384,0.052919467290242515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,16384,0.08329813480377198
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,16384,0.05085759957631429
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,7168,12288,0.0639520009358724
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,7168,12288,0.038675200939178464
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,12288,0.06375679969787598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,12288,0.03955093224843343
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,7168,10240,0.0329802672068278
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,7168,10240,0.05762879848480225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,10240,0.05358613332112631
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,10240,0.03466453154881795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,7168,8192,0.0513322671254476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,8192,0.04416000048319499
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,8192,0.028885332743326823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,7168,7168,0.02476373314857483
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,7168,7168,0.06785173416137695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,7168,0.038593065738677976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,7168,0.026614399751027425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,8192,65536,0.19529493649800617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,7168,6144,0.021513599157333373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,7168,6144,0.059477333227793375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,6144,0.023092265923817953
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,6144,0.039445332686106366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,7168,5120,0.015876266360282897
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,7168,5120,0.046692268053690596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,5120,0.028964267174402876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,5120,0.016135467092196147
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,8192,12288,0.04147306680679321
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,7168,4096,0.040454399585723874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,7168,4096,0.013410133123397828
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,4096,0.024069333076477052
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,8192,8192,0.02940906683603923
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,4096,0.012622933586438498
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,7168,3584,0.037811199824015304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,3584,0.02148800094922384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,3584,0.011986133456230164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,8192,7168,0.02648319999376933
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,7168,3072,0.03619840145111084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,7168,3072,0.011229866743087768
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,3072,0.017802667617797852
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,2560,0.009176533420880635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,3072,0.01106346646944682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,7168,2560,0.033973332246144614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,2560,0.010389332969983418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,7168,2048,0.031108266115188597
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,7168,2048,0.008778666456540424
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,2048,0.007956266899903615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,2048,0.009367466966311137
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,7168,1536,0.02908693353335063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,7168,1536,0.007769600053628286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,1536,0.006720000008742015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,1536,0.008515200018882752
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,7168,1024,0.026742400725682576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,7168,1024,0.0065984000762303666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,1024,0.005351466437180838
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,1024,0.0075647999842961625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,768,0.004808533191680908
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,7168,768,0.025409066677093507
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,7168,768,0.00621013343334198
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,768,0.007153066496054332
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,7168,512,0.02467306653658549
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,7168,512,0.005748266478379568
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,512,0.004110933343569437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,512,0.00674773355325063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,7168,256,0.022825600703557332
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,7168,256,0.005126399795214335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,256,0.0036746665835380556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,256,0.006259199976921081
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,7168,128,0.02180373271306356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,7168,128,0.004750933249791463
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,128,0.003370666752258936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,128,0.006030933558940887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,7168,64,0.021359999974568687
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,64,0.003172266731659571
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,64,0.006106666723887126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,7168,32,0.021921066443125407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,32,0.0032000000278155005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,8192,512,0.005897599955399831
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,32,0.006016000111897787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,6144,65536,0.15984427134195964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,65536,0.1738848050435384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,6144,16384,0.07098560333251953
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,6144,65536,0.21171199480692543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,65536,0.2847050666809082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,6144,16384,0.043030401070912674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,16384,0.04806400140126546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,16384,0.0773845354715983
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,6144,12288,0.059860265254974364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,12288,0.05803840160369873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,6144,10240,0.05376000006993612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,12288,0.03543359835942586
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,6144,10240,0.029418667157491047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,10240,0.032798933982849124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,10240,0.04882560173670451
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,6144,8192,0.024170666933059692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,6144,8192,0.049038934707641604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,8192,0.04016853173573812
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,8192,0.025362133979797363
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,6144,7168,0.0645354668299357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,6144,7168,0.022051199277242025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,7168,0.03427733182907104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,7168,0.023124267657597862
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,6144,6144,0.054744533697764074
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,6144,6144,0.017084799210230508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,6144,0.031146667400995892
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,6144,0.01694719990094503
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,6144,5120,0.014193066954612732
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,6144,5120,0.0460970679918925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,5120,0.026104533672332765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,5120,0.01437226633230845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,6144,4096,0.04035840034484863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,6144,4096,0.012171733379364013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,4096,0.022190932432810465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,4096,0.011997866630554199
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,6144,3584,0.038097067674001055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,6144,3584,0.011736533045768738
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,3584,0.01955946683883667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,3584,0.011654399832089742
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,6144,3072,0.03619946638743083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,6144,3072,0.010385066270828247
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,3072,0.010496000448862713
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,3072,0.010706133643786113
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,6144,2560,0.03384960095087687
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,2560,0.008887466788291932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,2560,0.009656533598899841
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,6144,2048,0.03101973334948222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,6144,2048,0.008462933699289958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,2048,0.007733333110809326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,2048,0.008769067128499348
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,7168,8192,0.027190399169921876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,6144,1536,0.02908586661020915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,6144,1536,0.007704533139864604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,1536,0.00811839997768402
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,1536,0.006579199930032094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,6144,1024,0.027130667368570966
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,6144,1024,0.006526933113733928
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,1024,0.0052597333987553915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,1024,0.007223466535409291
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,768,0.004709333181381226
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,768,0.0068896000583966565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,6144,768,0.02580053408940633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,6144,768,0.006082133452097575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,6144,512,0.02377706567446391
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,6144,512,0.005659733215967814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,512,0.00402453343073527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,512,0.006515199939409892
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,256,0.003573333223660787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,6144,256,0.022755199670791627
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,6144,256,0.005239466826121012
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,6144,128,0.004837333162625631
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,256,0.006202666461467743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,6144,128,0.02246826688448588
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,128,0.0033439998825391137
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,128,0.005899733304977417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,64,0.005932799975077311
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,7168,3584,0.01242453356583913
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,6144,64,0.021364265680313112
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,64,0.0031456001102924346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,6144,32,0.02105813423792521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,32,0.003155199935038885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,32,0.006050133208433787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,7168,2560,0.009868799646695455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,5120,65536,0.13567147254943848
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,5120,16384,0.041179732481638594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,5120,65536,0.1789514700571696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,5120,16384,0.06502399841944376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,65536,0.1304255962371826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,16384,0.06269013484319051
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,65536,0.23908692995707193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,16384,0.0382912000020345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,5120,12288,0.03127359946568807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,5120,12288,0.055472000439961755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,12288,0.04676373402277629
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,12288,0.03107946713765462
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,5120,10240,0.027850667635599773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,5120,10240,0.050204801559448245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,10240,0.03975786765416463
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,10240,0.028943999608357744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,5120,8192,0.04552213350931804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,5120,8192,0.02241493264834086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,8192,0.0322432001431783
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,8192,0.022077866395314536
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,5120,7168,0.058007466793060306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,7168,0.02879040042559306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,7168,0.016509866714477538
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,6144,0.013804800311724343
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,5120,6144,0.015784533818562825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,5120,6144,0.05185386737187704
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,6144,0.025384533405303954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,5120,5120,0.01411733329296112
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,5120,5120,0.04477013349533081
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,5120,0.021939200162887574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,5120,0.012401066223780314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,5120,4096,0.03998080094655355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,5120,4096,0.011927466591199238
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,4096,0.0144896000623703
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,4096,0.010899200042088827
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,5120,3584,0.03788906733194987
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,3584,0.009771733482678732
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,3584,0.010293333729108175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,5120,3072,0.0359007994333903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,5120,3072,0.010188800096511842
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,3072,0.008611200253168742
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,3072,0.009645866354306538
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,5120,2560,0.03421119848887126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,5120,2560,0.009547733267148336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,2560,0.007667199770609538
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,2560,0.00923520028591156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,6144,12288,0.03420799970626831
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,2048,0.006638933221499126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,5120,2048,0.03102826674779256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,2048,0.008261333405971526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,5120,1536,0.029667200644810994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,5120,1536,0.007640533149242401
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,1536,0.005691733459631602
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,1536,0.00780266672372818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,5120,1024,0.02670186758041382
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,1024,0.004833066463470459
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,1024,0.006961066524187725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,5120,768,0.025711999336878462
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,5120,768,0.005927466849486033
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,768,0.0043381333351135256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,768,0.006663466493288676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,5120,512,0.024485333760579427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,5120,512,0.005571199953556061
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,512,0.003930666546026865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,512,0.006342400113741558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,5120,256,0.02300800085067749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,5120,256,0.005089066425959269
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,256,0.0035573333501815797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,256,0.006436266501744588
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,5120,128,0.022456532716751097
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,5120,128,0.0047647997736930845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,128,0.003328000009059906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,128,0.005881600081920624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,5120,64,0.022001065810521445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,64,0.003126399964094162
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,64,0.005931733548641205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,5120,32,0.020986666282018028
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,32,0.0031114667654037476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,32,0.005896533528963724
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,4096,65536,0.12900479634602863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,65536,0.19413119951883953
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,4096,65536,0.162226136525472
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,6144,2560,0.009682133793830872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,65536,0.11021866798400878
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,4096,16384,0.05933119853337606
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,4096,16384,0.03606293201446533
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,16384,0.03298880060513814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,4096,12288,0.051113601525624594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,16384,0.05215893189112345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,4096,12288,0.029241599639256793
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,12288,0.02691520055135091
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,12288,0.04157546758651733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,4096,10240,0.046860798199971514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,10240,0.03519999980926514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,10240,0.02293973366419474
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,4096,8192,0.0194815993309021
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,4096,8192,0.036773331960042316
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,8192,0.029626667499542236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,8192,0.013894400000572205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,4096,7168,0.016133333245913185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,4096,7168,0.05141119956970215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,7168,0.026504532496134443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,7168,0.01299626628557841
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,4096,6144,0.0455839991569519
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,4096,6144,0.014566399653752646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,6144,0.023830399910608927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,6144,0.011700266599655151
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,4096,5120,0.041816532611846924
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,4096,5120,0.012813867131868998
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,5120,0.016145066420237223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,5120,0.010718933741251628
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,4096,4096,0.03859519958496094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,4096,4096,0.01100373367468516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,4096,0.009609599908192951
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,4096,0.008686932921409606
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,4096,3584,0.03635840018590291
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,4096,3584,0.010445866982142131
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,3584,0.007729066908359528
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,3584,0.009150933225949604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,4096,3072,0.03434346516927083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,4096,3072,0.009512533744176228
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,3072,0.007159466544787089
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,3072,0.008593066533406576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,4096,2560,0.03265920082728068
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,2560,0.006862933437029521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,2560,0.008081066608428954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,4096,2048,0.030959999561309813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,4096,2048,0.007797333101431529
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,2048,0.006116266548633576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,2048,0.007493333518505096
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,4096,1536,0.029063467184702558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,4096,1536,0.007152000069618225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,1536,0.0053951998551686605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,1536,0.007063466807206471
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,5120,7168,0.01872106591860453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,4096,1024,0.02661866744359334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,1024,0.004503466685612996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,1024,0.006663466493288676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,4096,768,0.025383466482162477
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,4096,768,0.005835733314355215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,768,0.004181333382924398
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,768,0.006477866570154827
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,4096,512,0.02427519957224528
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,4096,512,0.005510400235652924
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,512,0.003769599894682566
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,5120,3584,0.011329066753387452
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,512,0.006253866851329804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,4096,256,0.022645332415898643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,4096,256,0.0050911997755368555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,256,0.0034602666894594826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,256,0.006017066538333893
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,4096,128,0.021568000316619873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,4096,128,0.004820266862710317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,128,0.0031818665564060213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,128,0.005806933343410492
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,4096,64,0.02105813423792521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,64,0.0030080000559488933
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,64,0.005890133480230967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,4096,32,0.021843200922012328
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,32,0.003159466634194056
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,32,0.0058773333827654515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,5120,2048,0.008345599969228108
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3584,65536,0.11947200298309327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3584,65536,0.1469173272450765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,65536,0.09939839839935302
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,5120,1024,0.006393600006898243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3584,16384,0.05664106607437134
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3584,16384,0.03340799808502197
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,65536,0.17262825965881348
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,16384,0.030397866169611615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,16384,0.04672960042953491
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3584,12288,0.02746666669845581
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3584,12288,0.04953386783599854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,12288,0.02293333411216736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,12288,0.03678613503774007
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3584,10240,0.04138773282368978
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3584,10240,0.022562134265899658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,10240,0.032808534304300946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,10240,0.016266666849454246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3584,8192,0.01993173360824585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3584,8192,0.03442879915237427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,8192,0.026786132653554277
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,8192,0.013414399822553
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3584,7168,0.0502623995145162
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,7168,0.024331732590993246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,7168,0.012014933427174886
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3584,6144,0.04662293195724487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,6144,0.011141332983970641
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3584,6144,0.014910933375358582
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,6144,0.021361066897710165
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3584,5120,0.012898133198420206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,5120,0.011267200112342834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3584,5120,0.04215253194173177
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,5120,0.010309333602587383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3584,4096,0.037922132015228274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,4096,0.008657067020734151
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,4096,0.009168000022570292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3584,3584,0.03612693150838216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3584,3584,0.010406399766604107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,3584,0.00804799993832906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,3584,0.008824533224105835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3584,3072,0.035293865203857425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,4096,10240,0.024209066232045492
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,3072,0.007368533313274384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,3072,0.008612266182899475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3584,2560,0.03272320032119751
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3584,2560,0.00853013296922048
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,2560,0.006631466746330261
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,2560,0.008002133170763651
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3584,2048,0.03017599980036418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3584,2048,0.0075093333919843035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,2048,0.005826133489608765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,2048,0.0074879998962084455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3584,1536,0.028474666674931842
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3584,1536,0.0069930667678515124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,1536,0.005122133096059163
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,1536,0.007090133428573608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3584,1024,0.026428800821304322
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3584,1024,0.005997866888840993
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,1024,0.004474666714668274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,1024,0.006678399940331777
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3584,768,0.02494293252627055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3584,768,0.005677866439024607
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,768,0.006444799900054932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,768,0.004167466859022776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3584,512,0.024270933866500855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3584,512,0.005479466418425242
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,512,0.006320000191529592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,512,0.0037589333951473236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3584,256,0.022563199202219643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,256,0.0034186666210492453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3584,256,0.005026133358478546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,256,0.005987200140953064
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3584,128,0.021689599752426146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3584,128,0.004614399870236715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,128,0.0031360000371932983
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,128,0.006071466704209646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3584,64,0.02121493419011434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,64,0.0031306666632493338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,64,0.005906133353710175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,4096,2560,0.008734933535257975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3584,32,0.020885332425435384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,32,0.003115733216206233
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,32,0.005885866781075796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3072,65536,0.10933547019958496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,4096,1024,0.006298666695753734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3072,65536,0.1382570743560791
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,65536,0.14761066436767578
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,65536,0.09035733540852865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3072,16384,0.054765868186950686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3072,16384,0.03317760030428569
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,16384,0.027962666749954224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,16384,0.04326933224995931
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3072,12288,0.04721920092900594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3072,12288,0.026035199562708538
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,12288,0.03401813507080078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,12288,0.01758506695429484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3072,10240,0.03680213292439778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3072,10240,0.021501866976420085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,10240,0.029255467653274535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3072,8192,0.032408533493677776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,10240,0.015031466881434122
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3072,8192,0.019000534216562906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,8192,0.025116799275080363
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,8192,0.012552533547083536
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3072,7168,0.04961493412653605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3072,7168,0.016122666994730632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,7168,0.0221834659576416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,7168,0.011764267086982727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3072,6144,0.045510399341583255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3072,6144,0.014758400122324624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,6144,0.013363200426101684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3072,5120,0.04196266730626424
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,6144,0.010936533411343891
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3072,5120,0.012520533800125123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,5120,0.011941333611806233
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,5120,0.010246400038401287
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3072,4096,0.03802560170491536
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3072,4096,0.01087679962317149
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,4096,0.010060800115267436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,4096,0.009136000275611877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3072,3584,0.03640213410059611
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,3584,0.008144000172615051
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3072,3584,0.0100682665904363
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,3584,0.008794666330019633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3072,3072,0.03461546500523885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3072,3072,0.009457066655158997
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,3072,0.007333333293596904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,2560,0.006611200173695882
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,3072,0.008343467116355896
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3072,2560,0.03260266582171122
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,2560,0.007947733501593272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3072,2048,0.03020586570103963
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3072,2048,0.007406933108965556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,2048,0.005832533538341522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,2048,0.007415466507275899
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3072,1536,0.028457599878311157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,1536,0.007274666428565979
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3072,1536,0.0068245331446329755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,1536,0.00510506679614385
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3072,1024,0.026210133234659833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,1024,0.004538666705290476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,1024,0.006734933455785115
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3072,768,0.025127466519673663
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3072,768,0.0056981335083643595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3584,7168,0.01629866659641266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,768,0.0041696002086003625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,768,0.006479999919732411
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,512,0.0037845333417256674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3072,512,0.024205867449442545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,512,0.006227200229962667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3072,256,0.022562134265899658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3072,256,0.004953599969546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,256,0.0033791999022165934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3584,4096,0.011238400141398113
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,256,0.006078933179378509
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3072,128,0.022386133670806885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3072,128,0.004638933142026265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,128,0.0031530665854612983
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,128,0.005832533538341522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3072,64,0.021793067455291748
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,64,0.002961066613594691
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,64,0.005871999760468801
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,3072,32,0.02090880076090495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,32,0.0030805334448814393
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,32,0.005885866781075796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2560,65536,0.09929172992706299
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3584,3072,0.009986133376757304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2560,65536,0.13148480256398518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,65536,0.08296533425649008
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,65536,0.12503466606140137
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2560,16384,0.05251413186391195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2560,16384,0.030637866258621214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,16384,0.02555519938468933
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,16384,0.03668053150177002
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2560,12288,0.04313173294067383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2560,12288,0.023400533199310302
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,12288,0.028535467386245728
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,12288,0.01585813363393148
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2560,10240,0.036549333731333414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,10240,0.02469653288523356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,10240,0.014272000392278036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2560,8192,0.032545065879821776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2560,8192,0.017641599973042807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,8192,0.019729065895080566
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,8192,0.012488533059755962
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2560,7168,0.04992213249206543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,7168,0.012319999933242797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,7168,0.011653332908948263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2560,6144,0.04607359965642293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2560,6144,0.01400213340918223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,6144,0.011495467027028401
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,6144,0.01086186667283376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2560,5120,0.012329600254694621
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2560,5120,0.04226453304290771
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,5120,0.010241066416104633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,5120,0.010146133104960124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2560,4096,0.03811093171437581
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2560,4096,0.010056533416112264
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,4096,0.009078400333722432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,4096,0.009033600489298504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2560,3584,0.03629333178202311
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2560,3584,0.00941546658674876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,3584,0.008303999900817871
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,3584,0.008712533116340637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2560,3072,0.03454080025355021
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2560,3072,0.008658132950464885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,3072,0.00791786660750707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,3072,0.008278400202592214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2560,2560,0.007886933286984761
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2560,2560,0.032605866591135665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,2560,0.006537599861621857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,2560,0.007874133189519246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2560,2048,0.03079040050506592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,2048,0.005842133363087972
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,2048,0.007375999788443248
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2560,1536,0.028280532360076903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2560,1536,0.006449066599210103
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,1536,0.005130666494369507
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,1536,0.006994133194287618
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2560,1024,0.026375466585159303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2560,1024,0.005761066575845083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,1024,0.004462933540344239
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,1024,0.006605866551399231
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2560,768,0.025049599011739095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2560,768,0.005525333185990652
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,768,0.004113066693147024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,768,0.006486399968465169
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2560,512,0.023795199394226075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2560,512,0.00522986650466919
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,512,0.0037674665451049804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,512,0.006168533364931742
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2560,256,0.0226474662621816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2560,256,0.004845866560935974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,256,0.0033887999753157297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,256,0.005972266693909963
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2560,128,0.021594667434692384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2560,128,0.0046410664916038515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,128,0.003125333289305369
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,128,0.0058005332946777345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2560,64,0.0214954674243927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,64,0.0029397333661715193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,64,0.005827199916044871
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2560,32,0.021821866432825722
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,32,0.0030389333764712016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,32,0.005842133363087972
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2048,65536,0.10465386708577473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2048,65536,0.12593066692352295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,65536,0.10200213591257731
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,65536,0.07591040134429931
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2048,16384,0.048455464839935306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2048,16384,0.0281056006749471
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,16384,0.029521065950393676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,16384,0.019564799467722573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2048,12288,0.038172801335652665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2048,12288,0.022380799055099487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,12288,0.023890133698781332
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,12288,0.015875200430552162
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2048,10240,0.03485759894053141
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2048,10240,0.019942400852839152
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3072,2560,0.008529067039489746
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,10240,0.018922666708628334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,10240,0.01420799990495046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2048,8192,0.03133760094642639
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2048,8192,0.01609386702378591
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,8192,0.010890666643778484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,8192,0.012381866574287415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2048,7168,0.05072640180587769
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2048,7168,0.014482133587201438
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,7168,0.010283733407656353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,7168,0.01156160036722819
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2048,6144,0.045643734931945804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3072,1024,0.005967999994754791
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2048,6144,0.013371733824412027
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,6144,0.009417600433031718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3072,512,0.005351466437180838
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,6144,0.01074666678905487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2048,5120,0.04182826677958171
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2048,5120,0.011574400464693706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,5120,0.008759466807047527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2048,4096,0.03799039920171102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,5120,0.009980799754460652
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2048,4096,0.009782399733861287
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,4096,0.008002133170763651
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,4096,0.008929066856702169
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2048,3584,0.036055465539296463
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2048,3584,0.009309867024421692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,3584,0.007702399790287018
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,3584,0.008557867010434468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,3072,0.007356800138950348
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2048,3072,0.03429866631825765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2048,3072,0.008674133817354839
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,3072,0.00811520020167033
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2048,2560,0.03236906727155049
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2048,2560,0.008082133531570435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,2560,0.006380799909432728
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,2560,0.007716266810894013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2048,2048,0.029872000217437744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2048,2048,0.007333333293596904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,2048,0.005691733459631602
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,2048,0.0072618668278058365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2048,1536,0.028387200832366944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2048,1536,0.00666133314371109
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,1536,0.004986666639645894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,1536,0.006872533261775971
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2048,1024,0.02643520037333171
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2048,1024,0.005861333509286245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,1024,0.004355200131734212
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,1024,0.006450133522351582
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2048,768,0.024741333723068238
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2560,10240,0.020163200298945107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2048,768,0.00555733342965444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,512,0.00360000009338061
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,512,0.00603413333495458
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,768,0.003987200061480204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,768,0.006318933268388112
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2048,512,0.024175999561945598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2560,7168,0.015837867061297098
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2048,512,0.00525439977645874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2048,256,0.02258560061454773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2048,256,0.005004799862702688
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,256,0.003291733314593633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,256,0.0058783998092015585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2048,128,0.02234026590983073
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,128,0.0031082667410373688
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,128,0.005718400080998739
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2048,64,0.021144533157348634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,64,0.0029333333174387617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,64,0.005752533177534739
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,2048,32,0.021811199188232423
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,32,0.0029546665648619336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,32,0.005672533313433329
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1536,65536,0.09453226725260416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1536,65536,0.1195093313852946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,65536,0.0791040023167928
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,16384,0.024075732628504435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,65536,0.07018880049387613
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1536,12288,0.03592960039774577
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1536,16384,0.042454401652018234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2560,2048,0.007506133119265239
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,16384,0.019530665874481202
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1536,10240,0.03349226713180542
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1536,10240,0.01955519914627075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1536,12288,0.02209279934565226
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,12288,0.013117866714795432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,12288,0.01588053305943807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,10240,0.011571199695269267
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,10240,0.014201600352923074
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1536,8192,0.031225599845250446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1536,8192,0.017280000448226928
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,8192,0.009895466764767965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,8192,0.012404266993204753
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1536,7168,0.04994560082753499
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1536,7168,0.014728533228238425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,7168,0.009511466821034749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,7168,0.011590400338172912
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1536,6144,0.01332373321056366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1536,6144,0.04619306723276774
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,6144,0.008835200468699138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,6144,0.010809600353240967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1536,5120,0.04178666671117147
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1536,5120,0.011844266454378765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,5120,0.0082997332016627
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,5120,0.010019200046857198
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1536,4096,0.03819733460744222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1536,4096,0.009963732957839967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,4096,0.007534933090209961
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,4096,0.00886400043964386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1536,3584,0.036311467488606766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1536,3584,0.009369599819183349
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,3584,0.007717333237330119
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,3584,0.008515200018882752
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1536,3072,0.0343018651008606
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1536,3072,0.008789333701133727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1536,2560,0.007829333345095318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,3072,0.007433600227038066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,3072,0.008150400221347808
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1536,2560,0.03231253425280253
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,2560,0.006380799909432728
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,2560,0.007796266674995422
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1536,2048,0.030188800891240437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1536,2048,0.00697386662165324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,2048,0.005703466633955637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,1536,0.005008000135421753
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,2048,0.007298133273919423
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1536,1536,0.028347732623418172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1536,1536,0.006485333542029063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,1536,0.006919466455777486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1536,1024,0.026662399371465046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1536,1024,0.005704533557097117
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,1024,0.00432640016078949
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,1024,0.006486399968465169
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1536,768,0.02500266631444295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1536,768,0.005453866720199585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,768,0.003947733342647553
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,768,0.006297599772612255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1536,512,0.023766400416692098
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1536,512,0.0052159999807675685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,512,0.003668266783157984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,512,0.006045866509278615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1536,256,0.02263039946556091
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1536,256,0.004827733337879181
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,256,0.0032981333633263906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,256,0.0058890665570894875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,128,0.0057781333724657696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1536,128,0.02148800094922384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1536,128,0.004582400123278299
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,128,0.0030975999931494398
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1536,64,0.02172586719195048
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,64,0.0029535998900731405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,64,0.005737600227197012
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1536,32,0.020782933632532755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,32,0.0029557332396507262
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,32,0.005816533168156942
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1024,65536,0.07407466570536295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1024,65536,0.11320746739705403
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,65536,0.05615573326746622
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,65536,0.06732373237609864
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1024,16384,0.040275200208028154
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,16384,0.011436800161997478
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,16384,0.019142399231592812
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1024,12288,0.035750401020050046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1024,12288,0.015953066945075988
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,12288,0.009677867094675701
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,12288,0.015833600362141927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1024,10240,0.0331167995929718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1024,10240,0.013992533087730408
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,10240,0.009301333626111349
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,10240,0.014222932855288186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1024,8192,0.03128426671028137
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1024,8192,0.012238933642705282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1024,7168,0.049763198693593344
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,8192,0.00804373323917389
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,8192,0.012315733234087627
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,7168,0.007707733412583668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,7168,0.011595732967058818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1024,6144,0.04562453428904216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1024,6144,0.010261332988739014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,6144,0.007125333448251088
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,6144,0.01081706682840983
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1024,5120,0.04201279878616333
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1024,5120,0.009397332866986591
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2048,128,0.0046634669105211895
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,5120,0.007570133109887441
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,5120,0.010010666648546855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1024,4096,0.037912531693776445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1024,4096,0.008335999647776286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,4096,0.00713919997215271
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,4096,0.008903466661771138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1024,3584,0.03633493185043335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1024,3584,0.007952000200748443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,3584,0.007463466624418895
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,3584,0.008476799726486206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1024,3072,0.03426666657129924
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1024,3072,0.0074986666440963745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,3072,0.007162666817506154
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,3072,0.008117333551247915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1024,2560,0.0321557343006134
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1024,2560,0.0069013332327206925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,2560,0.006301866471767425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,2560,0.007754666606585185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1024,2048,0.030436267455418903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1024,2048,0.006361599763234456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,2048,0.005582933127880096
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,2048,0.007249066730340321
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1024,1536,0.028394667307535808
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1024,1536,0.006016000111897787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1536,16384,0.02795413335164388
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,1536,0.0049333333969116214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,1536,0.00689386675755183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1024,1024,0.026270933945973712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1024,1024,0.005569066603978475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,1024,0.0042805333932240805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,1024,0.006391466657320659
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1024,768,0.024987733364105223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1024,768,0.005328000088532766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,768,0.0039264000952243805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,768,0.006274133423964183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1024,512,0.023735467592875162
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,512,0.003576533248027166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,512,0.00606826643149058
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1024,256,0.022959999243418374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1024,256,0.00477866679430008
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,256,0.003253333270549774
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,256,0.005868799984455109
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1024,128,0.02231360077857971
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1024,128,0.00452159990866979
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,128,0.0030623999734719592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,128,0.005692799886067709
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1024,64,0.02174293398857117
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,64,0.0028746667007605235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,64,0.005746133128801982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,1024,32,0.02073919971783956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,32,0.002926933268706004
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,32,0.005713066458702088
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,768,65536,0.06993066469828288
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,768,65536,0.11140480041503906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,65536,0.045266131560007736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,768,16384,0.04088426828384399
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,65536,0.066484268506368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,768,16384,0.019477333625157675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,16384,0.019326933224995933
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,768,12288,0.03584533135096232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,16384,0.010140800476074218
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,768,12288,0.015852800011634825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,12288,0.01573973298072815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,12288,0.009193600217501322
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,768,10240,0.01400213340918223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,768,10240,0.03347520033518474
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,10240,0.008611200253168742
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,10240,0.014099199573198953
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,768,8192,0.031540266672770184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,768,8192,0.0120170662800471
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,8192,0.007783466577529907
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,8192,0.012337066729863485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,768,7168,0.050204801559448245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,7168,0.007293866574764251
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,7168,0.011542399724324543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,768,6144,0.04561493396759033
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,768,6144,0.010188800096511842
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,6144,0.007134933272997539
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,6144,0.010717866818110149
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,768,5120,0.041571199893951416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,768,5120,0.009154133001963298
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,5120,0.007502933343251546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,5120,0.009956266482671101
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,768,4096,0.03787519931793213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,768,4096,0.008261333405971526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,4096,0.0071050668756167095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,4096,0.008865066369374593
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,768,3584,0.036099199453989664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,768,3584,0.007796266674995422
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,3584,0.007505066692829132
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,3584,0.008477866649627686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,768,3072,0.0342410683631897
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,768,3072,0.007342933118343354
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,3072,0.007145600020885467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,3072,0.008144000172615051
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,2560,0.006284800171852112
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,768,2560,0.03209706743558248
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,768,2560,0.006870399912198384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,2560,0.00774186650911967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,768,2048,0.03015786608060201
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,768,2048,0.00647680014371872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,2048,0.00557226687669754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,2048,0.007282133400440216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,768,1536,0.028410667181015016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,1536,0.0048885335524876915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,1536,0.00688213308652242
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,768,1024,0.026333866516749065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,768,1024,0.005541333556175232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,1024,0.004263466596603394
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,1024,0.006504533191521962
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,768,768,0.005231999854246775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1024,16384,0.019729065895080566
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,768,768,0.024996266762415568
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,768,0.003885866701602936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,768,512,0.00498879998922348
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,768,0.006292266647020976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,768,512,0.02360639969507853
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,512,0.0036085332433382668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,512,0.005997866888840993
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,768,256,0.022935465971628825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,768,256,0.004646400113900503
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,256,0.00323840007185936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,256,0.005841066439946493
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,768,128,0.02230186661084493
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,768,128,0.0044170667727788295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,128,0.003048533449570338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,128,0.0056981335083643595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,768,64,0.02177066604296366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1024,7168,0.011192533373832702
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,64,0.002883200099070867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,64,0.005767466624577841
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,768,32,0.020898133516311646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,32,0.0029418667157491045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,32,0.005729066828886667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,512,65536,0.10718506971995037
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,512,65536,0.06874986489613852
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,65536,0.03372373183568318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,65536,0.0652021328608195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,512,16384,0.040445868174235025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,16384,0.008363733688990276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,16384,0.01920213301976522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,512,12288,0.03582933346430461
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,512,12288,0.016064000129699708
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,12288,0.0077450667818387345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,12288,0.015793066223462424
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,512,10240,0.03339093526204427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,10240,0.007683200140794118
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,10240,0.014186666409174601
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,512,8192,0.03142720063527425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,512,8192,0.012101333340009053
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,8192,0.007414400080839793
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,8192,0.01237546702226003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,512,7168,0.05008533398310343
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,512,7168,0.011133866508801778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,7168,0.00721919983625412
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,7168,0.011567999919255573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,512,6144,0.045406933625539145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,512,6144,0.010207999746004741
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,6144,0.0069482664267222095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,6144,0.010724266370137531
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,512,5120,0.04183573325475057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,512,5120,0.009252267082532246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1024,512,0.005008000135421753
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,5120,0.007366399963696797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,5120,0.009986133376757304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,512,4096,0.03773546616236369
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,512,4096,0.008284799754619598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,4096,0.006958933174610138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,4096,0.008831999699274699
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,512,3584,0.03605653444925944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,512,3584,0.007860266665617625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,3584,0.0075914666056633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,512,3072,0.034049065907796223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,3584,0.008505599697430928
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,512,3072,0.007405866682529449
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,3072,0.007065600156784058
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,3072,0.008105599880218506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,512,2560,0.0322432001431783
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,512,2560,0.006865066786607106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,2560,0.006209066510200501
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,2560,0.007779199878374736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,512,2048,0.02993599971135457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,512,2048,0.006372266511122386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,2048,0.005549866457780202
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,2048,0.007285333176453908
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,512,1536,0.028246400753657024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,512,1536,0.00595413347085317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,1536,0.004891733328501383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,1536,0.0068800002336502075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,512,1024,0.026220800479253133
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,512,1024,0.005418666700522105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,1024,0.004257066547870636
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,1024,0.006446933249632518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,512,768,0.024961066246032716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,512,768,0.005248000224431356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,768,0.003870933254559835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,768,0.006307200094064077
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,512,512,0.023692800601323446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,512,512,0.004925866425037384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,512,0.0035861333211263022
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,512,0.006077866752942403
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,512,256,0.022434133291244506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,512,256,0.004714666803677877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,256,0.0032437334458033243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,256,0.005875200033187866
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,512,128,0.021542400121688843
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,512,128,0.004527999957402547
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,128,0.003065599997838338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,128,0.005765333275000254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,512,64,0.021053866545359293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,64,0.002903466671705246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,64,0.005730133255322774
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,512,32,0.020886399348576865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,32,0.0029077333708604175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,32,0.005757866799831391
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,256,65536,0.10344853401184081
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,256,65536,0.06761706670125325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,65536,0.016049066185951234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,65536,0.06150506734848023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,256,16384,0.040088534355163574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,768,7168,0.01114026705423991
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,16384,0.0072736000021298725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,16384,0.019203199942906698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,256,12288,0.03595946629842122
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,12288,0.007136000196139018
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,12288,0.015734400351842245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,256,10240,0.03356159925460815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,256,10240,0.01397119959195455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,10240,0.007368533313274384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,10240,0.014057599504788718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,256,8192,0.03140693306922913
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,256,8192,0.012011733651161195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,8192,0.007256533205509186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,8192,0.01221440037091573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,256,7168,0.011018666625022887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,256,7168,0.05016853411992391
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,7168,0.007047466437021891
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,7168,0.011505066355069479
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,256,6144,0.045525332291920975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,256,6144,0.010141866405804952
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,6144,0.006863999863465626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,6144,0.01062506635983785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,256,5120,0.041945600509643556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,256,5120,0.009175466497739156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,768,1536,0.005970133344332377
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,5120,0.007282133400440216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,5120,0.009880533814430237
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,256,4096,0.037862400213877365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,256,4096,0.008185600241025288
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,4096,0.006853333115577698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,4096,0.008809600273768108
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,256,3584,0.03632320165634155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,256,3584,0.007772799829641979
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,3584,0.007402666906515758
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,3584,0.008396800359090168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,256,3072,0.03424959977467855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,3072,0.006958933174610138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,3072,0.008106666803359985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,256,2560,0.03214400013287862
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,256,2560,0.006821333368619282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,2560,0.006190933287143707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,2560,0.00768746683994929
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,256,2048,0.029868799448013305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,256,2048,0.006316799918810527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,2048,0.005478399991989136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,2048,0.007177599767843883
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,256,1536,0.027923200527826948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,256,1536,0.0059232001503308615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,1536,0.0048650667071342465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,1536,0.0067669332027435304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,256,1024,0.026195200284322102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,256,1024,0.005387733379999796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,1024,0.004214400053024292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,1024,0.0063967997829119366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,256,768,0.024733867247899374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,256,768,0.00520000010728836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,768,0.0038880000511805216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,768,0.006131199995676676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,256,512,0.023611734310785927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,256,512,0.004958933095137278
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,512,0.003583999971548716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,512,0.006018133461475372
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,256,256,0.02291946609814962
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,256,256,0.004579199850559235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,256,0.0032106667757034303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,256,0.0058677335580190025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,256,128,0.02145493427912394
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,256,128,0.004568533102671305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,128,0.003045333425203959
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,128,0.005671466886997223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,256,64,0.021028266350428263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,64,0.0028949332733949023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,64,0.005706666906674703
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,256,32,0.020768000682195028
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,32,0.0029045333464940387
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,32,0.005654400090376536
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,128,65536,0.10235839684804279
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,128,65536,0.0673642635345459
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,65536,0.013194666306177775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,65536,0.0597109317779541
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,128,16384,0.04047360022862752
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,128,16384,0.019486933946609497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,16384,0.007064533233642578
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,16384,0.019142399231592812
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,128,12288,0.035964798927307126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,128,12288,0.0158869336048762
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,12288,0.015589333573977151
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,12288,0.006967466572920482
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,128,10240,0.03344853321711223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,512,16384,0.019593600432078043
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,10240,0.0071936001380284624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,10240,0.01406719982624054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,128,8192,0.03128106594085693
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,128,8192,0.012072533369064331
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,8192,0.006922666728496551
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,8192,0.012180266777674358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,128,7168,0.05001493295033773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,512,10240,0.014058666427930197
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,7168,0.011527466773986816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,128,7168,0.011031466722488403
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,7168,0.006866133213043213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,128,6144,0.010133333007494609
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,128,6144,0.045577601591746016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,6144,0.006770133475462596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,6144,0.010618666807810467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,128,5120,0.042499200503031416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,128,5120,0.009099733829498292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,5120,0.0071839998165766404
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,5120,0.009805867075920105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,128,4096,0.03854933182398478
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,128,4096,0.008141866823037466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,4096,0.006650666892528534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,4096,0.008876799543698629
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,128,3584,0.03616426785786946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,128,3584,0.007707733412583668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,3584,0.008407466610272725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,3584,0.007231999933719635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,128,3072,0.03408106565475464
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,128,3072,0.007227733234564463
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,3072,0.006899199883143107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,3072,0.008044800162315369
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,128,2560,0.03192960023880005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,128,2560,0.006786133348941803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,2560,0.006206933160622915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,2560,0.007602133353551229
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,128,2048,0.030453334252039593
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,128,2048,0.006258133550484974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,2048,0.00547626664241155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,2048,0.007234133283297221
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,128,1536,0.028332799673080444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,128,1536,0.005871999760468801
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,1536,0.004852266609668731
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,1536,0.0068351998925209045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,128,1024,0.026165332396825152
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,128,1024,0.005343999962011973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,1024,0.0041909332076708475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,1024,0.006354133288065593
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,128,768,0.025283199548721314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,128,768,0.0051584000388781226
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,768,0.0038751999537150065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,768,0.00631039987007777
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,128,512,0.023667200406392416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,128,512,0.004807466765244802
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,512,0.0035648000737031303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,512,0.006041599810123444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,128,256,0.022294400135676067
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,128,256,0.004589866598447164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,256,0.005776000022888183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,256,0.003189333279927572
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,128,128,0.02229866584142049
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,128,128,0.004298666616280874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,128,0.0029994666576385496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,128,0.005672533313433329
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,128,64,0.02096959948539734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,64,0.005713066458702088
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,64,0.0028405333558718365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,128,32,0.020680532852808634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,32,0.002868266652027766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,32,0.005650133391221364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,65536,0.011538133025169373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,64,65536,0.10258666674296062
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,65536,0.05973226626714071
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,64,16384,0.04039786656697591
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,16384,0.006921599805355072
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,16384,0.019008000691731773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,64,12288,0.03587626616160075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,12288,0.006887466708819072
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,12288,0.015639467040697734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,64,10240,0.03352746566136678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,10240,0.00697813332080841
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,10240,0.014151466886202493
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,64,8192,0.03130773305892944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,8192,0.006884266436100006
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,8192,0.01225386659304301
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,64,7168,0.04989333152770996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,7168,0.0067210664351781205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,7168,0.01144533356030782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,64,6144,0.045902931690216066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,6144,0.006683733562628429
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,6144,0.010689066847165425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,64,5120,0.041934935251871745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,5120,0.006903466582298279
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,5120,0.009929600358009338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,64,4096,0.037904000282287596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,4096,0.006546133259932201
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,4096,0.008774399757385254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,256,16384,0.019661867618560792
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,256,12288,0.016049066185951234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,64,3584,0.03662186861038208
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,3584,0.006962133447329204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,3584,0.008504533767700195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,64,3072,0.03445546627044678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,2560,0.0076896001895268755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,3072,0.006692266464233399
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,64,2048,0.030007465680440264
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,3072,0.00804906686147054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,64,2560,0.032757333914438885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,2560,0.0061471998691558834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,2048,0.005513600011666616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,2048,0.007212799787521362
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,64,1536,0.02802666624387105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,1536,0.004818133513132731
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,1536,0.006819200019041698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,64,1024,0.02611733277638753
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,1024,0.00417493333419164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,1024,0.006412800153096516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,64,768,0.024833067258199056
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,768,0.003824000060558319
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,768,0.006295466423034668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,64,512,0.023447465896606446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,512,0.0035125332574049628
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,512,0.005990399916966756
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,64,256,0.022477867205937703
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,256,0.003160533308982849
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,256,3072,0.007315200070540111
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,256,0.005713066458702088
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,64,128,0.021461333831151327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,128,0.002961066613594691
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,128,0.005690666536490122
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,64,64,0.020963199933369956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,64,0.0028480000793933867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,64,0.005990399916966756
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,64,32,0.02072640061378479
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,32,0.002845866729815801
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,32,0.005635199944178263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,65536,0.011992533008257549
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,32,65536,0.10298453172047932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,65536,0.06011840105056763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,32,16384,0.04030933380126953
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,16384,0.0072202667593956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,16384,0.019003732999165853
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,32,12288,0.035674667358398436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,12288,0.0071733335653940845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,12288,0.015619200468063355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,32,10240,0.0335093339284261
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,10240,0.006982400019963582
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,10240,0.01397119959195455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,32,8192,0.03131733338038127
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,8192,0.006751999755700429
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,8192,0.012220799922943115
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,32,7168,0.05032000144322714
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,7168,0.0065738668044408154
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,7168,0.011416533589363098
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,32,6144,0.045628801981608076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,6144,0.006439466774463653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,6144,0.010584533214569092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,32,5120,0.04203946590423584
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,5120,0.006796800096829732
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,32,3584,0.036824532349904376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,5120,0.009828266501426697
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,32,4096,0.03783359924952189
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,4096,0.006381866832574208
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,4096,0.008758399883906047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,3584,0.006909866631031036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,32,2560,0.03219733238220215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,3584,0.008433066805203756
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,32,3072,0.034991999467213944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,3072,0.006525866687297821
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,3072,0.008036266764005024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,2560,0.006126933296521505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,2560,0.007654400169849395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,32,2048,0.02991146643956502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,2048,0.005480533341566721
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,2048,0.007237333556016285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,32,1536,0.02797546585400899
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,1536,0.004806399842103322
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,1536,0.006701866785685222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,32,1024,0.026187733809153242
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,1024,0.0041685332854588825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,128,10240,0.013864533106486002
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,512,0.0059690664211908976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,1024,0.006413866579532623
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,32,768,0.025252266724904375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,768,0.0037632000943024955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,768,0.0062720000743865965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,32,512,0.02344213326772054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,512,0.0034933333595593774
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,32,256,0.022396800915400188
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,256,0.0031456001102924346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,64,0.005618133147557576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,256,0.0058005332946777345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,32,128,0.022266666094462075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,128,0.0029237332443396253
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,128,0.00565226674079895
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,32,64,0.020940800507863365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,64,0.0028170667588710784
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,32,32,32,0.02065066695213318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,32,0.0028586665789286296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,32,0.005628799895445505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,65536,16384,0.3905173301696777
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,65536,16384,0.39269653956095374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,65536,12288,0.2723562558492025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,65536,12288,0.29975252151489257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,65536,16384,0.8039253234863282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,65536,16384,0.4344479878743489
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,65536,12288,0.5692810694376628
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,65536,12288,0.3080373446146647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,65536,10240,0.2503178596496582
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,65536,10240,0.2660000006357829
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,65536,8192,0.19856959978739422
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,65536,8192,0.21019627253214518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,65536,8192,0.22848213513692223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,65536,8192,0.3792640050252279
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,65536,10240,0.4735402743021647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,65536,7168,0.17883626619974774
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,65536,7168,0.17348160743713378
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,65536,7168,0.21598079999287925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,65536,7168,0.35718825658162434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,65536,6144,0.18934292793273927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,65536,6144,0.28616320292154945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,65536,6144,0.15520639419555665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,65536,5120,0.13185386657714843
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,65536,5120,0.16811413764953614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,65536,5120,0.25658987363179525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,65536,5120,0.1277728001276652
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,65536,4096,0.09658559958140055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,65536,4096,0.14026452700297037
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,65536,4096,0.11714239915211995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,65536,3584,0.08747093677520752
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,65536,4096,0.1935690720876058
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,65536,3584,0.12326719760894775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,65536,3584,0.09519466559092203
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,65536,3584,0.1688447952270508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,65536,3072,0.07614080111185709
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,65536,3072,0.11152533690134685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,65536,3072,0.1457472006479899
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,65536,3072,0.08276906808217367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,65536,2560,0.06354026794433594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,65536,2560,0.09773333072662353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,65536,2560,0.06803306738535563
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,65536,2560,0.12191999753316243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,65536,2048,0.08411306540171305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,65536,2048,0.0532970666885376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,65536,2048,0.09788800080617269
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,65536,2048,0.05599040190378825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,65536,1536,0.07166186968485513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,65536,1536,0.04163839817047119
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,65536,1536,0.07570133209228516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,65536,1536,0.04489493370056152
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,65536,1024,0.05853653351465861
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,65536,1024,0.05185386737187704
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,65536,1024,0.032304000854492185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,65536,768,0.053082664807637535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,65536,768,0.02405866583188375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,65536,768,0.04013546705245972
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,65536,768,0.02669866681098938
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,65536,512,0.012971733013788858
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,65536,512,0.041966934998830155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,65536,512,0.0290613333384196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,65536,512,0.01760639945665995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,65536,256,0.010206933816274006
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,65536,256,0.013928533593813578
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,65536,256,0.03271573384602865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,65536,256,0.007909333209196727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,65536,128,0.030961066484451294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,65536,128,0.005798399945100148
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,65536,128,0.01125440001487732
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,65536,64,0.031117866436640423
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,65536,64,0.004755199948946635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,65536,64,0.011559466520945232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,65536,32,0.03127786715825399
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,65536,32,0.004817066589991251
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,65536,32,0.011427199840545655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,16384,65536,0.3943466822306315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,65536,0.4261642773946126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,16384,65536,0.48110081354777023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,16384,16384,0.10353493690490723
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,65536,0.8030495961507162
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,16384,16384,0.14064745903015136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,16384,0.11925546328226726
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,16384,0.19942293167114258
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,16384,12288,0.11265386740366619
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,12288,0.08969279925028482
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,12288,0.14013973871866864
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,16384,10240,0.06594666639963785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,16384,10240,0.09771093527475992
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,10240,0.1164896011352539
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,10240,0.07436800003051758
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,16384,8192,0.051276799043019614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,16384,8192,0.08137386639912923
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,8192,0.05927679936091105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,8192,0.10177493095397949
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,16384,7168,0.04581866661707561
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,16384,7168,0.09434560139973959
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,7168,0.08921386400858561
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,7168,0.052738134066263834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,16384,6144,0.08485013643900553
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,16384,6144,0.04224106470743815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,65536,10240,0.22894506454467772
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,6144,0.07225706577301025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,6144,0.046665600935618085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,16384,5120,0.07770773569742838
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,16384,5120,0.03903253475824992
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,5120,0.06049173275629679
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,5120,0.04025493462880452
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,16384,4096,0.028807467222213744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,16384,4096,0.06727999846140544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,65536,6144,0.14043199221293132
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,4096,0.04947520097096761
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,4096,0.033131732543309526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,16384,3584,0.062009600798288975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,3584,0.02861119906107585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,16384,3584,0.02723840077718099
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,3584,0.050892798105875645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,16384,3072,0.023334399859110514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,16384,3072,0.057501868406931556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,3072,0.03808853228886922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,3072,0.02773546576499939
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,16384,2560,0.0538101315498352
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,16384,2560,0.019581866264343262
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,2560,0.03742080132166545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,2560,0.02160960038503011
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,16384,2048,0.038395734628041585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,2048,0.01534933348496755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,2048,0.026680533091227216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,16384,1536,0.034399998188018796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,16384,1536,0.009956266482671101
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,1536,0.021041067441304524
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,1536,0.012930132945378623
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,16384,1024,0.030741333961486816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,16384,1024,0.008498133222262064
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,1024,0.007867733140786488
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,1024,0.010663466652234395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,16384,768,0.02922026713689168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,16384,768,0.007480533421039581
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,768,0.0063967997829119366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,768,0.00958079993724823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,16384,512,0.02738453348477681
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,16384,512,0.006478933493296306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,512,0.005312000215053558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,512,0.00863039990266164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,65536,1024,0.02977386713027954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,16384,256,0.024770132700602212
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,256,0.004332800209522247
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,256,0.00765119989713033
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,16384,128,0.023809067408243813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,16384,128,0.005301333467165629
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,128,0.0037119999527931214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,128,0.00727893312772115
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,16384,64,0.02445440093676249
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,64,0.0034133332471052804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,64,0.007255466779073079
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,16384,32,0.024472532669703166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,32,0.007283199826876323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,65536,128,0.00808426688114802
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,32,0.0034805332620938623
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,12288,65536,0.28799041112264
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,65536,0.3073557217915853
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,12288,16384,0.10291626453399658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,12288,16384,0.07923413117726644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,12288,65536,0.383404795328776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,65536,0.5663871765136719
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,16384,0.08271360397338867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,16384,0.1459466616312663
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,12288,12288,0.0852458635965983
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,12288,12288,0.06727573076883951
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,12288,10240,0.05012906789779663
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,12288,0.06226239999135336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,12288,10240,0.07476800282796224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,12288,0.11318826675415039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,10240,0.08914986451466879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,10240,0.05347520112991333
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,12288,8192,0.04075946807861328
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,12288,8192,0.0654368003209432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,8192,0.07050666809082032
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,8192,0.04433173338572184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,12288,7168,0.036814932028452554
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,12288,7168,0.08039999802907308
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,16384,12288,0.07414506276448568
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,7168,0.06244266827901205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,7168,0.03935146729151408
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,12288,6144,0.0740821361541748
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,12288,6144,0.03239679932594299
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,6144,0.0598090648651123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,6144,0.03457280000050862
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,12288,5120,0.06451199849446615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,12288,5120,0.02793813347816467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,5120,0.04588266611099243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,5120,0.030376533667246502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,12288,4096,0.05683733224868774
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,12288,4096,0.023474133014678954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,4096,0.03810773293177287
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,4096,0.025124265750249224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,12288,3584,0.052001067002614344
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,12288,3584,0.02118720014890035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,3584,0.03364799817403157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,3584,0.022848000129063926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,12288,3072,0.04131840070088704
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,12288,3072,0.0172160009543101
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,3072,0.029550933837890626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,3072,0.016796799500783284
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,12288,2560,0.036348799864451095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,12288,2560,0.014578133821487427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,2560,0.02487466732660929
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,2560,0.01362773378690084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,12288,2048,0.03379093408584595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,12288,1536,0.030913066864013673
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,12288,2048,0.012165333827336628
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,2048,0.02312320073445638
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,2048,0.012485333283742269
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,12288,1536,0.010179199775060018
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,1536,0.009918933113416035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,1536,0.010814932982126872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,12288,1024,0.027841067314147948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,12288,1024,0.008087466657161712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,1024,0.007378133138020833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,1024,0.009267200032869976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,12288,768,0.02622293432553609
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,12288,768,0.007208533088366191
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,768,0.0061951999862988796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,768,0.008455466230710347
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,12288,512,0.02478933334350586
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,512,0.005083733300367991
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,512,0.0077002664407094315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,12288,256,0.022472532590230306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,12288,256,0.005479466418425242
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,256,0.004091733445723852
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,256,0.0069930667678515124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,16384,2048,0.012061867117881774
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,12288,128,0.02167146603266398
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,12288,128,0.005157333115736643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,128,0.003639466563860575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,128,0.00664213349421819
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,12288,32,0.022333866357803343
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,12288,64,0.021569067239761354
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,64,0.003357866654793421
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,64,0.00665280024210612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,32,0.0033728001018365227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,32,0.006693333387374878
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,16384,256,0.005824000140031179
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,10240,65536,0.23763839403788248
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,10240,16384,0.09224853515625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,65536,0.25980159441630046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,10240,65536,0.3128351847330729
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,16384,0.07107413609822591
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,65536,0.4764938672383626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,16384,0.1171989361445109
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,10240,12288,0.07605120340983072
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,10240,12288,0.05676586627960205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,12288,0.09181439876556396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,12288,0.06293119986852011
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,10240,10240,0.06720746358235677
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,10240,10240,0.042498131593068436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,10240,0.04757013320922852
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,10240,0.07465066909790039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,10240,8192,0.05978879928588867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,10240,8192,0.03496533234914144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,8192,0.060471467177073156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,8192,0.039111467202504475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,10240,7168,0.031413332621256514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,10240,7168,0.07450133164723714
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,7168,0.05359679857889811
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,7168,0.03468586603800456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,10240,6144,0.06762986977895101
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,10240,6144,0.02770986755688985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,6144,0.047408000628153486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,6144,0.031034666299819946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,10240,5120,0.06132693290710449
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,10240,5120,0.02434239983558655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,5120,0.03979093233744303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,5120,0.02718293269475301
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,4096,0.023357866207758586
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,10240,4096,0.019771732886632285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,10240,4096,0.052179201443990075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,4096,0.032245333989461264
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,10240,3584,0.04489066600799561
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,10240,3584,0.014972800016403198
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,3584,0.028668800989786785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,3584,0.01648319959640503
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,10240,3072,0.03882559935251872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,10240,3072,0.013034666577974955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,3072,0.025352533658345538
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,2560,0.021513599157333373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,3072,0.014525866508483887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,10240,2560,0.03593493302663167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,10240,2560,0.01188053290049235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,2560,0.012921599547068277
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,10240,2048,0.03309333324432373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,2048,0.01430400013923645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,2048,0.011464533209800721
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,10240,1536,0.030595199267069502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,10240,1536,0.008784000078837078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,1536,0.007666133344173431
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,1536,0.010372266173362732
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,10240,1024,0.02772480050722758
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,10240,1024,0.007708799839019775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,1024,0.006084266801675161
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,1024,0.008890666564305623
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,10240,768,0.026441599925359088
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,10240,768,0.007003733515739441
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,768,0.005374933282534281
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,768,0.008250666658083598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,10240,512,0.024541866779327393
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,10240,512,0.006213333209355672
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,10240,256,0.00555733342965444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,512,0.004553600152333578
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,512,0.007533866663773854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,10240,256,0.022629332542419434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,10240,128,0.005182933310667674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,256,0.003804799914360046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,256,0.006968533496061961
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,10240,128,0.021656533082326256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,128,0.0034272000193595886
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,128,0.006637866795063019
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,10240,64,0.02224853237469991
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,32,0.003236266722281774
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,32,0.006694399813810985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,64,0.00322026660044988
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,64,0.006585599978764851
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,10240,32,0.021913599967956544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,12288,512,0.006166400015354156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,8192,65536,0.19733120600382487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,65536,0.21652372678120932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,8192,16384,0.05353600184122721
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,8192,16384,0.07917546431223552
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,65536,0.3757386525472005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,8192,65536,0.2509898662567139
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,16384,0.06649706761042276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,16384,0.09838720162709555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,8192,12288,0.06685760021209716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,8192,12288,0.04652693271636963
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,12288,0.051115731398264565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,12288,0.0830666700998942
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,8192,10240,0.060020267963409424
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,10240,0.05938560167948405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,10240,0.044300798575083414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,8192,8192,0.05299413204193115
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,8192,8192,0.029182932774225873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,8192,0.048691201210021975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,8192,0.036484265327453615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,8192,7168,0.06906560262044271
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,8192,7168,0.02622399926185608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,7168,0.042896000544230144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,7168,0.033004800478617355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,8192,6144,0.06314880053202311
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,6144,0.03753493229548137
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,6144,0.029259733359018963
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,8192,5120,0.05575573444366455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,8192,5120,0.020082134008407592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,5120,0.021522132555643718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,5120,0.03699626525243123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,8192,4096,0.04197226762771607
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,8192,4096,0.014386133352915446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,10240,16384,0.06574079990386963
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,4096,0.026272000869115193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,4096,0.01702293356259664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,8192,3584,0.0379253347714742
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,8192,3584,0.01328426698843638
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,3584,0.02346773346265157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,3584,0.015871999661127727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,8192,3072,0.03590293327967326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,8192,3072,0.012004266182581585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,3072,0.020549333095550536
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,3072,0.013677866260210673
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,8192,2560,0.03345066706339518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,8192,2560,0.01067626674969991
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,2560,0.01577173372109731
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,2560,0.012650666634241739
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,8192,2048,0.03155413269996643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,2048,0.00788800021012624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,2048,0.011291733384132386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,8192,1536,0.029282132784525555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,8192,1536,0.008189866443475087
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,1536,0.006816000243028005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,1536,0.009965866804122925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,8192,1024,0.0270687997341156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,8192,1024,0.007083733379840851
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,1024,0.005358933409055074
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,1024,0.008622933427492778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,8192,768,0.02527359922726949
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,8192,768,0.00653653343518575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,768,0.004772266745567322
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,768,0.007874133189519246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,8192,512,0.024306132396062215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,8192,512,0.005791999896367391
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,512,0.004260266820589701
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,512,0.007309866448243459
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,8192,256,0.022613332668940226
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,8192,256,0.005294933418432872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,256,0.0037802666425704955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,256,0.006844800213972728
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,8192,128,0.021809067328770956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,128,0.0065311998128890995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,10240,2048,0.010364799698193868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,8192,64,0.02140480081240336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,8192,128,0.004972800115744273
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,128,0.003385599950949351
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,64,0.003173333406448364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,64,0.00643093337615331
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,8192,32,0.02187839945157369
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,32,0.0032330666979153953
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,32,0.006623999774456024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,7168,65536,0.17280960083007812
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,65536,0.19126933415730793
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,7168,16384,0.07213013172149658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,7168,65536,0.2332533359527588
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,7168,16384,0.05092373291651407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,65536,0.32274665832519533
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,16384,0.05193599859873453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,16384,0.09171199798583984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,7168,12288,0.06688746611277262
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,12288,0.07107199827829996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,12288,0.04047466516494751
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,7168,10240,0.03331093390782674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,7168,10240,0.05636800130208334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,10240,0.05731306473414103
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,10240,0.03519253333409627
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,7168,8192,0.050040535132090246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,7168,8192,0.026825600862503053
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,8192,0.04325013160705567
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,8192,0.029309866825739543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,7168,7168,0.06602986653645834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,7168,7168,0.02434879938761393
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,7168,0.038251733779907225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,7168,0.02649066646893819
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,7168,6144,0.059706668059031166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,6144,0.03370026747385661
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,6144,0.02109973430633545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,7168,5120,0.047076265017191574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,7168,5120,0.017146666844685875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,5120,0.028593067328135175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,5120,0.016718933979670204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,7168,4096,0.04099626541137695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,7168,4096,0.014018133282661438
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,4096,0.023666133483250938
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,4096,0.013462400436401368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,7168,3584,0.03829439878463745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,7168,3584,0.012866133451461792
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,3584,0.02116159995396932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,3584,0.012653866410255432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,7168,3072,0.03614826599756877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,8192,10240,0.03558506568272908
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,3072,0.016736000776290894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,3072,0.011677866180737812
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,7168,2560,0.03378986517588298
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,7168,2560,0.01034879982471466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,2560,0.008748799562454224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,2560,0.010522666573524474
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,7168,2048,0.031199999650319415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,2048,0.007458133498827617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,2048,0.00949013332525889
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,7168,1536,0.02895680069923401
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,7168,1536,0.007826133569081625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,1536,0.006417066852251689
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,8192,6144,0.023266132672627768
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,1536,0.008565333485603333
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,7168,1024,0.006771199901898702
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,7168,1024,0.02667306661605835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,1024,0.005129600067933401
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,1024,0.0076117331782976794
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,7168,768,0.02517333428064982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,768,0.004603733122348785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,7168,768,0.00633493314186732
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,768,0.007106133302052816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,7168,512,0.023916800816853843
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,512,0.003993600110212962
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,7168,512,0.005805866420269012
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,512,0.006708266834417979
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,7168,256,0.022010666131973267
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,7168,256,0.005148800214131674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,256,0.0035445332527160645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,256,0.006178133189678192
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,7168,128,0.022057600816090903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,128,0.00329066663980484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,128,0.005983999868233999
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,8192,2048,0.009418666362762451
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,7168,64,0.02113173405329386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,64,0.0031136001149813337
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,64,0.00595413347085317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,7168,32,0.021437867482503255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,32,0.003124266614516576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,32,0.00600853314002355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,6144,65536,0.16621012687683107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,6144,65536,0.19426347414652506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,65536,0.2836117426554362
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,65536,0.17672425905863445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,6144,16384,0.049192531903584795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,6144,16384,0.07229653199513754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,16384,0.04771519899368286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,16384,0.0758016029993693
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,6144,12288,0.05782613356908163
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,6144,12288,0.035336534182230636
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,12288,0.05685760180155436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,12288,0.03466666539510091
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,6144,10240,0.052442665894826254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,6144,10240,0.02962133288383484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,10240,0.0472981333732605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,10240,0.03232640027999878
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,6144,8192,0.024513065814971924
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,6144,8192,0.047918931643168135
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,8192,0.03814506530761719
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,8192,0.025337600708007814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,6144,7168,0.022453333934148154
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,6144,7168,0.06362560192743937
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,7168,0.03501760164896647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,7168,0.02409706711769104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,6144,6144,0.053191467126210534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,6144,6144,0.018178133169809978
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,6144,0.029978666702906293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,6144,0.017527467012405394
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,6144,5120,0.04535146554311116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,6144,5120,0.01544533371925354
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,5120,0.02581546703974406
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,5120,0.014962133765220643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,6144,4096,0.03995413382848104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,6144,4096,0.013058132926623025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,4096,0.021860265731811525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,4096,0.012389333049456278
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,6144,3584,0.03849386771519979
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,6144,3584,0.012365866700808208
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,3584,0.018938666582107543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,3584,0.012012799580891926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,7168,12288,0.03838613430658976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,6144,3072,0.03576426506042481
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,6144,3072,0.011251200238863628
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,3072,0.010752000411351522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,3072,0.009965866804122925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,6144,2560,0.03356906572977702
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,2560,0.008633599678675333
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,6144,2560,0.010238933563232421
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,2560,0.010107733805974324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,6144,2048,0.03127786715825399
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,6144,2048,0.008887466788291932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,2048,0.007446399827798207
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,2048,0.00921066701412201
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,6144,1536,0.028970666726430255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,1536,0.006126933296521505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,1536,0.008158933122952778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,7168,6144,0.02198186715443929
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,6144,1024,0.02645546595255534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,6144,1024,0.007028266787528992
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,1024,0.005021866659323374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,1024,0.007461333274841308
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,6144,768,0.02502400080362956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,6144,768,0.006367999811967213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,768,0.006870399912198384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,768,0.004520533482233683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,6144,512,0.023909332354863484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,6144,512,0.0057429333527882894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,512,0.003967999915281932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,512,0.006687999765078227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,6144,256,0.022073600689570108
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,6144,256,0.005346133311589559
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,256,0.0035349334279696143
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,256,0.006185600161552429
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,128,0.0032842665910720824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,6144,128,0.021358933051427206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,7168,3072,0.011762133240699768
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,128,0.0059797331690788266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,6144,64,0.021062399943669638
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,64,0.003047466774781545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,64,0.006042666733264923
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,7168,2048,0.009051733215649923
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,6144,32,0.021871999899546305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,32,0.003115733216206233
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,32,0.00591786652803421
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,5120,65536,0.1395594596862793
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,65536,0.13673173586527507
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,5120,65536,0.17875199317932128
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,5120,16384,0.061121066411336265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,65536,0.23372586568196616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,16384,0.060331734021504727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,16384,0.039061331748962404
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,5120,12288,0.05299946864446005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,5120,12288,0.03203306595484416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,12288,0.04643413225809733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,12288,0.03110186656316121
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,5120,10240,0.04829866488774617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,5120,10240,0.02730026642481486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,10240,0.03927573362986247
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,5120,8192,0.04357866843541463
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,10240,0.026693334182103474
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,5120,8192,0.022227199872334798
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,8192,0.03195093274116516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,8192,0.01879040002822876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,5120,7168,0.05546346505482992
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,5120,7168,0.018921599785486857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,7168,128,0.00483840008576711
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,7168,0.02852906584739685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,7168,0.017027199268341064
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,5120,6144,0.049466665585835776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,5120,6144,0.016108799974123636
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,6144,0.025092266003290814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,6144,0.01453439990679423
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,5120,5120,0.04411733150482178
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,5120,5120,0.01516266663869222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,5120,0.02197973330815633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,5120,0.01304639975229899
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,5120,4096,0.03997759819030762
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,5120,4096,0.012843733032544455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,4096,0.014311466614405313
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,4096,0.011796266833941142
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,5120,3584,0.03770986795425415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,3584,0.009324799974759419
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,3584,0.010745599865913391
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,5120,3072,0.035740800698598224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,5120,3072,0.010805333654085796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,3072,0.008172800143559773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,3072,0.010079999764760334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,5120,2560,0.03374613523483276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,2560,0.0074869334697723385
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,2560,0.009365333120028178
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,5120,2048,0.030947200457255047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,5120,2048,0.008700799942016602
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,2048,0.006403199831644694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,2048,0.008743466933568318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,5120,1536,0.028884265820185346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,1536,0.0056309332450230915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,1536,0.007785599927107494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,5120,1024,0.02673706610997518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,5120,1024,0.00658133327960968
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,1024,0.004643199841181437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,1024,0.007228800157705943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,5120,768,0.025332266092300416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,5120,768,0.006072533130645752
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,768,0.004154666761557261
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,768,0.006863999863465626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,5120,512,0.023800534009933472
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,5120,512,0.005551999807357788
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,512,0.003790933390458425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,512,0.0065984000762303666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,5120,256,0.022782933712005616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,5120,256,0.005076266825199127
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,256,0.0034186666210492453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,256,0.006121600170930227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,5120,128,0.021908267339070638
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,5120,128,0.00488319993019104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,128,0.0032138665517171226
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,128,0.005915733178456625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,5120,64,0.02178773283958435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,64,0.0030410667260487873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,6144,1536,0.00792746643225352
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,64,0.005964800218741099
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,5120,32,0.02153600056966146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,32,0.0030570665995279947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,32,0.0058005332946777345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,4096,65536,0.139463472366333
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,4096,65536,0.15414293607076007
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,65536,0.18688319524129232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,65536,0.10969599882761638
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,4096,16384,0.05617920160293579
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,4096,16384,0.038705066839853926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,16384,0.051692799727121985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,16384,0.03301013310750325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,6144,128,0.005029333134492239
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,4096,12288,0.04885760148366292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,4096,12288,0.029985066254933673
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,12288,0.04058666626612346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,12288,0.026401066780090333
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,4096,10240,0.044949332873026535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,4096,10240,0.0256223996480306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,10240,0.034713598092397054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,10240,0.02013546625773112
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,4096,8192,0.032307199637095135
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,4096,8192,0.01955839991569519
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,8192,0.028909866015116376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,8192,0.014729600151379904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,4096,7168,0.04920533498128255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,4096,7168,0.017550933361053466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,7168,0.026321067412694292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,5120,16384,0.04573653141657512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,7168,0.01334933340549469
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,4096,6144,0.045850666364034016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,4096,6144,0.01580693324406942
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,4096,5120,0.04174506664276123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,6144,0.023626667261123658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,6144,0.01237226625283559
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,4096,5120,0.014036267002423605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,5120,0.01141439974308014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,5120,0.018093866109848023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,4096,4096,0.038151466846466066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,4096,4096,0.012065066893895467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,4096,0.009756799538930256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,4096,0.009826133648554485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,4096,3584,0.03591573238372803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,4096,3584,0.011169067025184632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,3584,0.009147733449935913
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,3584,0.009367466966311137
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,4096,3072,0.03397013346354167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,4096,3072,0.010191999872525533
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,3072,0.007389866809050243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,3072,0.008909866213798523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,4096,2560,0.03234133323033651
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,2560,0.0061930666367212926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,2560,0.008390399813652038
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,4096,2048,0.030159999926884968
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,4096,2048,0.008162133395671844
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,2048,0.0055285334587097164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,2048,0.007771733403205872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,4096,1536,0.028178133567174274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,4096,1536,0.007449600100517273
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,1536,0.004996266464392344
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,1536,0.007245866457621257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,4096,1024,0.0260917325814565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,4096,1024,0.006397866706053417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,5120,3584,0.012171733379364013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,1024,0.004391466577847799
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,1024,0.006705066561698914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,4096,768,0.02459519902865092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,4096,768,0.006072533130645752
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,768,0.004041599979003271
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,768,0.006388266881306966
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,4096,512,0.02391786575317383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,5120,2560,0.009819733103116353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,512,0.006253866851329804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,512,0.0037578667203585305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,4096,256,0.021820799509684245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,4096,256,0.005106133222579956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,256,0.0033952000240484873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,256,0.006016000111897787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,5120,1536,0.007795199751853943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,4096,128,0.021228800217310585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,4096,128,0.004725333551565806
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,128,0.0031413334111372627
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,128,0.005801600217819214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,4096,64,0.02104746699333191
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,64,0.0030069333811601003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,64,0.0060032000144322716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,4096,32,0.021781333287556968
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,32,0.0031167998909950255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,32,0.005826133489608765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3584,65536,0.12949972947438557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3584,65536,0.1400053342183431
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,65536,0.0987061341603597
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3584,16384,0.039826134840647384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,65536,0.17562133471171062
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3584,16384,0.053556267420450845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,16384,0.04571733474731445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,16384,0.030243200063705445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3584,12288,0.04627626736958822
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3584,12288,0.028667734066645308
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,12288,0.03596373399098714
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,12288,0.024465066194534302
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3584,10240,0.03678613503774007
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3584,10240,0.0231989324092865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,10240,0.0322762668132782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,10240,0.016780799627304076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,8192,0.026272000869115193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,8192,0.014252799749374389
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3584,8192,0.03139413396517436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3584,8192,0.019579732418060304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3584,7168,0.049556267261505124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3584,7168,0.017590399583180746
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,7168,0.024145066738128662
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,7168,0.012641066312789917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3584,6144,0.04528213342030843
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,6144,0.01981546680132548
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,6144,0.012285866340001424
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3584,5120,0.04174720048904419
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,5120,0.010896000266075134
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,5120,0.011249066392580668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3584,4096,0.037724800904591876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3584,4096,0.012078932921091716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,4096,0.009596799810727436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,4096,0.009907199939092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3584,3584,0.036185598373413085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3584,3584,0.011201066772143047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,3584,0.00907306671142578
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,3584,0.009404800335566203
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3584,3072,0.034031999111175534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3584,3072,0.010222933689753215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,3072,0.007204266885916392
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,3072,0.009035733342170716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3584,2560,0.03227199912071228
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3584,2560,0.008965333302815754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,2560,0.006494933366775512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,2560,0.008414933085441589
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3584,2048,0.030244266986846922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3584,2048,0.007964799801508587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,2048,0.005794133245944977
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,2048,0.007850666840871174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3584,1536,0.0280021329720815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3584,1536,0.007150933146476746
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,1536,0.005130666494369507
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,1536,0.007255466779073079
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3584,1024,0.026225066184997557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3584,1024,0.006200533111890157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,1024,0.004437333345413208
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,1024,0.0067104001839955645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3584,768,0.024727465709050496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3584,768,0.005815466741720835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,768,0.004065066576004028
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,768,0.006433066725730896
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3584,512,0.023670399188995363
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3584,512,0.005447466671466827
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,512,0.003718400001525879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,512,0.0062496001521746315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,256,0.003402666747570038
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3584,256,0.022026666005452476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3584,256,0.004906666775544485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,256,0.005964800218741099
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3584,128,0.021191465854644775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3584,128,0.004964266717433929
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,128,0.0030858665704727173
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,128,0.005829333265622457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,64,0.005735466877619425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3584,64,0.020961066087086998
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,64,0.0029824001093705496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,4096,2560,0.009354666868845622
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3584,32,0.0217685341835022
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,32,0.0030250666042168934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,32,0.0057322666049003605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3072,65536,0.12483200232187908
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,65536,0.14775466918945312
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,65536,0.08978986740112305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3072,16384,0.05150933265686035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3072,16384,0.037418667475382486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,16384,0.027987200021743774
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,16384,0.04081706603368123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3072,12288,0.02755413254102071
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,12288,0.03217173417409261
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3072,12288,0.041842134793599446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,12288,0.018055466810862224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3072,10240,0.03314346671104431
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,4096,512,0.005430399874846141
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3072,10240,0.021846399704615275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,10240,0.016293332974116007
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,10240,0.02786986629168193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3072,8192,0.030691200494766237
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3072,8192,0.018665599822998046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,8192,0.02440213362375895
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,8192,0.012839466333389282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3072,7168,0.04907093445460002
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3072,7168,0.017100799083709716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,7168,0.020166399081548055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,7168,0.012115200360616047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,6144,0.012301866213480632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3072,6144,0.04523733456929525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3072,6144,0.015380266308784484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,6144,0.011109333237012227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3072,5120,0.04120106697082519
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3072,5120,0.013249066472053529
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,5120,0.011070932944615681
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,5120,0.010221866766611735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3072,4096,0.03784639835357666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3072,4096,0.011530666550000509
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,4096,0.00949013332525889
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,4096,0.009093333284060161
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3072,3584,0.03592106501261393
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,3584,0.009054932991663616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,3584,0.008721066514650981
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3072,3072,0.03399680058161418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,3072,0.008452266454696655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,3072,0.008547199765841167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3072,2560,0.032313599189122515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3072,2560,0.008776533603668212
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,2560,0.006435200075308482
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,2560,0.007970133423805236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3072,2048,0.02988053361574809
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3072,2048,0.007842133442560833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,2048,0.0057429333527882894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,2048,0.007469866673151653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3072,1536,0.007159466544787089
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3072,1536,0.028039467334747315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,1536,0.005089066425959269
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,1536,0.0070720002055168155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3072,1024,0.026045866807301837
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3072,1024,0.006211199859778086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,1024,0.004401066899299621
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,1024,0.006621866424878438
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3072,768,0.024650667111078897
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3072,768,0.005820799867312113
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,768,0.004049066702524821
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,768,0.00644053320089976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3072,512,0.02385279933611552
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3072,512,0.005443199972311656
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,512,0.00373333344856898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3584,6144,0.015963733196258545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,512,0.006232533355553945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3072,256,0.021953066190083824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3072,256,0.004860800007979075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,256,0.0033770665526390077
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,256,0.005977599819501241
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3072,128,0.021180800596872964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3072,128,0.004654933512210846
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,128,0.003124266614516576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3584,5120,0.014018133282661438
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,128,0.005787733197212219
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3072,64,0.020925867557525634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,3072,32,0.021350399653116862
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,64,0.0029525332152843474
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,64,0.005756799876689911
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,32,0.0030293333033720653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,32,0.00576853354771932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2560,65536,0.10652480125427247
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2560,65536,0.11731200218200684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,65536,0.12025173505147298
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,65536,0.08097919623057047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2560,16384,0.049029334386189775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2560,16384,0.03567999998728434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,16384,0.03472533226013184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,16384,0.02351893385251363
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2560,12288,0.03908053239186605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2560,12288,0.025499733289082845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,12288,0.027896533409754436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,12288,0.015947733322779337
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2560,10240,0.03255680004755656
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2560,10240,0.019719467560450236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,10240,0.02421440084775289
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,10240,0.014270933469136557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2560,8192,0.030483200152715044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2560,8192,0.019338667392730713
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,8192,0.017771732807159425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,8192,0.012760532895723977
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2560,7168,0.04891839822133382
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2560,7168,0.01757546663284302
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2560,6144,0.015586133797963461
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,7168,0.011874133348464965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,7168,0.011648000280062357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2560,6144,0.04503466685612996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2560,5120,0.04179733196894328
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,6144,0.010442666212717692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,6144,0.010800000031789143
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2560,5120,0.01360106666882833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,5120,0.009654399752616883
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,5120,0.010073600212732951
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2560,4096,0.03743253151575725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2560,4096,0.011827199657758077
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,4096,0.008557867010434468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,4096,0.009006933371225993
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2560,3584,0.035605335235595705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2560,3584,0.010762666662534077
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,3584,0.008124800026416778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,3584,0.008663466572761536
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2560,3072,0.03373759984970093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2560,3072,0.009845333298047383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,3072,0.007743999858697255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,3072,0.008236800134181977
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2560,2560,0.03221120039621989
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2560,2560,0.008770133058230083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,2560,0.006490666667620342
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,2560,0.007837866743405659
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2560,2048,0.029918932914733888
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3072,65536,0.11673493385314941
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2560,2048,0.0077461332082748415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,2048,0.005764266848564148
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,2048,0.007417599856853485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2560,1536,0.02808000048001607
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,1536,0.004999466737111409
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,1536,0.00697920024394989
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2560,1024,0.02580053408940633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2560,1024,0.006011733412742614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,1024,0.004344533383846283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,1024,0.006551466882228851
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2560,768,0.02487466732660929
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,768,0.004012800008058548
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,768,0.006420266628265381
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2560,512,0.023635200659434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2560,512,0.0053610667586326596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,512,0.0036650667587916053
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,512,0.006182399888833364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2560,256,0.022004266579945884
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2560,256,0.004874666531880697
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,256,0.0033621333539485933
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,256,0.005924266576766968
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2560,128,0.021170133352279664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2560,128,0.00459199994802475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,128,0.0030794667700926462
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,128,0.005772800246874491
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2560,64,0.02090239922205607
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,64,0.002962133288383484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3072,3584,0.010662399729092916
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,64,0.0057322666049003605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2560,32,0.021724800268809
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,32,0.002974933385848999
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,32,0.0058037335673968
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2048,65536,0.11957333087921143
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2048,65536,0.11199466387430827
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,65536,0.09881172974904379
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,65536,0.0741493304570516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3072,3072,0.00978559950987498
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2048,16384,0.044582398732503255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,16384,0.028843732674916585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,16384,0.019266132513682047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2048,12288,0.03439146677652995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2048,12288,0.024106667439142863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,12288,0.023346134026845298
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,12288,0.01569066643714905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2048,10240,0.0331712007522583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2048,10240,0.020463999112447104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,10240,0.014140799641609192
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,10240,0.017836799224217735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2048,8192,0.03054506580034892
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2048,8192,0.017108267545700072
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,8192,0.010089600086212158
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,8192,0.012270933389663697
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2048,7168,0.049268265565236405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2048,7168,0.015633066495259605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,7168,0.010051199793815612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,7168,0.0115146666765213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2048,6144,0.045510399341583255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2048,6144,0.013939199844996133
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,6144,0.009014399846394856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,6144,0.01074773371219635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2048,5120,0.04133653243382772
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2048,5120,0.012530133128166199
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2048,4096,0.03774079879124959
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,5120,0.008563199639320373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,5120,0.009937066833178203
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2048,4096,0.010800000031789143
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,4096,0.007837866743405659
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,4096,0.008876799543698629
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2048,3584,0.03585919936498006
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2048,3584,0.010393599669138592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,3584,0.007542400062084198
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,3584,0.008453333377838134
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2048,3072,0.0339626669883728
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2048,3072,0.009398399790128072
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,3072,0.007241599758466085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,2560,0.006331733365853627
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,3072,0.008055466910203297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,2560,0.00774186650911967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2048,2560,0.03223680059115092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2048,2560,0.008506666620572407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2048,2048,0.02981013258298238
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2048,2048,0.007638399799664815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,2048,0.005660800139109293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,2048,0.007210666437943776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2048,1536,0.028078933556874592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2048,1536,0.006905599931875865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,1536,0.004975999891757965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,1536,0.006826666494210561
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2048,1024,0.02605653405189514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2048,1024,0.006111999849478403
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,1024,0.004282666742801667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,1024,0.0064074665307998655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2048,768,0.024655999739964803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2048,768,0.0056874667604764305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,768,0.003945599993069967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,768,0.0062730665008227035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2048,512,0.02364906668663025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,512,0.0035914666950702667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,512,0.006046933432420095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2048,256,0.02193386753400167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,256,0.0032927999893824257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,256,0.005799466868241628
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2048,128,0.021754666169484457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2048,128,0.004716800153255462
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,128,0.003052799900372823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,128,0.005693866809209188
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2048,64,0.021670399109522502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,64,0.0029279999434947968
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,64,0.005730133255322774
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,2048,32,0.021485867102940877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,32,0.002918400118748347
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,32,0.0056970665852228795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1536,65536,0.10595520337422688
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1536,65536,0.10113173325856525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,65536,0.07661120096842447
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,65536,0.06904853185017903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1536,16384,0.038097067674001055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1536,16384,0.03234346707661946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,16384,0.023433599869410196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,16384,0.019181867440541588
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1536,12288,0.034209068616231283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1536,12288,0.023541333278020223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,12288,0.012641066312789917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,12288,0.015748266379038492
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1536,10240,0.032509867350260416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1536,10240,0.020052266120910645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,10240,0.011389866471290588
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,10240,0.014254933595657349
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2560,1536,0.007064533233642578
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1536,8192,0.030612266063690184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1536,8192,0.01705706715583801
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,8192,0.009597866733868917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,8192,0.01232426663239797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1536,7168,0.049114668369293214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2560,768,0.0057781333724657696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1536,7168,0.015471999843915304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,7168,0.009175466497739156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,7168,0.011549866199493409
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1536,6144,0.04534080028533936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1536,6144,0.0139765332142512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,6144,0.008589866757392883
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,6144,0.010762666662534077
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1536,5120,0.041530664761861166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,5120,0.007962666451931
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,5120,0.009948800007502239
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1536,4096,0.0380181352297465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,4096,0.007351466516653697
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,4096,0.00888320008913676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1536,3584,0.03604053258895874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1536,3584,0.009909333785374959
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,3584,0.0075914666056633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,3584,0.008526933193206788
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1536,3072,0.03382720152537028
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1536,3072,0.009018666545550029
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,3072,0.0072053333123524976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,3072,0.008186666667461396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1536,2560,0.03204266627629598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1536,2560,0.007968000074227651
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,2560,0.006325333317120869
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,2560,0.007760000228881836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1536,2048,0.029785599311192828
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1536,2048,0.007397333284219106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,2048,0.005613866448402405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2048,16384,0.032739200194676715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,2048,0.0072629332542419435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1536,1536,0.027932800849278766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1536,1536,0.006783999999364217
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,1536,0.004925866425037384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,1536,0.006875733534495036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1536,1024,0.02580053408940633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1536,1024,0.005962666869163513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,1024,0.0042581334710121155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,1024,0.0065098668138186145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,768,0.003927466770013174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1536,768,0.0247978667418162
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1536,768,0.005580799778302511
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,768,0.00631466656923294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1536,512,0.005277866621812185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,512,0.003622400015592575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1536,512,0.023526400327682495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,512,0.006150400141874949
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1536,256,0.02253546714782715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1536,256,0.004795733094215393
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,256,0.0032831999162832894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,256,0.005869866907596588
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1536,128,0.021126399437586464
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1536,128,0.004700799783070883
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,128,0.0030591999491055804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,128,0.0056426664193471275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1536,64,0.020907733837763467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,64,0.0029440000653266907
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,64,0.005721599857012431
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1536,32,0.021393066644668578
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,32,0.0029343999922275543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,32,0.005691733459631602
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1024,65536,0.10007893244425456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1024,65536,0.08877973556518555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,65536,0.05159680048624674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,65536,0.06659946839014688
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1024,16384,0.038126933574676516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1024,16384,0.027124265829722088
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,16384,0.019181867440541588
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,16384,0.011225600043932598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1024,12288,0.034213332335154216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1024,12288,0.02063466707865397
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,12288,0.009594666957855224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,12288,0.015761066476504007
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1024,10240,0.032503465811411544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,10240,0.008615466952323913
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,10240,0.01411946713924408
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1024,8192,0.030511999130249025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1024,8192,0.016285866498947144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,8192,0.00790293316046397
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,8192,0.012299733360608418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1024,7168,0.04889813264211019
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1024,7168,0.015018666783968607
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,7168,0.007652266820271809
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,7168,0.011552000045776367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1024,6144,0.04496106704076131
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1024,6144,0.013709867000579834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,6144,0.007049599786599477
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2048,512,0.005380266904830932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,6144,0.010687999924023946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1024,5120,0.04159359931945801
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2048,256,0.005038933455944061
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,5120,0.007429333527882893
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,5120,0.009872000416119893
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1024,4096,0.037843199570973714
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1024,4096,0.01032426655292511
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,4096,0.007238399982452392
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,4096,0.008804266651471455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1024,3584,0.03584106763203938
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1024,3584,0.009248000383377076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,3584,0.007381333410739899
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,3584,0.00844693382581075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1024,3072,0.034067201614379886
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1024,3072,0.008719999591509502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,3072,0.007005866865317028
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,3072,0.008012799918651581
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1024,2560,0.03215679923693339
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1024,2560,0.007161599894364675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,2560,0.006302933394908905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,2560,0.007624533275763195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1024,2048,0.02994453310966492
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1024,2048,0.006785066425800323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,2048,0.005622399846712748
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,2048,0.007189333438873291
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1024,1536,0.028056534131368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1024,1536,0.005981866518656413
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,1536,0.0048981333772341405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,1536,0.006792533397674561
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1024,1024,0.026075732707977296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1024,1024,0.005624533196290334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,1024,0.004249600072701773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,1024,0.006411733229955037
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1024,768,0.024943999449412026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,768,0.006202666461467743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1024,768,0.005416533350944519
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,768,0.003884800026814143
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1024,512,0.023654399315516154
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1024,512,0.00508480022350947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,512,0.0036127999424934386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,512,0.0060127998391787205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1024,256,0.02252266605695089
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1024,256,0.004747733473777771
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,256,0.0032874666154384612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,256,0.005831466615200042
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1024,128,0.021716266870498657
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1024,128,0.004563199977080027
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,128,0.00308693324526151
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,128,0.005689600110054016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1536,5120,0.012289067109425861
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1024,64,0.021667200326919555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,64,0.0028768000503381092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,64,0.005740800003210703
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,1024,32,0.02131839990615845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,32,0.002887466549873352
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,32,0.005685333410898844
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1536,4096,0.011215999722480774
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,768,65536,0.09717973073323569
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,768,65536,0.07666453520456949
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,768,16384,0.03817066748936971
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,65536,0.04095146656036377
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,65536,0.0662506659825643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,16384,0.009737599889437358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,16384,0.019243733088175455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,768,12288,0.034083199501037595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,768,12288,0.015753600001335143
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,12288,0.008771199981371562
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,12288,0.015797332922617594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,768,10240,0.0323093334833781
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,768,10240,0.013857066631317139
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,10240,0.008180266618728638
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,10240,0.014126933614412942
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,768,8192,0.03051626682281494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,768,8192,0.011942399541536967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,8192,0.0075989335775375364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,768,7168,0.04884693225224813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,8192,0.012331733107566833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,768,7168,0.01097813347975413
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,7168,0.007238399982452392
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,7168,0.011495467027028401
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,768,6144,0.04553279876708984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,768,6144,0.01007466713587443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,6144,0.007019733389218648
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,6144,0.010718933741251628
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,768,5120,0.041688533624013265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,768,5120,0.009245866537094116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,5120,0.00740479975938797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,5120,0.009950932860374451
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,768,4096,0.038045867284138994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,768,4096,0.008280533552169799
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,4096,0.006970666845639547
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,4096,0.008826667070388794
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,768,3584,0.035810132821400956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,768,3584,0.007761066655317943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,3584,0.007313066720962524
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,768,3072,0.034142935276031496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,768,3072,0.007310933371384938
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,3584,0.008564266562461852
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,3072,0.00703359991312027
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,3072,0.008168533444404602
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,768,2560,0.03216213385264079
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,768,2560,0.006886399785677592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,2560,0.006257066627343495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,2560,0.007719466586907704
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,768,2048,0.006357333560784657
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,768,2048,0.030039467414220172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,2048,0.005544533332188925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,2048,0.007235200206438701
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,768,1536,0.028069333235422773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,768,1536,0.006010666489601135
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,1536,0.004955733319123586
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,1536,0.006877866884072621
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,768,1024,0.025834665695826216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,768,1024,0.00544213354587555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,1024,0.00421973317861557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,1024,0.006483200192451477
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,768,768,0.02462293306986491
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,768,0.0038762666285037995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,768,0.006258133550484974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,768,512,0.02362133264541626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1024,10240,0.01757546663284302
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,512,0.003568000098069509
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,512,0.006046933432420095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,768,256,0.021897600094477335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,768,256,0.004781866570313772
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,256,0.003221333275238673
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,256,0.0058005332946777345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,768,128,0.02127466599146525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,768,128,0.004636799792448679
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,128,0.0030154667794704436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,128,0.005679999788602194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,768,64,0.02107306718826294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,64,0.002885333448648453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,64,0.005713066458702088
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,768,32,0.021699200073877968
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,32,0.002914133419593175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,32,0.005702400207519531
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,512,65536,0.09437973499298095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,65536,0.02977173328399658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,65536,0.06485333442687988
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1024,5120,0.011949867010116577
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,512,16384,0.0380021333694458
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,512,16384,0.019355734189351402
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,16384,0.008142933249473572
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,16384,0.019197867314020792
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,512,12288,0.03413333495457967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,512,12288,0.015773866573969522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,12288,0.007474133372306823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,12288,0.015749333302179973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,512,10240,0.03223359982172648
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,512,10240,0.013853866855303446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,10240,0.007397333284219106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,10240,0.01415786643822988
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,512,8192,0.030434133609135945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,512,8192,0.011986133456230164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,8192,0.007292800148328145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,8192,0.012286933263142903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,512,7168,0.04937493403752645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,512,7168,0.011027200023333232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,7168,0.007090133428573608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,7168,0.011497599879900615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,512,6144,0.04543999830881755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,6144,0.006903466582298279
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,6144,0.010682666301727295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,512,5120,0.04134933153788249
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,512,5120,0.00916373332341512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,5120,0.007367466886838277
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,5120,0.009909333785374959
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,512,4096,0.0374570647875468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,512,4096,0.008267733454704284
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,4096,0.006836266815662384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,4096,0.008849066495895386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,512,3584,0.035889065265655516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,512,3584,0.007788800199826558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,3584,0.007393066585063934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,3584,0.008470400174458822
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,512,3072,0.03374933401743571
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,512,3072,0.007309866448243459
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,3072,0.007068799932797749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,3072,0.008110933502515157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,512,2560,0.03211413423220317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,512,2560,0.006857599814732869
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,2560,0.00629013329744339
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,2560,0.00767680009206136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,512,2048,0.029960532983144123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,2048,0.005598933498064677
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,2048,0.007206400235493978
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,512,1536,0.028119466702143353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,512,1536,0.005907199780146281
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,1536,0.004934399823347728
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,1536,0.006806399921576183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,512,1024,0.025845332940419512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,1024,0.004266666869322458
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,1024,0.00643093337615331
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,512,768,0.024564266204833984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,768,16384,0.020204800367355346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,768,0.003921066721280416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,768,0.006318933268388112
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,512,512,0.02320853273073832
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,512,512,0.004923733572165171
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,512,0.003565866748491923
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,512,0.006042666733264923
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,512,256,0.021691733598709108
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,512,256,0.004681600133577982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,256,0.0032149332265059153
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,256,0.005767466624577841
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,512,128,0.02104426622390747
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,512,128,0.004452266792456309
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,128,0.003010133405526479
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,128,0.005644799768924713
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,512,64,0.020855466524759926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,64,0.0028949332733949023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,64,0.005692799886067709
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,512,32,0.021233065923055013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,32,0.002914133419593175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,32,0.005657599866390228
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,256,65536,0.08978559970855712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,256,65536,0.06624746719996134
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,65536,0.011907200018564861
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,65536,0.05982933441797892
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,256,16384,0.03806826670964559
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,256,16384,0.019454934199651084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,16384,0.007099733253320058
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,16384,0.01909653345743815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,256,12288,0.034168533484141034
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,256,12288,0.015812266866366068
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,12288,0.006966400146484375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,12288,0.015588266650835672
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,768,512,0.005070933202902476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,256,10240,0.032230399052302044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,256,10240,0.013778133193651834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,768,768,0.005202133456865946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,10240,0.007074133555094402
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,10240,0.014008532961209616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,256,8192,0.030618667602539062
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,256,8192,0.011947733163833619
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,8192,0.007045333087444305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,8192,0.012195199728012085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,256,7168,0.049252267678578696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,256,7168,0.011004799604415893
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,7168,0.006910933554172516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,7168,0.011397332946459452
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,256,6144,0.04570879936218262
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,256,6144,0.0100490669409434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,6144,0.010634666681289673
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,6144,0.0067893331249554946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,256,5120,0.009093333284060161
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,256,5120,0.04163519938786824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,5120,0.007246933380762736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,5120,0.009848533074061076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,256,4096,0.037664000193277994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,256,4096,0.008227199812730153
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,4096,0.006810666620731353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,4096,0.00874133308728536
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,256,3584,0.03584106763203938
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,256,3584,0.007771733403205872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,3584,0.007286400099595388
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,3584,0.008342400193214417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,256,3072,0.03369919856389363
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,256,3072,0.007287466526031494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,3072,0.006908800204594929
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,3072,0.00798933357000351
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,256,2560,0.03177173336346944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,512,65536,0.06649813254674276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,256,2560,0.006781866649786632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,2560,0.0061930666367212926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,2560,0.007597866654396057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,256,2048,0.029843199253082275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,256,2048,0.006321066617965698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,2048,0.005527466535568237
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,2048,0.007079466680685679
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,256,1536,0.027819732824961346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,256,1536,0.00588266650835673
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,1536,0.004859733581542969
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,1536,0.006723199784755707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,256,1024,0.027114667495091754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,1024,0.004192000130812327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,1024,0.006673066814740498
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,256,768,0.024513065814971924
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,256,768,0.00514986664056778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,768,0.0038218667109807336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,768,0.006201600035031637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,256,512,0.023285333315531412
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,256,512,0.004839466512203216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,512,0.0035082665582497918
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,512,0.005967999994754791
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,256,256,0.021667200326919555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,256,256,0.004619733492533366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,512,6144,0.01007466713587443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,256,0.003173333406448364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,256,0.005817600091298421
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,256,128,0.02118720014890035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,256,128,0.004446933170159658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,128,0.005570133527119955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,128,0.0030432000756263735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,256,64,0.020883200565973918
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,64,0.002820266783237457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,64,0.005731200178464254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,256,32,0.021191465854644775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,32,0.002906666696071625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,32,0.0055861334005991616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,128,65536,0.0887061357498169
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,128,65536,0.06632106701533
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,65536,0.011121066411336263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,65536,0.05985173384348551
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,128,16384,0.03826773166656494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,128,16384,0.019423999389012656
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,16384,0.00703359991312027
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,128,12288,0.03449813524881999
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,128,12288,0.01576959987481435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,512,2048,0.006354133288065593
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,16384,0.018911999464035035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,12288,0.0070783997575442
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,12288,0.0156768004099528
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,128,10240,0.03221333424250285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,128,10240,0.01369599997997284
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,10240,0.007233066856861115
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,8192,0.006981333096822103
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,10240,0.01395626664161682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,128,8192,0.030909866094589233
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,128,8192,0.01193386713663737
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,512,1024,0.0053962667783101406
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,8192,0.012146133184432983
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,128,7168,0.04944106737772624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,128,7168,0.010886399944623312
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,7168,0.006914133330186208
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,7168,0.011411199967066448
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,128,6144,0.010003200173377991
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,128,6144,0.045816532770792645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,6144,0.0065984000762303666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,512,768,0.005231999854246775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,6144,0.01060693363348643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,128,5120,0.041606398423512776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,128,5120,0.009093333284060161
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,5120,0.006877866884072621
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,5120,0.009842133522033692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,128,4096,0.037588266531626384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,128,4096,0.008132266501585644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,4096,0.006572799881299336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,4096,0.008789333701133727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,128,3584,0.035795199871063235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,128,3584,0.00767680009206136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,3584,0.006958933174610138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,3584,0.008289066453774769
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,128,2560,0.006737066805362702
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,128,3072,0.033659732341766356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,128,3072,0.007176533341407776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,3072,0.006826666494210561
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,3072,0.007979733248551685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,128,2560,0.032245333989461264
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,2560,0.007627733548482259
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,2560,0.005938133100668589
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,128,2048,0.029553065697352093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,128,2048,0.006183466811974844
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,2048,0.005381333331267038
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,2048,0.007101866602897644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,128,1536,0.027841067314147948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,128,1536,0.0058794667323430385
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,1536,0.004716800153255462
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,1536,0.006715733309586842
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,128,1024,0.025990400711695356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,128,1024,0.005284266670544943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,1024,0.004053333401679992
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,1024,0.006342400113741558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,128,768,0.024514132738113405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,768,0.003832533210515976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,768,0.0061034664511680605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,512,0.00595306654771169
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,128,512,0.023270400365193684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,128,512,0.004713599880536398
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,512,0.003472000112136205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,128,256,0.021628799041112264
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,128,256,0.004588800172011057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,256,0.0031061333914597826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,256,0.0057536001006762184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,128,128,0.021118932962417604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,128,128,0.004345599810282389
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,128,0.003017599880695343
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,64,0.005590400099754334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,128,0.005589333176612854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,128,64,0.020925867557525634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,64,0.0028181334336598715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,128,32,0.021461333831151327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,32,0.002809600035349528
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,32,0.005558399856090546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,64,65536,0.08684373696645101
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,65536,0.00923520028591156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,65536,0.060038399696350095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,64,16384,0.038097067674001055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,16384,0.0067775999506314594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,16384,0.01886826753616333
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,64,12288,0.03439360062281291
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,12288,0.015602133671442666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,12288,0.006875733534495036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,64,10240,0.03226026693979899
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,10240,0.00681386689345042
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,10240,0.014009599884351095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,64,8192,0.030422399441401165
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,8192,0.006739200154940288
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,8192,0.012388267119725545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,64,7168,0.049092264970143636
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,7168,0.00660159985224406
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,7168,0.011392000317573547
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,64,6144,0.045468799273173016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,6144,0.006517333288987477
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,6144,0.010643200079600016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,64,5120,0.041255466143290204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,5120,0.006833066542943318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,5120,0.009802666306495667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,64,4096,0.037546666463216145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,4096,0.006489600241184235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,4096,0.008775466680526733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,64,3584,0.03580160140991211
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,3584,0.006903466582298279
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,3584,0.008368000388145447
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,64,3072,0.03385920127232869
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,3072,0.00661653329928716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,3072,0.008087466657161712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,64,2560,0.03188160061836243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,2560,0.005820799867312113
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,2560,0.007626666625340779
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,64,2048,0.029756800333658857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,2048,0.005284266670544943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,2048,0.007242666681607564
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,64,1536,0.0277834673722585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,1536,0.004620799918969473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,1536,0.006790400048096975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,64,1024,0.025548799832661943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,1024,0.004013866682847341
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,1024,0.006377600133419037
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,64,768,0.024564266204833984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,768,0.0036757332583268487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,64,256,0.021642667055130006
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,768,0.006152533491452535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,64,512,0.02320853273073832
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,512,0.0034304000437259674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,512,0.006004266440868378
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,256,0.0030752000709374744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,256,0.0057205334305763245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,64,128,0.021194666624069214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,128,0.0029525332152843474
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,256,1024,0.005351466437180838
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,128,0.005674666663010915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,64,64,0.021144533157348634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,64,0.0027701333165168762
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,64,0.0055861334005991616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,64,32,0.021338667472203574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,32,0.0027829334139823914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,32,0.005601066847642263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,32,65536,0.08777706623077393
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,65536,0.008653866251309712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,65536,0.05948586861292521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,32,16384,0.038097067674001055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,16384,0.006941866874694824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,16384,0.01893226703008016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,32,12288,0.0344650665918986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,12288,0.006897066533565521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,12288,0.015685333808263143
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,32,10240,0.032340266307195026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,10240,0.006771199901898702
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,10240,0.013939199844996133
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,32,8192,0.030576000610987347
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,8192,0.006674133241176605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,8192,0.012223999698956807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,32,7168,0.04939733346303304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,7168,0.0065984000762303666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,7168,0.011390933394432068
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,32,6144,0.045544532934824626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,5120,0.01011946698029836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,6144,0.006425599753856659
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,6144,0.01064639985561371
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,32,5120,0.04160000085830688
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,5120,0.006772266825040181
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,32,4096,0.03763200044631958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,4096,0.0064416001240412395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,4096,0.008742400010426839
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,32,3584,0.035955198605855304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,3072,0.007965866724650066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,3584,0.0068234667181968685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,3584,0.008384000261624653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,32,3072,0.034068266550699874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,3072,0.006516266862551372
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,32,2560,0.03207359910011291
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,2560,0.005861333509286245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,2560,0.007537066439787547
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,32,2048,0.029875200986862183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,2048,0.005252266426881155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,2048,0.007032533486684163
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,32,1536,0.027974400917689008
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,1536,0.004638933142026265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,1536,0.006714666883150737
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,32,1024,0.02564053336779277
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,1024,0.004018133382002512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,1024,0.006321066617965698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,32,768,0.024757333596547446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,768,0.003700266778469086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,768,0.006156800190607706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,32,512,0.023297067483266196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,512,0.0033621333539485933
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,512,0.00594346672296524
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,32,256,0.02173653244972229
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,256,0.0031167998909950255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,256,0.005735466877619425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,32,128,0.021132800976435342
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,128,0.002940800040960312
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,128,0.005508266886075338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,32,64,0.02071466644605001
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,64,0.0028042666614055633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,64,0.005656533439954122
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,16,32,32,0.021182932456334434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,32,0.002752000093460083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,32,0.005580799778302511
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,65536,16384,0.391213862101237
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,65536,12288,0.27085545857747395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,65536,16384,0.3999402681986491
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,65536,12288,0.35258452097574866
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,128,768,0.005019733309745788
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,65536,16384,0.4318154652913411
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,65536,12288,0.5677194595336914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,65536,16384,0.7862357457478841
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,65536,12288,0.31593173344930015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,65536,10240,0.23139413197835287
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,65536,8192,0.1858133316040039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,65536,10240,0.2515893300374349
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,65536,10240,0.2939359982808431
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,65536,10240,0.47336638768513994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,65536,8192,0.3774976094563802
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,65536,8192,0.19738346735636395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,65536,7168,0.16137280464172363
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,65536,8192,0.22358080546061196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,65536,6144,0.13911147117614747
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,65536,7168,0.3299381256103516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,65536,7168,0.176692263285319
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,65536,7168,0.19794774055480957
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,65536,6144,0.1543615976969401
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,65536,6144,0.1735007921854655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,65536,6144,0.2852383931477865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,65536,5120,0.1358751932779948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,65536,5120,0.1622111956278483
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,65536,5120,0.236899201075236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,65536,4096,0.09465280373891195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,65536,4096,0.13632426261901856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,65536,4096,0.10427626768747966
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,65536,4096,0.1938389301300049
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,65536,3584,0.08979413509368897
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,65536,3584,0.1263914664586385
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,65536,3584,0.09242773056030273
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,65536,3584,0.16797226270039875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,65536,3072,0.07526720364888509
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,65536,3072,0.10686506430308025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,65536,3072,0.14459625879923504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,65536,3072,0.0821557362874349
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,65536,2560,0.06275626818339029
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,65536,2560,0.09475519657135009
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,65536,2560,0.12130026817321778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,65536,2560,0.06864213148752848
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,65536,2048,0.055940266450246176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,65536,2048,0.08190506299336751
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,65536,2048,0.09718186855316162
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,65536,1536,0.06966079870859782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,65536,1536,0.04146133263905843
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,65536,1536,0.07455680370330811
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,65536,1536,0.04451626539230347
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,65536,1024,0.029986133178075153
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,65536,1024,0.05669866800308228
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,65536,1024,0.050918400287628174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,65536,1024,0.03208213249842326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,65536,768,0.05226346651713053
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,65536,768,0.02673706610997518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,65536,768,0.024213333924611412
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,65536,768,0.03963306744893392
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,65536,512,0.039741865793863934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,65536,512,0.013879467050234476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,65536,512,0.031623466809590654
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,65536,512,0.017459199825922648
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,65536,256,0.03241173426310222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,65536,256,0.00788800021012624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,65536,256,0.013617066542307535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,65536,128,0.03069973389307658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,65536,128,0.0075989335775375364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,65536,128,0.0056085333228111265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,65536,128,0.011260799566904704
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,65536,64,0.031541333595911665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,65536,64,0.004602666695912679
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,65536,64,0.011339733004570007
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,65536,32,0.02987733284632365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,65536,32,0.004571733375390371
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,65536,32,0.011355732878049214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,16384,65536,0.39622294108072914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,65536,0.4252511978149414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,16384,16384,0.10272853374481201
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,16384,65536,0.4706997235616048
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,65536,0.8015935897827149
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,16384,16384,0.14953385988871257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,16384,0.1852458635965983
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,16384,0.11751573085784912
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,16384,12288,0.07772160371144612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,12288,0.1406592051188151
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,16384,12288,0.11205653349558513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,12288,0.08064533074696859
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,16384,10240,0.0932149330774943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,16384,10240,0.06567893425623575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,10240,0.11699519952138264
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,10240,0.0746997356414795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,16384,8192,0.08074986934661865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,8192,0.05930879910786947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,8192,0.10116373697916667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,16384,7168,0.08659520149230956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,7168,0.0830677350362142
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,7168,0.05223146677017212
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,16384,6144,0.03986560106277466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,16384,6144,0.07837973435719808
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,6144,0.07229119936625163
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,6144,0.04660053253173828
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,16384,5120,0.07005120118459066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,5120,0.06046186685562134
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,5120,0.04183573325475057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,16384,4096,0.06013226509094238
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,16384,4096,0.029927466313044233
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,4096,0.04937386512756348
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,4096,0.0343509316444397
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,16384,3584,0.0564245343208313
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,16384,3584,0.027305599053700762
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,3584,0.04351466496785482
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,3584,0.028913066784540815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,16384,3072,0.05221439997355143
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,16384,3072,0.02418773372968038
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,3072,0.038024532794952395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,65536,5120,0.12132373650868733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,3072,0.02584853370984395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,16384,2560,0.04761600097020467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,16384,2560,0.01815999945004781
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,2560,0.03440426588058472
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,2560,0.020702934265136717
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,16384,2048,0.03543893496195476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,16384,2048,0.0127018670241038
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,2048,0.027798400322596235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,2048,0.015564800302187601
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,16384,1536,0.03192426760991414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,16384,1536,0.010545066992441813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,1536,0.021887999773025513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,1536,0.013126400113105775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,16384,1024,0.028534400463104247
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,16384,1024,0.008385066191355388
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,1024,0.007763200004895528
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,1024,0.011321600278218586
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,16384,768,0.02765120069185893
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,16384,768,0.007844266792138416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,768,0.006296533346176148
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,768,0.010056533416112264
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,65536,2048,0.053091200192769375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,16384,512,0.026659200588862102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,16384,512,0.00654720018307368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,512,0.005243733525276184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,512,0.00867306689421336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,16384,256,0.025047467152277632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,16384,256,0.005747200051943461
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,256,0.00417493333419164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,256,0.0076906666159629825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,16384,128,0.023804799715677897
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,128,0.003669333209594091
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,128,0.007247999807198842
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,16384,64,0.024653865893681844
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,64,0.003345066557327906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,64,0.007239466905593872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,16384,32,0.02263360023498535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,32,0.003335466732581457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,32,0.007294933497905731
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,65536,256,0.009265066186587016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,12288,65536,0.2960927963256836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,12288,65536,0.379582945505778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,65536,0.5634016036987305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,65536,0.3087050755818685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,12288,16384,0.11206293106079102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,16384,0.09233173529307047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,16384,0.1470688025156657
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,12288,12288,0.08502613703409831
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,12288,12288,0.06199573278427124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,12288,0.062337064743041994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,12288,0.10640640258789062
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,12288,10240,0.07655359903971354
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,12288,10240,0.052731732527414955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,10240,0.08679893016815185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,10240,0.053592534859975186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,12288,8192,0.06461653312047323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,8192,0.07704106966654459
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,8192,0.0432426651318868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,12288,7168,0.0705130656560262
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,12288,7168,0.03814613421758016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,7168,0.0620138684908549
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,7168,0.03968533277511597
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,12288,6144,0.06415146589279175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,12288,6144,0.033403734366099044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,16384,8192,0.05043520132700602
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,6144,0.05588693221410116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,6144,0.0346837321917216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,12288,5120,0.05700586636861166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,12288,5120,0.02911146680514018
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,5120,0.04549013376235962
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,5120,0.03024959961573283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,12288,4096,0.05172266562779745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,16384,7168,0.045380266507466634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,12288,4096,0.023796266317367552
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,4096,0.025193599859873454
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,4096,0.037918933232625324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,16384,5120,0.034305067857106526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,3584,0.03724160194396973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,12288,3584,0.0471125324567159
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,12288,3584,0.02258560061454773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,3584,0.022887466351191203
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,12288,3072,0.03646080096562703
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,12288,3072,0.01838399966557821
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,3072,0.028884265820185346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,3072,0.01626240015029907
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,12288,2560,0.03234240015347799
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,12288,2560,0.01609386702378591
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,2560,0.02480319937070211
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,2560,0.01365013321240743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,12288,2048,0.029883732398351032
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,2048,0.021079466740290324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,2048,0.01237546702226003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,12288,1536,0.0284768005212148
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,12288,1536,0.010804266730944315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,1536,0.009922132889429728
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,1536,0.010733866691589355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,12288,1024,0.026281599203745527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,12288,1024,0.008193066716194153
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,1024,0.007216000060240428
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,1024,0.009452799956003826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,12288,768,0.025033599138259886
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,12288,768,0.007549866537253062
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,768,0.006131199995676676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,768,0.008569600184758504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,12288,512,0.02348053256670634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,12288,512,0.006267733375231425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,512,0.005032533407211303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,12288,256,0.0056533331672350565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,512,0.007696000238259633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,12288,256,0.02221226692199707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,256,0.00425600012143453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,256,0.006972800195217133
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,12288,128,0.02220373352368673
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,12288,128,0.0050453335046768185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,128,0.003640533238649368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,128,0.0067071999112765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,64,0.0034122665723164878
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,12288,64,0.02236479918162028
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,64,0.0066890666882197065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,12288,32,0.02053440014521281
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,32,0.0033215999603271483
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,32,0.006680533289909363
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,16384,128,0.005143466591835022
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,10240,65536,0.23799360593159996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,10240,16384,0.08919573624928792
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,65536,0.26108479499816895
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,10240,65536,0.3048351923624674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,65536,0.47733014424641923
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,16384,0.13052480220794677
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,16384,0.0816543976465861
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,10240,12288,0.08159039815266927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,10240,12288,0.050443732738494874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,12288,0.09023253122965494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,12288,0.05378559827804565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,10240,10240,0.04212053219477336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,10240,10240,0.06686293284098307
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,10240,0.0754197359085083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,10240,0.04728533426920573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,10240,8192,0.034302934010823564
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,10240,8192,0.05900373458862305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,8192,0.06082666714986166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,8192,0.038261334101359054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,10240,7168,0.06372479995091757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,10240,7168,0.030861866474151612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,7168,0.05373333295186361
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,7168,0.0346997340520223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,12288,16384,0.08040746847788492
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,10240,6144,0.05825599829355875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,10240,6144,0.027687466144561766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,6144,0.04875626564025879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,10240,5120,0.052705065409342444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,6144,0.030989867448806763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,5120,0.040941866238911946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,5120,0.02728533347447713
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,10240,4096,0.02141973376274109
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,10240,4096,0.04520959854125976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,4096,0.01988053321838379
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,4096,0.0322432001431783
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,12288,8192,0.04321066538492839
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,10240,3584,0.03823999961217244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,3584,0.02874133388201396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,3584,0.015793066223462424
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,10240,3072,0.03510613441467285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,10240,3072,0.013318399588267008
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,3072,0.025115732351938886
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,3072,0.014260266224543253
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,10240,2560,0.012643200159072877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,10240,2560,0.0321888009707133
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,2560,0.02342080076535543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,2560,0.012837333480517068
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,10240,2048,0.030264532566070555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,10240,2048,0.010545066992441813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,2048,0.01528320014476776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,2048,0.011523200074831645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,10240,1536,0.02815679907798767
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,10240,1536,0.00944533348083496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,1536,0.007551999886830647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,1536,0.010071466366449993
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,10240,1024,0.02593280076980591
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,10240,1024,0.00766293356815974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,1024,0.0059797331690788266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,1024,0.00902826686700185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,10240,768,0.025101866324742633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,10240,768,0.007369600236415863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,768,0.005253333350022634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,768,0.008213333288828532
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,10240,512,0.023989333709081014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,512,0.007524266839027405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,10240,512,0.0063360000650088
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,512,0.0044501334428787235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,10240,256,0.02265920042991638
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,10240,256,0.005635199944178263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,256,0.003755733370780945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,256,0.0069365332523981735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,12288,2048,0.012777599692344665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,10240,128,0.02226240038871765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,10240,128,0.005109333495299021
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,128,0.00344106654326121
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,128,0.006656000018119812
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,10240,64,0.022125866015752158
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,64,0.0031509332358837126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,64,0.006577066580454509
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,10240,32,0.020549333095550536
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,32,0.00315733328461647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,32,0.006638933221499126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,8192,65536,0.19422400792439778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,8192,16384,0.07780799865722657
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,65536,0.2430976072947184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,8192,65536,0.25187519391377766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,8192,16384,0.05373866558074951
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,65536,0.3743562698364258
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,16384,0.0924394687016805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,16384,0.07076799869537354
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,8192,12288,0.041716265678405764
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,8192,12288,0.07235413392384847
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,12288,0.046817068258921304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,12288,0.07834986845652261
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,8192,10240,0.058829867839813234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,8192,10240,0.035163732369740804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,10240,0.059622399012247716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,10240,0.043918931484222413
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,8192,8192,0.051999998092651364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,8192,8192,0.02882986664772034
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,8192,0.048072532812754316
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,8192,0.03627413511276245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,8192,7168,0.058481065432230625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,7168,0.04248426755269368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,7168,0.032536532481511435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,8192,6144,0.05407253503799438
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,8192,6144,0.02350506583849589
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,6144,0.03733119964599609
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,6144,0.026664533217748004
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,8192,5120,0.048438398043314616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,8192,5120,0.018691200017929076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,5120,0.03178133368492127
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,5120,0.021989333629608154
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,8192,4096,0.03477333386739095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,8192,4096,0.014798933267593383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,4096,0.02580053408940633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,10240,16384,0.0646453340848287
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,4096,0.017179733514785765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,8192,3584,0.03289493322372437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,8192,3584,0.013793067137400309
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,3584,0.014792533715566
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,3584,0.023124267657597862
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,8192,3072,0.03116053342819214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,8192,3072,0.012239999572436015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,3072,0.020488532384236653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,3072,0.013929599523544311
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,2560,0.013523200154304504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,8192,2560,0.029679999748865767
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,8192,2560,0.011729066570599873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,2560,0.012239999572436015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,8192,2048,0.028310400247573853
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,8192,2048,0.00978559950987498
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,2048,0.007437866429487865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,2048,0.011303466558456422
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,8192,1536,0.008831999699274699
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,8192,1536,0.026055467128753663
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,1536,0.006442666550477346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,1536,0.009910399715105694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,8192,1024,0.025125332673390705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,8192,1024,0.007161599894364675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,10240,5120,0.024022400379180908
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,1024,0.00522986650466919
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,1024,0.008494933446248371
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,8192,768,0.024225066105524697
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,8192,768,0.006989866495132446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,768,0.004825599988301595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,10240,3584,0.01606186628341675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,768,0.008120533327261608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,8192,512,0.022776534159978233
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,8192,512,0.005894400179386139
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,512,0.004164266586303711
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,512,0.007401599983374278
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,8192,256,0.02240426739056905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,256,0.0036831999818483984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,256,0.006681600213050842
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,8192,128,0.02143146594365438
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,8192,128,0.00487253318230311
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,128,0.003387733300526937
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,128,0.006586666901906331
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,8192,64,0.021960532665252684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,64,0.003156266609827677
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,64,0.0065184002121289565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,8192,32,0.020426666736602782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,32,0.0031658666829268134
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,32,0.006528000036875407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,7168,65536,0.18079573313395184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,7168,16384,0.049021867911020915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,65536,0.18749653498331706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,7168,16384,0.07082666556040446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,7168,65536,0.21699946721394858
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,65536,0.3231040000915527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,16384,0.08506666819254557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,16384,0.05225386619567871
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,7168,12288,0.06312106847763062
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,12288,0.04143893321355184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,12288,0.07611306508382162
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,7168,10240,0.05465279817581177
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,7168,10240,0.032968533039093015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,10240,0.05585493246714274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,10240,0.03487893342971802
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,8192,0.047388799985249835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,7168,8192,0.04932693243026733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,7168,8192,0.027118933200836182
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,8192,0.028905600309371948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,7168,7168,0.055238401889801024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,7168,7168,0.024174932638804117
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,7168,0.041076266765594484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,7168,0.02612053354581197
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,7168,6144,0.05126399993896484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,7168,6144,0.02187946637471517
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,6144,0.03742826779683431
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,6144,0.021808000405629475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,7168,5120,0.040387201309204104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,5120,0.016731733083724977
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,5120,0.03134933312733968
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,7168,4096,0.03430399894714355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,4096,0.02350613276163737
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,4096,0.013192533453305563
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,7168,3584,0.03249280055363973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,7168,3584,0.013220266501108805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,3584,0.02105813423792521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,3584,0.012474667032559712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,7168,3072,0.01202453374862671
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,7168,3072,0.03191253344217936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,3072,0.01842133402824402
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,3072,0.011751466989517212
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,7168,2560,0.02959679961204529
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,7168,2560,0.010973866780598958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,2560,0.008713600039482117
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,2560,0.01074666678905487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,8192,7168,0.026125866174697875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,7168,2048,0.02807040015856425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,7168,2048,0.009108266234397889
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,2048,0.009563733140627544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,2048,0.007257600128650665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,7168,1536,0.026599466800689697
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,7168,1536,0.008344533046086629
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,1536,0.0062613333264986675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,1536,0.008763733506202697
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,7168,1024,0.024680533011754355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,7168,1024,0.006856533388296763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,1024,0.005145599941412607
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,1024,0.007816533247629803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,7168,768,0.023804799715677897
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,7168,768,0.006626133124033611
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,768,0.004541866481304169
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,768,0.007249066730340321
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,7168,512,0.023401600122451783
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,7168,512,0.005901866654555003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,512,0.004033066580692927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,512,0.006799999872843425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,7168,256,0.021721599499384563
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,256,0.0034656000634034475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,256,0.006237866481145223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,7168,128,0.021893332401911415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,128,0.0032373333970705668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,128,0.006018133461475372
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,7168,64,0.02207039992014567
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,64,0.0033461332321166994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,64,0.005983999868233999
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,7168,32,0.020195200045903524
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,32,0.0030933332939942675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,32,0.005924266576766968
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,6144,65536,0.168179194132487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,6144,65536,0.1964362621307373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,65536,0.176254940032959
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,65536,0.28143040339152015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,8192,256,0.0053269331653912864
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,6144,16384,0.04856746594111125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,6144,16384,0.06820480028788248
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,16384,0.07759040196736654
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,16384,0.04846400022506714
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,6144,12288,0.05777813196182251
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,6144,12288,0.03566933472951253
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,12288,0.0372383991877238
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,12288,0.057530665397644044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,6144,10240,0.050714667638142905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,6144,10240,0.030943999687830608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,10240,0.04771093527475993
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,10240,0.033336532115936277
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,6144,8192,0.046854400634765626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,8192,0.027065600951512652
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,8192,0.03925333420435588
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,6144,7168,0.051946667830149326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,7168,0.033986131350199386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,7168,0.021972266832987468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,6144,6144,0.04414186477661133
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,7168,12288,0.039450665314992264
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,6144,6144,0.01895573337872823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,6144,0.018467199802398682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,6144,0.030237867434819536
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,6144,5120,0.03765866756439209
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,5120,0.02542613347371419
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,5120,0.014781866470972696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,6144,4096,0.034723198413848876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,6144,4096,0.013474133610725404
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,4096,0.021600000063578286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,4096,0.012618666887283326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,6144,3584,0.033212800820668534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,6144,3584,0.01295253336429596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,3584,0.017892267306645712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,3584,0.011962667107582092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,7168,5120,0.017859200636545815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,6144,3072,0.03158400058746338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,6144,3072,0.011490133404731751
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,3072,0.009551999966303508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,3072,0.010626133282979329
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,7168,4096,0.014476799964904785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,6144,2560,0.02977493405342102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,6144,2560,0.01118933359781901
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,2560,0.008428800106048583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,2560,0.010043733318646749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,6144,2048,0.02802133361498515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,6144,2048,0.009144533673922222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,2048,0.007179733117421467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,2048,0.008914132912953694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,6144,1536,0.026472532749176027
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,6144,1536,0.00879253347714742
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,1536,0.006017066538333893
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,1536,0.008220799763997396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,6144,1024,0.024924800793329874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,6144,1024,0.006949333349863689
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,1024,0.004930133124192556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,1024,0.007353599866231282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,6144,768,0.023690666755040488
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,6144,768,0.0065760001540184024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,768,0.004412800073623657
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,768,0.0068906664848327635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,6144,512,0.02252053419748942
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,6144,512,0.005724800129731497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,512,0.003885866701602936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,512,0.006545066833496094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,6144,256,0.02153279980023702
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,6144,256,0.005201066533724466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,256,0.0034293333689371743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,256,0.0060586666067441305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,6144,128,0.021267199516296388
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,6144,128,0.004871466755867004
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,128,0.00322026660044988
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,128,0.005857066810131073
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,6144,64,0.021683200200398763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,64,0.003010133405526479
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,64,0.005854933460553487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,7168,256,0.00522986650466919
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,6144,32,0.020137600104014077
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,32,0.0030229332546393077
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,32,0.005924266576766968
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,7168,128,0.0047872001926104225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,5120,65536,0.14812374114990234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,5120,65536,0.1664896011352539
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,65536,0.23601706822713217
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,5120,16384,0.05986666679382324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,65536,0.13655145963033039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,5120,16384,0.04598613182703654
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,16384,0.05987199942270914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,16384,0.03788906733194987
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,5120,12288,0.05243200063705444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,12288,0.04633173147837321
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,5120,12288,0.034083199501037595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,12288,0.03019839922587077
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,5120,10240,0.04830933411916097
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,5120,10240,0.029676800966262816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,10240,0.03887573480606079
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,5120,8192,0.042080001036326094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,10240,0.02643306652704875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,5120,8192,0.023226666450500488
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,8192,0.031838933626810714
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,8192,0.01900906761487325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,5120,7168,0.04512106577555339
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,5120,7168,0.020593067010243736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,7168,0.02821439901987712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,7168,0.016422399878501893
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,5120,6144,0.03989866574605306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,6144,0.024922666947046916
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,5120,5120,0.03598293463389079
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,6144,0.014858667055765787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,5120,5120,0.015089066823323569
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,5120,0.02127679983774821
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,5120,0.012814933061599731
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,5120,4096,0.033801599343617754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,5120,4096,0.012743467092514038
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,4096,0.01700693368911743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,4096,0.011194666226704914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,5120,3584,0.03207040031750997
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,5120,3584,0.012651733557383218
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,3584,0.008850133419036866
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,6144,8192,0.025194666783014935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,3584,0.010551466544469198
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,5120,3072,0.030712532997131347
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,5120,3072,0.010945066809654236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,3072,0.007985066870848339
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,3072,0.009928533434867859
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,5120,2560,0.029073067506154376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,6144,7168,0.022711465756098427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,2560,0.007359999914964039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,2560,0.009160533547401428
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,5120,2048,0.027784534295399982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,5120,2048,0.008933333555857341
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,2048,0.006218666831652323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,2048,0.008641067147254943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,5120,1536,0.026128000020980834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,5120,1536,0.008178133269151051
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,1536,0.0054730668663978575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,1536,0.007772799829641979
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,6144,5120,0.01625599960486094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,5120,1024,0.02482879956563314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,5120,1024,0.006757333377997081
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,1024,0.004647466540336609
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,1024,0.007163733243942261
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,5120,768,0.023727999130884806
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,5120,768,0.006487466891606649
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,768,0.00417493333419164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,768,0.006771199901898702
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,5120,512,0.0226197342077891
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,5120,512,0.005560533205668131
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,512,0.003717333326737086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,512,0.00647680014371872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,5120,256,0.022275199492772423
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,5120,256,0.005092266698678335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,256,0.0034080001215140024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,256,0.006098133325576782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,5120,128,0.021226666371027627
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,5120,128,0.004668800036112467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,128,0.0031637333333492277
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,64,0.005820799867312113
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,128,0.0058218667904535925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,5120,32,0.020320000251134236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,5120,64,0.021670399109522502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,64,0.0029557332396507262
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,32,0.002998399982849757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,32,0.0058559998869895935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,4096,65536,0.1460266590118408
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,4096,65536,0.14763092994689941
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,65536,0.19473066329956054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,65536,0.10883413155873616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,4096,16384,0.05444373289744059
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,16384,0.032654933134714764
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,16384,0.05086400111516317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,4096,12288,0.04764480193456014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,4096,12288,0.030922667185465498
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,12288,0.0401802659034729
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,4096,10240,0.04322453339894612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,12288,0.026532266537348432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,4096,10240,0.02667733430862427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,10240,0.03432426850001018
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,10240,0.020615466435750327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,4096,8192,0.03329493403434754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,4096,8192,0.020844799280166627
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,8192,0.014174933234850565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,8192,0.028410667181015016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,4096,7168,0.039000535011291505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,4096,7168,0.018528000513712565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,7168,0.02576213280359904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,7168,0.013178666432698568
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,4096,6144,0.03647146622339885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,4096,6144,0.016851200660069784
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,6144,0.023009065786997476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,6144,0.012567466497421265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,4096,5120,0.03415893316268921
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,4096,5120,0.014755200346310934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,5120,0.017244799931844076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,5120,0.011171199878056844
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,4096,4096,0.031617067257563275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,4096,4096,0.012519466876983642
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,4096,0.008853333195050557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,4096,0.009790933132171631
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,4096,3584,0.03025493423144023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,4096,3584,0.012666666507720947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,3584,0.007783466577529907
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,3584,0.009264000256856282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,4096,3072,0.029036800066630047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,3072,0.0071370666225751235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,3072,0.008925867080688477
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,4096,2560,0.027879466613133747
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,4096,2560,0.01035520037015279
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,2560,0.006512000163396199
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,2560,0.008332799871762593
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,4096,2048,0.02686400016148885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,2048,0.005779199798901876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,2048,0.0077248002092043565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,4096,1536,0.025562665859858197
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,4096,1536,0.00806933343410492
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,1536,0.005026133358478546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,1536,0.007113599777221679
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,4096,1024,0.024282666047414143
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,5120,6144,0.017653334140777587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,4096,1024,0.006436266501744588
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,1024,0.004368000229199727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,1024,0.006695466736952464
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,4096,768,0.02337706685066223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,4096,768,0.006578133503595989
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,768,0.004058666775623957
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,768,0.00645653357108434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,4096,512,0.022299732764561972
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,4096,512,0.005505066613356272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,512,0.003673599908749262
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,512,0.006244266529877981
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,256,0.0033418667813142145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,256,0.0059562668204307554
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,4096,256,0.021496532360712688
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,4096,256,0.005041066805521647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,5120,2560,0.010685867071151734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,4096,128,0.02113173405329386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,4096,128,0.004693333307902018
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,128,0.00311253344019254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,128,0.00573333352804184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,4096,64,0.021620267629623414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,64,0.002942933390537898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,64,0.005760000149408976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,4096,32,0.020102399587631225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,32,0.0029663999875386557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,32,0.0057546665271123254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3584,65536,0.13370879491170246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3584,65536,0.13371094067891437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,65536,0.09822933673858643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3584,16384,0.051867731412251795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,65536,0.16502186457316081
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3584,16384,0.0367850661277771
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,16384,0.029845333099365233
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,16384,0.04566613435745239
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3584,12288,0.04611626863479614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,12288,0.02427306572596232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,12288,0.036907732486724854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3584,10240,0.03466560045878093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3584,10240,0.025179733832677204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,10240,0.031880533695220946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,10240,0.016360533237457276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3584,8192,0.03051946759223938
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3584,8192,0.019138133525848387
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,8192,0.025719465812047322
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,8192,0.013365333278973898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3584,7168,0.038777601718902585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3584,7168,0.01820053259531657
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,7168,0.023752532402674355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,7168,0.01313706636428833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3584,6144,0.036115201314290364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,6144,0.019595734278361
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,6144,0.011916800340016683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3584,5120,0.03407893180847168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3584,5120,0.014375467101732889
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,5120,0.01123413344224294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,5120,0.010611200332641601
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3584,4096,0.03174293239911397
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3584,4096,0.012342400352160136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,4096,0.008470400174458822
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,4096,0.009417600433031718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3584,3584,0.03057173291842143
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3584,3584,0.012406399846076966
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,3584,0.007840000092983246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,3584,0.008940800031026205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3584,3072,0.029384533564249676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,4096,16384,0.04019733270009358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3584,3072,0.010288000106811523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,3072,0.007110400001207988
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,3072,0.008494933446248371
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3584,2560,0.027960532903671266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,2560,0.006469333171844482
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3584,2560,0.009136000275611877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,2560,0.007985066870848339
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3584,2048,0.026843732595443724
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3584,2048,0.007982933521270752
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,2048,0.005703466633955637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,2048,0.007499733567237854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3584,1536,0.025764266649882
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3584,1536,0.0075434664885203045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,1536,0.005067733426888784
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,1536,0.007156266768773396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3584,1024,0.02411093314488729
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,1024,0.004369066655635833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3584,1024,0.006326400240262349
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,1024,0.006588799754778545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3584,768,0.02368746598561605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,768,0.004090666770935059
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,768,0.006388266881306966
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3584,512,0.022521599133809408
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3584,512,0.005548800031344095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,512,0.006206933160622915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,512,0.0036906667053699495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3584,256,0.021496532360712688
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3584,256,0.00499839981396993
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,256,0.0033439998825391137
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,256,0.005964800218741099
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,4096,3072,0.010546132922172546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,128,0.0031199999153614043
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3584,128,0.02141439914703369
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3584,128,0.004589866598447164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,128,0.005691733459631602
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3584,64,0.021642667055130006
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,64,0.0029343999922275543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,64,0.005734399954477946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3584,32,0.020045866568883262
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,32,0.002942933390537898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,32,0.005762133498986562
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,4096,2048,0.008475733796755473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3072,65536,0.12473066647847493
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3072,65536,0.1203221321105957
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,65536,0.08904213110605876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,65536,0.14636054039001464
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3072,16384,0.03808533350626628
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3072,16384,0.049060265223185226
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,16384,0.027328000466028853
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,16384,0.04072106679280599
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3072,12288,0.038260265191396074
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,12288,0.03198826710383097
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,12288,0.01799359917640686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3072,10240,0.03185813426971436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3072,10240,0.02278719941775004
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,10240,0.02782613237698873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,10240,0.015521066387494406
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3072,8192,0.030004266897837324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3072,8192,0.019447465737660728
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,8192,0.023060266176859537
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3072,7168,0.03868480126063029
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,8192,0.013285332918167114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3072,7168,0.01739733417828878
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,7168,0.02038080096244812
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,7168,0.012470400333404541
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3072,6144,0.036354132493336994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3072,6144,0.015774933497111003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,6144,0.011912533640861511
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,6144,0.01193173329035441
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3072,5120,0.033958399295806886
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3072,5120,0.01376426617304484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,5120,0.010466133554776508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,5120,0.010624000430107116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3072,4096,0.031640533606211343
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,4096,0.009147733449935913
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,4096,0.009480533003807069
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3072,3584,0.03038933277130127
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3072,3584,0.011794132987658183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,3584,0.008813866972923278
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,3584,0.009060266613960265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3072,3072,0.029234133164087933
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,3072,0.008250666658083598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,3072,0.008472533027331034
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3072,2560,0.028165332476298016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3584,12288,0.02983040014902751
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3072,2560,0.008980266253153483
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,2560,0.006451199948787689
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,2560,0.007962666451931
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3072,2048,0.026871466636657716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3072,2048,0.007820799946784973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,2048,0.005772800246874491
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,2048,0.0074325333038965866
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3072,1536,0.02508693337440491
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3072,1536,0.007477333148320516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,1536,0.005094400048255921
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,1536,0.007006933291753133
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3072,1024,0.023915733893712363
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3072,1024,0.006263466676076253
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,1024,0.004368000229199727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,1024,0.006646400193373363
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3072,768,0.023558400074640908
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3072,768,0.006043733159701029
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,768,0.00405973345041275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3584,6144,0.016380799810091655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,768,0.006439466774463653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3072,512,0.022422399123509726
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3072,512,0.005357866485913595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,512,0.003668266783157984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,512,0.006226133306821187
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3072,256,0.021385600169499717
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3072,256,0.004935466746489207
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,256,0.0033781332274278007
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,256,0.005869866907596588
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3072,128,0.004537599782148996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,128,0.005686399837334951
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3072,128,0.02139520049095154
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,128,0.0031285333136717476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3072,64,0.021603200833002725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,64,0.0029343999922275543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,64,0.005758933226267497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,3072,32,0.020091732343037925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,32,0.0029418667157491045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,32,0.005739733576774597
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2560,65536,0.11187840302785237
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2560,65536,0.10990293025970459
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,65536,0.1212661345799764
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,65536,0.08127466837565103
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2560,16384,0.04685013294219971
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2560,16384,0.03566293319066365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,16384,0.034392531712849936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,16384,0.024106667439142863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2560,12288,0.03777173360188802
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,12288,0.02754986683527629
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,12288,0.01618026693662008
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2560,10240,0.03338773250579834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2560,10240,0.02326719959576925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,10240,0.0237226665019989
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,10240,0.014210133751233419
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2560,8192,0.019963733355204263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2560,8192,0.02994133234024048
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,8192,0.01700693368911743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,8192,0.012562132875124612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3584,768,0.0062165334820747375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2560,7168,0.03833386500676473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2560,7168,0.016204800208409628
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,7168,0.011177600423494975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,7168,0.011703466375668842
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2560,6144,0.03577386538187663
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2560,6144,0.014465066790580749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,6144,0.010161067048708599
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,6144,0.010866133371988933
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2560,5120,0.033530668417612715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2560,5120,0.012619733810424805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,5120,0.009577600161234538
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,5120,0.010034132997194927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2560,4096,0.031454932689666745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2560,4096,0.010959999759991963
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,4096,0.008267733454704284
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,4096,0.008967467149098714
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2560,3584,0.030282666285832722
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2560,3584,0.010624000430107116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,3584,0.00782719999551773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,3584,0.00856213370958964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2560,3072,0.028709334135055543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2560,3072,0.009176533420880635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,3072,0.007433600227038066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,3072,0.008226133386294047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2560,2560,0.02794133424758911
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,2560,0.007810133198897044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2560,2560,0.008500267068545024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2560,2048,0.026476800441741943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,2560,0.006348800162474315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2560,2048,0.007439999779065449
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,2048,0.005681066711743673
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3072,12288,0.030184533198674517
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,2048,0.007288533449172974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2560,1536,0.025737599531809492
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2560,1536,0.007292800148328145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,1536,0.005003733436266581
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,1536,0.006921599805355072
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2560,1024,0.024190932512283325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2560,1024,0.006137600044409434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,1024,0.004316799839337667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,1024,0.006571733454863231
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2560,768,0.02362133264541626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,768,0.006331733365853627
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2560,768,0.006051200131575266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2560,512,0.005310933291912079
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,768,0.003986133386691412
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,512,0.006101333101590474
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2560,512,0.022512000799179078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2560,256,0.021363200743993123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,512,0.003647999962170919
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2560,256,0.004789333542188009
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,256,0.0032927999893824257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,256,0.005871999760468801
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2560,128,0.02112320065498352
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2560,128,0.004428799947102865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,128,0.0030847998956839246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,128,0.005659733215967814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2560,64,0.021477333704630532
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,64,0.0028768000503381092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3072,4096,0.011982933680216471
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,64,0.0057536001006762184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3072,3072,0.009989333152770997
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2560,32,0.02002453406651815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,32,0.00290133332212766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2048,65536,0.10285546779632568
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,32,0.005726933479309082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2048,65536,0.12138453324635823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,65536,0.09839786688486735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,65536,0.0739039977391561
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2048,16384,0.04075733423233032
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2048,16384,0.03129706581433614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,16384,0.028391466538111372
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,16384,0.01925546725591024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2048,12288,0.03397866487503052
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2048,12288,0.025204267104466754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,12288,0.022702932357788086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,12288,0.01565333306789398
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2048,10240,0.031754666566848756
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2048,10240,0.02172693411509196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,10240,0.017077332735061644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,10240,0.014089600245157877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2048,8192,0.030024532477060956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,8192,0.009641599655151368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,8192,0.012316800157229106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2048,7168,0.038165334860483804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,7168,0.009297066926956176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,7168,0.011500799655914306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2048,6144,0.03583999872207642
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2048,6144,0.014806399742762247
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,6144,0.008659199873606364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,6144,0.010674132903416952
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2048,5120,0.03386773268381755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2048,5120,0.013192533453305563
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2048,4096,0.01113813320795695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,5120,0.0084906667470932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,5120,0.009852799773216247
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2048,4096,0.0315775990486145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,4096,0.007718400160471599
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,4096,0.008801066875457763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2048,3584,0.030504532655080158
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2048,3584,0.011628799637158712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,3584,0.007501866420110066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,3072,0.007029333213965099
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,3584,0.008414933085441589
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,3072,0.008036266764005024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2048,3072,0.029266132911046343
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2048,3072,0.009640533725420635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2048,2560,0.02810346682866414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2048,2560,0.00913813312848409
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,2560,0.006297599772612255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,2560,0.0076447998483975725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2048,2048,0.02672426700592041
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2048,2048,0.007852800190448761
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,2048,0.005578666428724925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2560,12288,0.027592533826828004
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,2048,0.007210666437943776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2048,1536,0.024936532974243163
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,1536,0.00489386667807897
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,1536,0.0067552000284194945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2048,1024,0.02395840088526408
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2048,1024,0.0062496001521746315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,1024,0.004233600199222564
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,1024,0.006397866706053417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2048,768,0.023169066508611044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2048,768,0.006265600025653839
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,768,0.003945599993069967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,768,0.0062165334820747375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2048,512,0.02229866584142049
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2048,512,0.005387733379999796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,512,0.003576533248027166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,512,0.0060479998588562015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,256,0.003239466746648153
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2048,256,0.021413334210713706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2048,256,0.004982399940490723
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,256,0.005804799993832906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2048,128,0.021121066808700562
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2048,128,0.004600533346335093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,128,0.003005866706371307
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,128,0.005643733342488607
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2048,64,0.021565866470336915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,64,0.0028949332733949023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,64,0.0056415999929110205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,2048,32,0.01999146739641825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,32,0.002919466545184453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,32,0.0056639999151229855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1536,65536,0.0955946683883667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,65536,0.0736074686050415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,65536,0.06848533153533935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1536,16384,0.03720106681187947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1536,16384,0.029585067431132
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,16384,0.022881066799163817
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,16384,0.01910933256149292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1536,12288,0.033769599596659344
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1536,12288,0.024153600136439003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,12288,0.01188053290049235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,12288,0.015723733107248943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1536,10240,0.03178240060806274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1536,10240,0.020734934012095134
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,10240,0.010959999759991963
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,10240,0.01409173309803009
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1536,8192,0.029973334074020384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1536,8192,0.017767467101415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,8192,0.009278933207194012
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,8192,0.012268799543380737
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1536,7168,0.03836479981740316
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1536,7168,0.016437333822250367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,7168,0.00890773336092631
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,7168,0.011550933122634888
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1536,6144,0.03581759929656982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1536,6144,0.014498133460680643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,6144,0.008392533659934998
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1536,5120,0.0335264007250468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,6144,0.010692266623179118
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1536,5120,0.012795733412106833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,5120,0.007968000074227651
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,5120,0.009965866804122925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1536,4096,0.03164693315823873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,4096,0.007147733370463054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,4096,0.008876799543698629
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1536,3584,0.03049386739730835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1536,3584,0.01079360047976176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,3584,0.007493333518505096
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,3584,0.00841813286145528
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1536,3072,0.02884693344434102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1536,3072,0.009340799848238627
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,3072,0.007080533107121785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,3072,0.008017066617806752
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1536,2560,0.02791253328323364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1536,2560,0.008289066453774769
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,2560,0.0062389334042867025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,2560,0.007732266684373219
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1536,2048,0.026743467648824053
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,2048,0.005595733225345611
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,2048,0.007223466535409291
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1536,1536,0.025242666403452556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1536,1536,0.007008000214894612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,1536,0.00490880012512207
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,1536,0.006839466591676076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1536,1024,0.02405440012613932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1536,1024,0.0060032000144322716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,1024,0.004223999877770742
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,1024,0.00653653343518575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1536,768,0.023525333404541014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1536,768,0.0058335999647776285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,768,0.003909333298603693
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,768,0.006263466676076253
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1536,512,0.022296533981959025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1536,512,0.0053610667586326596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,512,0.003748266647259394
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,512,0.006077866752942403
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1536,256,0.021441066265106203
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1536,256,0.004802133142948151
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2048,8192,0.018609066804250084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,256,0.0032757334411144257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2048,7168,0.01662506659825643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,256,0.005866666634877523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1536,128,0.021197867393493653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1536,128,0.004519466559092204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,128,0.0030271999537944795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,128,0.005723733206590017
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1536,64,0.021657600005467733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,64,0.0028949332733949023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,64,0.005702400207519531
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1536,32,0.020163200298945107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,32,0.0028736000259717304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,32,0.0057205334305763245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1024,65536,0.09191466967264811
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1024,65536,0.09257493019104004
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,65536,0.05138453245162964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,65536,0.06682133674621582
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1024,16384,0.037666134039560956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1024,16384,0.02876053253809611
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,16384,0.01058666706085205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,16384,0.019144533077875774
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1024,12288,0.033638401826222734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1024,12288,0.022507733106613158
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,12288,0.009194667140642803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,12288,0.015596800049146018
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1024,10240,0.03201066652933757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1024,10240,0.01928640007972717
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,10240,0.008376533786455791
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,10240,0.014141866564750671
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1024,8192,0.030271999041239422
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1024,8192,0.0172106663386027
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,8192,0.007799466451009114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,8192,0.012243200341860454
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1024,7168,0.03830186525980632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1024,7168,0.016106667121251424
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,7168,0.007461333274841308
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2048,1536,0.007577600081761678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,7168,0.011433600385983785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1024,6144,0.03599466482798259
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1024,6144,0.013559466600418091
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,6144,0.006925866504510244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,6144,0.010629333058993022
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1024,5120,0.03376106818517049
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1024,5120,0.011851732929547627
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,5120,0.007287466526031494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,5120,0.009803733229637146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1024,4096,0.03150720000267029
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1024,4096,0.010188800096511842
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,4096,0.00688213308652242
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,4096,0.008746666709582011
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1024,3584,0.030470399061838786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1024,3584,0.009769599636395771
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,3584,0.00729066679875056
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,3584,0.008373333017031352
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,3072,0.007062399884064992
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1024,3072,0.02913813392321269
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1024,3072,0.008658132950464885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,3072,0.00795839975277583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,2560,0.00624533345301946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1024,2560,0.027721599737803145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,2560,0.00758186678091685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1024,2048,0.006986666719118755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1024,2048,0.026686932643254595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,2048,0.005610666672388713
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,2048,0.007098666826883952
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1024,1536,0.02540053327878316
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1024,1536,0.006353066861629486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,1536,0.004920533299446106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,1536,0.006758399804433187
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1024,1024,0.023861332734425863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1024,1024,0.005589333176612854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,1024,0.004228266576925913
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,1024,0.00631573349237442
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1024,768,0.023490132888158162
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1536,65536,0.10719467004140218
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1024,768,0.005621333420276642
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,768,0.003924266745646795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,768,0.006137600044409434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1024,512,0.022296533981959025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1024,512,0.005086933573087057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,512,0.0035487999518712364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,512,0.0059232001503308615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1024,256,0.02136639952659607
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1024,256,0.004817066589991251
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,256,0.0032757334411144257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,256,0.005817600091298421
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1024,128,0.021022933721542358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1024,128,0.004341333111127218
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,128,0.003058133274316788
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,128,0.005634133517742157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1024,64,0.0216213325659434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,64,0.0028757333755493166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,32,0.005682133138179779
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,64,0.005624533196290334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,1024,32,0.020060799519220986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,32,0.0028746667007605235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,768,65536,0.08961386680603027
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,768,65536,0.07916159629821777
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,65536,0.03998719851175944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,65536,0.06560106674830118
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,768,16384,0.037058134873708085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1536,4096,0.010984533031781514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,16384,0.009530666470527648
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,16384,0.018998400370279948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,768,12288,0.03377493222554524
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,768,12288,0.017375999689102174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,12288,0.008564266562461852
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,12288,0.015639467040697734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,768,10240,0.03179840048154195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,768,10240,0.015753600001335143
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,10240,0.0081045334537824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,10240,0.014108799894650779
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,768,8192,0.030308266480763752
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,8192,0.007448533177375793
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,8192,0.012227200468381246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,768,7168,0.03817066748936971
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,768,7168,0.013370666901270548
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,7168,0.007020799815654755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,7168,0.01144426663716634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1536,2048,0.007606400052706401
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,768,6144,0.03592106501261393
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,768,6144,0.01214400033156077
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,6144,0.006810666620731353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,6144,0.01072746713956197
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,768,5120,0.03391679922739665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,768,5120,0.01042560040950775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,5120,0.007272533575693766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,5120,0.009979733824729919
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,768,4096,0.031573333342870075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,768,4096,0.008892800410588582
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,4096,0.006849066913127899
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,4096,0.008880000313123066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,768,3584,0.03053119977315267
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,768,3584,0.008701866865158081
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,3584,0.007209600011507671
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,3584,0.008505599697430928
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,768,3072,0.0291978657245636
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,768,3072,0.007684266567230225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,3072,0.006923733154932658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,3072,0.008099199831485748
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,768,2560,0.027837866544723512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,2560,0.00621973325808843
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,2560,0.0076906666159629825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,768,2048,0.02666986584663391
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,768,2048,0.006436266501744588
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,2048,0.0056085333228111265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,2048,0.007227733234564463
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,768,1536,0.02526400089263916
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,768,1536,0.0060703997810681665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,1536,0.004910933474699656
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,1536,0.006828799843788147
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,768,1024,0.023745065927505492
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,768,1024,0.005468800167242686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,1024,0.004189866781234741
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,1024,0.006446933249632518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,768,768,0.005212800204753875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,768,768,0.023332266012827556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,768,0.0038442666331926978
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,768,0.006260266900062561
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,768,512,0.022629332542419434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,768,512,0.005070933202902476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,512,0.0035573333501815797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,512,0.006039466460545858
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,768,256,0.02210879921913147
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,768,256,0.004642133414745331
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,256,0.003230933348337809
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,256,0.005794133245944977
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,128,0.0032266666491826378
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,768,128,0.021345067024230956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,768,128,0.0044725333650906885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,128,0.005667200187842051
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,768,64,0.02146986722946167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,64,0.0028597332537174227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,64,0.005700266857941946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,768,32,0.019972266753514607
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,32,0.0028618666032950084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,32,0.005716266731421152
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,512,65536,0.08631253242492676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,512,65536,0.06749973297119141
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,65536,0.02925119996070862
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,65536,0.06444799900054932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,512,16384,0.03722026745478312
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,512,16384,0.019382399320602418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,16384,0.007974400122960409
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,16384,0.019107200702031455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,512,12288,0.03361279964447021
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,512,12288,0.015753600001335143
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,12288,0.007294933497905731
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,12288,0.015608533223470052
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,512,10240,0.03177280028661092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,512,10240,0.013882666826248169
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,10240,0.00719893326361974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,10240,0.014108799894650779
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,512,8192,0.03002133369445801
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,512,8192,0.011907200018564861
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,8192,0.00701333334048589
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1024,2560,0.007919999957084655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,8192,0.012235732873280843
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,512,7168,0.037853864828745525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,512,7168,0.011029332876205444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,7168,0.006906666855017345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,7168,0.011547733346621196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,512,6144,0.036134401957194015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,512,6144,0.010121599833170573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,6144,0.006748799979686737
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,6144,0.010550399621327717
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,512,5120,0.03379840056101481
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,512,5120,0.009218133489290873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,5120,0.007157333195209503
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,5120,0.009946667154630025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,512,4096,0.03138773242632548
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,512,4096,0.008224000036716462
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,4096,0.006749866902828217
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,4096,0.008842666943868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,512,3584,0.03015786608060201
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,512,3584,0.0077450667818387345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,3584,0.007368533313274384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,3584,0.008461866776148479
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,512,3072,0.028677332401275634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,512,3072,0.007222400108973186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,3072,0.006937600175539653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,3072,0.008014933268229166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,512,2560,0.02762453357378642
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,512,2560,0.006727466483910878
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,2560,0.006275199850400289
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,2560,0.007650133470694225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,512,2048,0.026284799973169966
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,2048,0.005627733469009399
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,2048,0.0072405333320299785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,512,1536,0.02528746724128723
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,1536,0.004915200173854828
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,1536,0.006781866649786632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,512,1024,0.023834667603174844
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,512,1024,0.005352533360322317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,1024,0.004221866528193155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,1024,0.006353066861629486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,512,768,0.023268266518910726
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,512,768,0.005179733534653981
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,768,0.003857066730658213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,768,0.0061951999862988796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,512,512,0.02209279934565226
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,512,512,0.004936533172925314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,512,0.0035648000737031303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,512,0.006040533383687338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,512,256,0.021270400285720824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,512,256,0.004673066735267639
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,256,0.003193599979082743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,256,0.005747200051943461
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,512,128,0.021185066302617392
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,512,128,0.0044149334232012425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,768,16384,0.02323946754137675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,128,0.0030037333567937215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,128,0.005658666789531708
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,512,64,0.02137493292490641
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,64,0.005637333293755849
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,64,0.0029088000456492106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,512,32,0.019825067122777304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,32,0.0028864001234372456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,32,0.0056533331672350565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,256,65536,0.0658570647239685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,256,65536,0.08350826899210612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,65536,0.011399466792742412
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,256,16384,0.03707520167032878
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,65536,0.059256533781687416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,256,16384,0.01944640080134074
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,768,8192,0.014863999684651694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,16384,0.007012266914049785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,16384,0.01889386574427287
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,256,12288,0.03364906708399455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,256,12288,0.015914666652679443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,12288,0.006816000243028005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,12288,0.01551040013631185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,256,10240,0.03158400058746338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,256,10240,0.013766400019327798
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,10240,0.007119999825954437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,10240,0.013874133427937826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,256,8192,0.02988693316777547
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,256,8192,0.01192639966805776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,8192,0.00695253312587738
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,8192,0.0121888001759847
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,256,7168,0.03808000087738037
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,256,7168,0.011043199896812439
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,7168,0.006810666620731353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,7168,0.011326932907104492
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,256,6144,0.03621973196665446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,6144,0.006770133475462596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,6144,0.010470400253931682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,5120,0.007177599767843883
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,768,2560,0.007283199826876323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,256,5120,0.03385173479715983
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,256,5120,0.00918826659520467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,5120,0.00976746678352356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,256,4096,0.031142399708429976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,256,4096,0.008166400094827015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,4096,0.006716800232728322
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,4096,0.008770133058230083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,256,3584,0.030210133393605548
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,3584,0.007141333321730297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,3584,0.008309333523114523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,256,3072,0.029181865851084392
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,256,3072,0.007256533205509186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,3072,0.0068234667181968685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,3072,0.00798826664686203
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,256,2560,0.02796906630198161
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,256,2560,0.0067658667763074234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,2560,0.006233599781990051
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,2560,0.007541333138942718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,256,2048,0.026732800404230754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,256,2048,0.0062720000743865965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,2048,0.0055285334587097164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,2048,0.0070602665344874065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,256,1536,0.0257749338944753
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,256,1536,0.005888000130653381
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,1536,0.0048767998814582825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,1536,0.006673066814740498
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,256,1024,0.024312533934911094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,256,1024,0.005182933310667674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,1024,0.00421973317861557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,1024,0.0064085334539413456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,256,768,0.023306665817896526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,256,768,0.005117866893609365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,768,0.003916800022125244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,768,0.006075733403364817
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,256,512,0.022101332743962608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,512,0.0035082665582497918
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,256,512,0.004828799764315287
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,512,0.00590826670328776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,256,256,0.021261866887410483
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,256,256,0.004609066744645437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,256,0.0031360000371932983
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,256,0.005670399963855743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,256,128,0.021319466829299926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,256,128,0.004363733530044556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,128,0.0029525332152843474
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,128,0.005662933488686879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,256,64,0.02145386735598246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,64,0.0028266665836175283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,64,0.0056757330894470215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,256,32,0.01996799906094869
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,32,0.0028319999575614927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,32,0.0056415999929110205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,128,65536,0.08151573340098063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,128,65536,0.06548053423563639
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,65536,0.008686932921409606
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,65536,0.05921386480331421
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,128,16384,0.03713066577911377
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,16384,0.00681386689345042
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,16384,0.018772266308466592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,128,12288,0.03356373310089111
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,128,12288,0.01563093364238739
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,12288,0.006717866659164429
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,12288,0.015406933426856995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,128,10240,0.03173866669336955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,128,10240,0.013678933183352152
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,10240,0.006844800213972728
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,10240,0.013863466183344521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,128,8192,0.030054400364557903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,128,8192,0.011834667126337687
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,8192,0.006750933329264323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,8192,0.012052266796429952
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,128,7168,0.03780373334884644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,7168,0.006637866795063019
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,7168,0.011282133062680562
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,128,6144,0.035547733306884766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,128,6144,0.01000426709651947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,6144,0.006483200192451477
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,6144,0.01051093339920044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,128,5120,0.03373866478602092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,128,5120,0.009091200431187947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,5120,0.0068351998925209045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,5120,0.009781333804130554
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,128,4096,0.031309866905212404
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,128,4096,0.00811413327852885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,4096,0.006431999802589417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,4096,0.008684800068537394
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,128,3584,0.030202666918436687
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,128,3584,0.007663999994595845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,3584,0.006874666611353557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,3584,0.008332799871762593
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,128,3072,0.028816000620524092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,128,3072,0.007231999933719635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,3072,0.00658133327960968
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,3072,0.007910400132338206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,128,2560,0.027720532814661664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,128,2560,0.006713599960009258
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,2560,0.005921066800753275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,512,2048,0.006281599899133046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,2560,0.007501866420110066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,128,2048,0.026246400674184163
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,128,2048,0.006190933287143707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,2048,0.005299200117588043
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,2048,0.007011199990908305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,128,1536,0.025290666023890178
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,128,1536,0.005806933343410492
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,1536,0.004690133531888326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,1536,0.006684799989064534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,128,1024,0.02386346658070882
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,128,1024,0.005206400156021118
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,1024,0.004019200056791306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,1024,0.0062837332487106325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,128,768,0.023221333821614586
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,128,768,0.004941866795221964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,768,0.003705599904060364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,768,0.006064000229040781
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,512,1536,0.005963733295599619
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,128,512,0.022216532627741495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,128,512,0.004717866579691568
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,512,0.003425066669782003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,512,0.005892266829808554
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,128,256,0.021294933557510377
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,128,256,0.004543999830881754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,256,0.0030847998956839246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,128,128,0.004278400043646494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,256,0.005718400080998739
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,128,128,0.02087999979654948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,128,0.0029397333661715193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,128,0.005562666555245717
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,128,64,0.021504000822703043
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,64,0.005570133527119955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,64,0.002791466563940048
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,128,32,0.01995519995689392
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,32,0.002829866607983907
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,32,0.005543466905752818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,64,65536,0.08131306966145833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,65536,0.008846933643023174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,65536,0.059545600414276124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,64,16384,0.037197868029276535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,16384,0.006658133367697399
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,16384,0.018889600038528444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,64,12288,0.0336298664410909
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,12288,0.006588799754778545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,12288,0.01544319987297058
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,64,10240,0.03179413278897603
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,10240,0.00678719977537791
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,10240,0.013894400000572205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,64,8192,0.02993706663449605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,8192,0.006715733309586842
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,8192,0.012074666221936543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,64,7168,0.03828266859054565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,7168,0.006548266609509785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,7168,0.011303466558456422
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,64,6144,0.035870933532714845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,6144,0.006434133152167003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,6144,0.010492799679438274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,64,5120,0.033521068096160886
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,5120,0.009709866841634114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,5120,0.006854400038719177
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,64,4096,0.0313589334487915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,4096,0.006363733112812043
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,4096,0.008726400136947633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,64,3584,0.030344533920288085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,3584,0.006836266815662384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,3584,0.008335999647776286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,64,3072,0.029002666473388672
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,3072,0.0065653334061304735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,3072,0.007846400141716003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,64,2560,0.028018132845560713
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,2560,0.005789866546789805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,2560,0.007526400188604991
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,64,2048,0.026348799467086792
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,2048,0.005284266670544943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,2048,0.007047466437021891
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,64,1536,0.025240532557169598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,1536,0.0046528001626332605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,1536,0.006911999980608622
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,1024,0.00626986672480901
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,64,1024,0.023836799462636313
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,1024,0.0039594667653242745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,768,0.0036661334335803984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,64,768,0.02327679991722107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,768,0.006084266801675161
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,64,512,0.02204906741778056
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,512,0.0033471999069054925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,256,6144,0.010011733571688334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,256,0.005656533439954122
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,512,0.0059114664793014525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,64,256,0.021293866634368896
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,256,0.0030730667213598887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,64,128,0.02101759910583496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,128,0.00290133332212766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,128,0.005569066603978475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,64,64,0.021501866976420085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,64,0.0027317332724730173
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,64,0.00550186683734258
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,64,32,0.019783467054367065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,256,3584,0.007771733403205872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,32,0.0027872001131375628
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,32,0.0056096002459526065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,32,65536,0.08181866804758707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,65536,0.008905599514643352
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,65536,0.059531732400258386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,32,16384,0.037036800384521486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,16384,0.006682666639486949
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,16384,0.01881600022315979
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,32,12288,0.033625598748524985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,12288,0.00652159998814265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,12288,0.015428266922632852
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,8192,0.006577066580454509
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,32,10240,0.03161173264185588
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,10240,0.006742399930953979
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,10240,0.013851733009020487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,32,8192,0.030044800043106078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,8192,0.012049067020416259
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,32,7168,0.03793066740036011
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,7168,0.0065087998906771345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,7168,0.011386666695276897
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,32,6144,0.0358570655186971
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,6144,0.006356266637643178
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,6144,0.010539733370145162
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,32,5120,0.03376106818517049
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,5120,0.006774400174617767
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,5120,0.009741866588592529
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,32,4096,0.031404799222946166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,4096,0.0063733334342638654
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,4096,0.008654933174451191
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,32,3584,0.030384000142415362
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,3584,0.0068245331446329755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,3584,0.008266666531562805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,32,3072,0.028938666979471846
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,3072,0.006676266590754191
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,3072,0.007831466694672901
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,32,2560,0.027502934137980144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,2560,0.0058378666639328
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,2560,0.007501866420110066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,32,2048,0.026266666253407796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,2048,0.005196799834569296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,2048,0.0070933332045873005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,32,1536,0.025417600075403852
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,1536,0.004631466666857402
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,1536,0.006615466872851054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,32,1024,0.023884799083073935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,1024,0.0039893334110577905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,1024,0.006287999947865804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,32,768,0.02334400018056234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,768,0.0036618667344252265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,768,0.006033066908518473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,32,512,0.022011733055114745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,256,0.0030570665995279947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,512,0.0033791999022165934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,512,0.005850666761398315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,32,256,0.02109973430633545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,256,0.005659733215967814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,32,128,0.02102186679840088
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,128,0.0029205332199732465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,64,0.005513600011666616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,128,0.005547733108202616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,128,16384,0.019448532660802206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,32,64,0.02130133310953776
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,64,0.002792533238728841
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,128,7168,0.010913067062695821
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,8,32,32,0.019989333550135293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,32,0.0027093333502610523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,32,0.005549866457780202
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,65536,16384,0.3931253433227539
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,65536,16384,0.3838122685750326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,65536,12288,0.27259305318196614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,65536,12288,0.2993194580078125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,65536,16384,0.42032534281412764
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,65536,16384,0.8034581502278646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,65536,12288,0.5661098480224609
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,65536,12288,0.35234025319417317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,65536,10240,0.22672425905863441
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,65536,10240,0.24853439331054689
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,65536,10240,0.30323521296183265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,65536,10240,0.4714186668395996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,65536,8192,0.2230613390604655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,65536,8192,0.21750826835632325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,65536,7168,0.16199572881062824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,65536,8192,0.37745065689086915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,65536,7168,0.20461014111836753
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,65536,7168,0.17913494110107422
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,65536,7168,0.3292672157287598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,65536,6144,0.138701868057251
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,65536,6144,0.1782719930013021
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,65536,6144,0.1488789399464925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,65536,6144,0.28460054397583007
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,65536,5120,0.11877439816792805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,65536,5120,0.15260906219482423
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,65536,5120,0.1254208008448283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,65536,5120,0.23673280080159506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,65536,4096,0.09420693715413411
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,65536,4096,0.1335327943166097
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,65536,4096,0.10477333068847657
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,65536,4096,0.1978069305419922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,65536,3584,0.0856106678644816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,65536,3584,0.11860053539276123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,65536,3584,0.09071146647135417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,65536,3584,0.16744960149129232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,65536,3072,0.10437546571095783
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,65536,3072,0.14416213035583497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,65536,3072,0.07739840348561605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,65536,2560,0.12159466743469238
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,65536,2560,0.06394879817962647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,65536,2560,0.08788373470306396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,65536,2560,0.07828906377156576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,65536,2048,0.0793503999710083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,65536,2048,0.050885331630706784
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,65536,2048,0.09666133721669515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,65536,2048,0.056094932556152347
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,65536,1536,0.06763306458791098
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,65536,1536,0.04360746542612712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,65536,1536,0.07448639869689941
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,65536,1536,0.042743468284606935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,65536,1024,0.05423893531163534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,65536,768,0.05016533136367798
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,65536,1024,0.050494933128356935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,65536,768,0.03904213507970174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,65536,1024,0.03354133367538452
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,65536,768,0.02376533349355062
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,65536,768,0.02951786716779073
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,65536,512,0.03848106861114502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,65536,512,0.014150399963061014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,65536,512,0.016999467213948568
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,65536,512,0.03110506733258565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,65536,256,0.031863466898600264
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,65536,256,0.007621333499749501
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,65536,256,0.013834666212399802
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,65536,128,0.03136000037193298
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,65536,128,0.007448533177375793
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,65536,128,0.00559146652619044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,65536,128,0.011361066500345867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,65536,64,0.029819732904434203
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,65536,64,0.006372266511122386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,65536,64,0.011322666207949321
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,65536,32,0.02871146599451701
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,65536,32,0.0066101332505544026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,65536,32,0.01123413344224294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,16384,65536,0.4076053301493327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,16384,65536,0.4677589416503906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,65536,0.8003434499104818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,65536,0.45284481048583985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,16384,16384,0.10165867010752361
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,16384,16384,0.1382143974304199
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,16384,0.11662399768829346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,16384,0.2052618662516276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,16384,12288,0.11109333038330078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,16384,12288,0.07848959763844808
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,12288,0.0908128023147583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,12288,0.14259947141011556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,16384,10240,0.09241920312245687
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,10240,0.06602346499760946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,10240,0.12586987018585205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,16384,8192,0.07976746559143066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,16384,8192,0.050596264998118076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,8192,0.10179839928944905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,8192,0.05921386480331421
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,16384,7168,0.07749546368916829
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,16384,7168,0.04851839939753215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,7168,0.08285120328267416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,7168,0.05129813353220621
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,16384,6144,0.07254186471303305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,16384,6144,0.03960106770197551
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,6144,0.07247040271759034
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,65536,8192,0.18480532964070637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,6144,0.047277867794036865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,16384,5120,0.034068266550699874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,16384,5120,0.06280853350957236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,5120,0.060697599252065026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,5120,0.04118186632792155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,16384,4096,0.05605973402659098
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,4096,0.049080534776051836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,16384,4096,0.028457599878311157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,4096,0.03252373337745666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,16384,3584,0.05356906652450562
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,3584,0.04370773235956828
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,3584,0.029167999823888142
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,16384,3072,0.05015679995218912
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,16384,3072,0.023027199506759643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,3072,0.03770986795425415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,3072,0.025553067525227863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,16384,2560,0.04678506851196289
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,16384,2560,0.021861332654953002
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,2560,0.034127998352050784
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,2560,0.01923840045928955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,16384,2048,0.03342613379160563
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,16384,2048,0.012474667032559712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,65536,3072,0.07896320025126138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,2048,0.026235733429590864
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,2048,0.016173866391181946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,16384,1536,0.03051519989967346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,16384,1536,0.012313600381215413
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,1536,0.020594133933385213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,1536,0.013589333494504294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,16384,1024,0.028074665864308672
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,16384,1024,0.008746666709582011
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,1024,0.00809386670589447
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,16384,768,0.008028799792130788
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,1024,0.011378133296966552
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,16384,768,0.027112533648808796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,768,0.0066549330949783325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,65536,1024,0.028883200883865357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,768,0.010282666484514872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,16384,512,0.025245867172876996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,16384,512,0.006785066425800323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,512,0.005453866720199585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,512,0.00883733332157135
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,16384,256,0.024196267127990723
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,16384,256,0.005676800012588501
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,256,0.004261333247025808
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,256,0.0076789334416389465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,16384,128,0.02439360022544861
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,65536,256,0.009614933530489604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,16384,128,0.00510506679614385
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,128,0.0036778666079044344
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,128,0.0071829333901405334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,16384,64,0.022829866409301756
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,64,0.00323840007185936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,64,0.007149866720040639
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,16384,32,0.021643733978271483
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,32,0.003337600082159042
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,32,0.0072736000021298725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,65536,0.305075200398763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,12288,16384,0.08666240374247233
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,12288,65536,0.29989760716756186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,12288,16384,0.10500693321228027
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,65536,0.5641994476318359
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,16384,0.15775146484375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,12288,65536,0.32115093866984046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,16384,0.08120853106180827
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,12288,12288,0.08674346605936686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,12288,0.10753066539764404
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,12288,0.07197653452555339
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,12288,10240,0.07337173620859781
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,12288,10240,0.05571626822153727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,10240,0.08840533097585043
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,10240,0.05366933345794678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,12288,8192,0.04509226481119792
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,12288,8192,0.06457066535949707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,8192,0.07104213237762451
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,8192,0.04357440074284871
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,12288,7168,0.06338880062103272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,7168,0.06231573422749838
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,7168,0.03944746653238933
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,12288,6144,0.05841279824574789
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,12288,6144,0.03549226522445679
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,6144,0.05407786766688029
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,6144,0.0345578670501709
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,12288,5120,0.030949334303538006
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,12288,5120,0.05266773303349813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,5120,0.045765332380930585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,5120,0.03036160071690877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,12288,4096,0.04749759833017985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,12288,4096,0.025225599606831867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,4096,0.03676480054855347
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,4096,0.02480106751124064
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,16384,10240,0.06228586832682291
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,12288,3584,0.044086400667826334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,3584,0.033565866947174075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,3584,0.022452267011006673
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,12288,3072,0.01892906626065572
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,12288,3072,0.03412160078684489
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,3072,0.029345067342122395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,3072,0.01609386702378591
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,12288,2560,0.03011946678161621
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,12288,2560,0.016898133357365928
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,2560,0.024652800957361855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,2560,0.013896532853444419
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,12288,2048,0.013719466328620911
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,12288,2048,0.028689066569010418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,2048,0.020300799608230592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,2048,0.013115732868512472
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,1536,0.010462933778762817
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,12288,1536,0.027353600660959883
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,1536,0.011356799801190694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,12288,1024,0.0247871994972229
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,12288,1024,0.008410666386286418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,1024,0.008076799909273784
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,1024,0.009413333733876546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,12288,768,0.024576000372568765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,16384,3584,0.026096000274022417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,12288,768,0.0078015998005867004
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,768,0.0069248000780741375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,768,0.008596266309420269
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,12288,512,0.006412800153096516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,12288,512,0.022900267442067464
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,512,0.005468800167242686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,512,0.0077802668015162155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,12288,256,0.021733333667119346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,256,0.004134399940570196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,256,0.006925866504510244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,12288,128,0.022127999862035116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,128,0.003737599899371465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,128,0.0065749332308769224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,12288,128,0.005015466610590617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,12288,64,0.020923733711242676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,64,0.0033344000577926634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,64,0.006557866434256236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,12288,32,0.01971199909845988
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,32,0.0033461332321166994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,32,0.006569600105285645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,10240,65536,0.2415466626485189
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,10240,65536,0.3094122568766276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,65536,0.4603455861409505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,65536,0.2609525362650553
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,10240,16384,0.07298986911773682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,10240,16384,0.09339626630147299
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,16384,0.11729493141174316
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,16384,0.06965546607971192
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,10240,12288,0.050670933723449704
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,10240,12288,0.08134506543477377
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,12288,0.08975466887156168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,12288,0.053875199953715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,10240,10240,0.04234133164087932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,10240,10240,0.06641813516616821
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,10240,0.07513919671376547
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,10240,0.046778666973114016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,10240,8192,0.05853013197580973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,10240,8192,0.03418773412704468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,8192,0.06084479888280233
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,8192,0.03741759856541951
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,10240,7168,0.05797546704610189
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,10240,7168,0.03153813282648722
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,7168,0.05518399874369303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,7168,0.0344490647315979
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,10240,6144,0.05298133293787638
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,10240,6144,0.027661865949630736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,6144,0.047203199068705244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,6144,0.0303872009118398
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,10240,5120,0.048750933011372885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,10240,5120,0.024016000827153525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,12288,12288,0.06483946641286215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,5120,0.039267198244730635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,5120,0.02684906721115112
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,10240,4096,0.01970133384068807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,10240,4096,0.04175573190053304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,4096,0.03663253386815389
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,4096,0.019375999768575035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,10240,3584,0.0365941325823466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,10240,3584,0.017230933904647826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,3584,0.028551467259724933
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,3584,0.015964800119400026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,10240,3072,0.030851199229558306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,10240,3072,0.014847999811172486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,3072,0.024991999069849648
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,3072,0.014045866330464682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,12288,7168,0.04036906560262044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,10240,2560,0.029677865902582805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,2560,0.0212501327196757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,2560,0.012894933422406515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,10240,2048,0.028355199098587035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,10240,2048,0.010600533088048298
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,2048,0.015568000078201295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,2048,0.012210133671760558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,10240,1536,0.0264629324277242
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,10240,1536,0.011169067025184632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,1536,0.008055466910203297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,1536,0.010657067100207012
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,10240,1024,0.024638932943344117
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,10240,1024,0.008083199958006541
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,1024,0.006386133531729381
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,1024,0.009077333410580953
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,10240,768,0.023870933055877685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,10240,768,0.0075530668099721265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,10240,512,0.022746666272481283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,768,0.005468800167242686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,12288,3584,0.024793599049250284
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,768,0.00823359986146291
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,10240,512,0.006417066852251689
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,512,0.004580266773700714
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,512,0.007542400062084198
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,10240,256,0.021782400210698445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,10240,256,0.005513600011666616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,256,0.0038399999340375268
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,256,0.006832000116507213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,10240,128,0.022184532880783082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,10240,128,0.005111466844876607
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,128,0.0034101332227389016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,64,0.006551466882228851
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,128,0.0065610667069753
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,10240,64,0.020653865734736123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,32,0.006595199803511302
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,12288,1536,0.012405332922935487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,64,0.0031199999153614043
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,10240,32,0.019361066818237304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,32,0.0031701333820819853
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,8192,65536,0.20968640645345055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,8192,16384,0.07711359659830729
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,65536,0.23789760271708169
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,16384,0.09321280320485434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,8192,65536,0.23885760307312012
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,65536,0.37408854166666666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,16384,0.05888746579488119
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,8192,12288,0.06913066705067952
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,12288,256,0.005500799914201101
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,8192,12288,0.042029865582784015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,8192,10240,0.058696532249450685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,12288,0.05617920160293579
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,12288,0.08295786380767822
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,8192,10240,0.035923198858896895
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,10240,0.05886933406194052
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,10240,0.04081706603368123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,8192,8192,0.05198613405227661
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,8192,8192,0.028987733523050944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,8192,0.04810133377710978
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,8192,0.03562239805857341
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,8192,7168,0.02720639904340108
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,8192,7168,0.0522271990776062
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,7168,0.04240640004475911
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,7168,0.03102826674779256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,8192,6144,0.04899093310038249
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,8192,6144,0.023075199127197264
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,6144,0.037043201923370364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,6144,0.02717333237330119
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,8192,5120,0.043044265111287436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,8192,5120,0.021171200275421142
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,5120,0.03150826692581177
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,5120,0.022453333934148154
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,8192,4096,0.032884265979131064
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,8192,4096,0.015397333105405173
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,4096,0.02571733395258586
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,4096,0.01607146660486857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,8192,3584,0.029798400402069092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,8192,3584,0.01767359972000122
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,8192,3072,0.01392213304837545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,3584,0.023014400402704874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,3584,0.014863999684651694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,8192,3072,0.028119466702143353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,3072,0.020257065693537392
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,3072,0.013697066903114319
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,8192,2560,0.02759466568628947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,8192,2560,0.014005333185195923
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,2560,0.01570453345775604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,2560,0.012615467111269632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,2048,0.011683199803034465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,8192,2048,0.02645439902941386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,8192,1536,0.025382399559020996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,8192,2048,0.010197333494822184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,2048,0.008517332871754964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,8192,1536,0.010382933417956035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,1536,0.007179733117421467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,1536,0.009804800152778625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,8192,1024,0.024075732628504435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,1024,0.005799466868241628
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,1024,0.008467200398445129
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,8192,768,0.023172267278035483
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,8192,768,0.007062399884064992
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,768,0.004972800115744273
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,768,0.007912533481915791
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,8192,512,0.022524799903233847
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,8192,512,0.0060479998588562015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,512,0.004343466460704803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,512,0.007292800148328145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,8192,256,0.021545600891113282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,8192,256,0.005248000224431356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,256,0.003717333326737086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,256,0.006795733173688252
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,8192,128,0.022029866774876915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,10240,2560,0.014526933431625366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,8192,128,0.004807466765244802
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,128,0.0033621333539485933
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,128,0.006451199948787689
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,8192,64,0.02047146757443746
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,64,0.003089066594839096
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,64,0.006387199958165486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,8192,32,0.01955733299255371
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,32,0.0030975999931494398
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,32,0.006311466793219249
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,7168,16384,0.07307946681976318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,65536,0.18852267265319825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,7168,65536,0.21754560470581055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,65536,0.32090880076090494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,16384,0.08534719944000244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,16384,0.056645333766937256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,7168,12288,0.04058133363723755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,12288,0.03912746508916219
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,12288,0.06541440089543661
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,7168,12288,0.06533333460489908
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,7168,10240,0.05531839927037557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,7168,10240,0.03248320023218791
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,10240,0.05575999816258749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,10240,0.03436479965845744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,7168,8192,0.0487391988436381
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,7168,8192,0.02685653368631999
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,8192,0.04563200076421102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,8192,0.02869759996732076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,7168,7168,0.04960213502248128
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,7168,7168,0.02502400080362956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,7168,0.040812798341115314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,7168,0.025891200701395674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,7168,6144,0.045960533618927005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,7168,6144,0.02246826688448588
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,6144,0.03681173324584961
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,6144,0.02180160085360209
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,7168,5120,0.03544106483459473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,7168,5120,0.018972800175348917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,5120,0.030859732627868654
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,5120,0.015782399972279867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,7168,4096,0.03133866588274638
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,8192,16384,0.054698665936787925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,7168,4096,0.014612266421318054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,4096,0.023101866245269775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,4096,0.013153066237767538
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,7168,3584,0.029946666955947877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,7168,3584,0.014495999614397685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,3584,0.021153066555658975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,3584,0.012308266758918763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,7168,3072,0.02824746568997701
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,7168,3072,0.012784000237782797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,3072,0.018312533696492515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,3072,0.011757866541544596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,7168,2560,0.02739306688308716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,7168,2560,0.012865066528320312
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,2560,0.009691733121871948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,2560,0.010680533448855082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,7168,2048,0.026897066831588747
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,7168,2048,0.009331199526786804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,2048,0.008412800232569377
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,2048,0.009920000036557516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,7168,1536,0.025201066335042314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,7168,1536,0.009711999694506328
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,7168,1024,0.023740800221761067
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,1536,0.006799999872843425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,1536,0.00886293351650238
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,7168,1024,0.00705813318490982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,1024,0.0056415999929110205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,1024,0.0077461332082748415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,7168,768,0.022782933712005616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,768,0.004924799998601278
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,768,0.007216000060240428
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,7168,512,0.022129066785176597
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,7168,512,0.005888000130653381
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,512,0.00418453315893809
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,512,0.0067221333583196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,7168,256,0.02164586583773295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,7168,256,0.005109333495299021
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,256,0.003601066768169403
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,256,0.006185600161552429
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,7168,128,0.021960532665252684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,7168,128,0.004747733473777771
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,128,0.003302400062481562
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,128,0.005860266586144766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,7168,64,0.02028053402900696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,64,0.005873066683610281
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,64,0.0030271999537944795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,7168,32,0.019028266270955406
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,32,0.0030410667260487873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,32,0.005874133110046387
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,8192,1024,0.007561600208282471
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,6144,65536,0.17773760159810384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,6144,65536,0.18203520774841309
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,65536,0.1759936014811198
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,65536,0.2841610590616862
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,6144,16384,0.06564693450927735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,6144,16384,0.050424532095591224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,16384,0.07760960261027018
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,16384,0.04722453355789184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,6144,12288,0.037963732083638506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,6144,12288,0.05637973149617513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,12288,0.058100267251332605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,12288,0.03410666783650716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,6144,10240,0.05063680013020834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,6144,10240,0.03094826738039653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,10240,0.04700266520182292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,10240,0.03034026622772217
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,6144,8192,0.04641600052515666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,6144,8192,0.025243733326594037
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,8192,0.039638400077819824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,8192,0.025230934222539265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,6144,7168,0.046027731895446775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,6144,7168,0.02466239929199219
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,7168,0.035087998708089194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,7168,0.022194133202234904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,6144,6144,0.037172265847524005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,7168,16384,0.05012480020523071
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,6144,6144,0.01890666683514913
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,6144,0.02940160036087036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,6144,0.01690346598625183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,7168,65536,0.17996266682942708
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,6144,5120,0.033463466167449954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,6144,5120,0.01817173361778259
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,5120,0.025962666670481367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,5120,0.014717866977055868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,6144,4096,0.030898133913675945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,6144,4096,0.014051199952761332
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,4096,0.012604799866676331
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,4096,0.024333866437276204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,6144,3584,0.029645866155624388
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,6144,3584,0.016793600718180337
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,3584,0.017106133699417114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,3584,0.011390933394432068
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,6144,3072,0.028167466322580974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,6144,3072,0.013275733590126038
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,3072,0.010757333040237427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,3072,0.010955733060836793
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,6144,2560,0.02770773371060689
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,6144,2560,0.013269333044687906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,2560,0.009278933207194012
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,2560,0.010388267040252686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,6144,2048,0.026792534192403156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,6144,2048,0.009379200140635173
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,2048,0.008071466783682505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,2048,0.009518933296203614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,6144,1536,0.024965333938598632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,6144,1536,0.010282666484514872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,1536,0.006671999891599019
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,1536,0.008541867136955261
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,6144,768,0.006986666719118755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,6144,1024,0.023560533920923867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,1024,0.005371733506520589
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,1024,0.007558399935563405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,6144,768,0.023116799195607503
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,768,0.004694400231043497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,768,0.007112533350785573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,6144,512,0.022087466716766358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,512,0.004062933226426443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,256,0.003532800078392029
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,512,0.006548266609509785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,6144,256,0.021661865711212158
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,256,0.006156800190607706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,6144,128,0.021891200542449953
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,128,0.005916800101598104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,6144,128,0.004833066463470459
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,6144,64,0.02027413249015808
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,128,0.003257599969704946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,64,0.0032543999453385672
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,64,0.005866666634877523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,6144,32,0.01924906571706136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,32,0.003020799905061722
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,32,0.005840000013510386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,7168,768,0.00676800012588501
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,5120,65536,0.15522027015686035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,5120,65536,0.17015786170959474
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,65536,0.23424320220947265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,5120,16384,0.059114666779836025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,65536,0.1354581356048584
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,5120,16384,0.04419413407643636
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,16384,0.05935253302256266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,16384,0.038626134395599365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,5120,12288,0.051787734031677246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,5120,12288,0.034355199337005614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,12288,0.04591466585795085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,12288,0.030548266569773358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,5120,10240,0.047172268231709794
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,5120,10240,0.028859732548395793
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,10240,0.038777601718902585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,10240,0.02662186622619629
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,5120,8192,0.03999360005060832
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,5120,8192,0.023202133178710938
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,8192,0.031930667161941526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,8192,0.019687465826670327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,5120,7168,0.039876266320546465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,5120,7168,0.0204693337281545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,7168,0.028424533208211263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,7168,0.016177067160606386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,5120,6144,0.034618667761484784
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,5120,6144,0.017607466379801432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,6144,0.02506133317947388
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,6144,0.014195199807484946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,5120,5120,0.03269866704940796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,5120,0.021297067403793335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,5120,5120,0.016244266430536905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,5120,0.012845866878827415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,5120,4096,0.03036053379376729
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,5120,4096,0.013804800311724343
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,4096,0.014033066232999167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,4096,0.011228799819946289
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,5120,3584,0.0293013334274292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,5120,3584,0.013877333203951518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,3584,0.0100490669409434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,3584,0.01050879955291748
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,5120,3072,0.028094933430353804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,3072,0.008835200468699138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,3072,0.010186666250228881
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,5120,2560,0.02719893256823222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,5120,2560,0.012391466895739238
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,2560,0.00782293329636256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,2560,0.00949013332525889
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,5120,2048,0.025778132677078246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,2048,0.0069013332327206925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,2048,0.009031466643015544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,5120,1536,0.024836266040802003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,5120,1536,0.009495466947555542
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,1536,0.005993600189685822
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,1536,0.00803413341442744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,5120,1024,0.023347200949986775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,5120,1024,0.006861866513888042
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,1024,0.004866133133570353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,1024,0.00719893326361974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,5120,768,0.02338026762008667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,768,0.004409599800904592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,5120,768,0.006569600105285645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,768,0.006922666728496551
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,5120,512,0.022012799978256226
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,512,0.003885866701602936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,5120,512,0.005755733450253805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,512,0.00639573335647583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,5120,256,0.02151893377304077
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,5120,256,0.005097599824269613
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,256,0.003403733422358831
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,256,0.0060576001803080235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,5120,128,0.021653334299723305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,5120,128,0.004681600133577982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,128,0.0032170665760835014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,128,0.005836800237496694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,5120,64,0.020292266209920248
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,64,0.002924799919128418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,64,0.005829333265622457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,5120,32,0.018938666582107543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,32,0.0029824001093705496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,32,0.005832533538341522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,6144,1024,0.007378133138020833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,4096,65536,0.16184959411621094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,4096,65536,0.13233919938405353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,65536,0.1945802688598633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,4096,16384,0.05391253232955932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,65536,0.10843093395233154
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,4096,16384,0.04098453521728516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,16384,0.03224213322003682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,4096,12288,0.04706133206685384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,16384,0.051430400212605795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,6144,512,0.005992533266544342
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,4096,12288,0.032781867186228435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,12288,0.03963413238525391
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,12288,0.025997867186864216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,4096,10240,0.04204800128936768
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,4096,10240,0.02741546630859375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,6144,256,0.005301333467165629
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,10240,0.03402666648228963
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,10240,0.019080533583958944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,4096,8192,0.03232640027999878
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,8192,0.028166399399439497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,8192,0.014174933234850565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,4096,7168,0.033165866136550905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,4096,7168,0.020028799772262573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,7168,0.025464532772699992
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,7168,0.013052800297737121
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,4096,6144,0.03136639992396037
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,4096,6144,0.017258665959040322
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,6144,0.02278613249460856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,6144,0.012152533729871113
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,4096,5120,0.029897600412368774
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,4096,5120,0.016477866967519125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,5120,0.017412267128626504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,5120,0.011073066790898641
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,4096,4096,0.028386133909225463
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,4096,4096,0.01317759950955709
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,4096,0.010583466291427613
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,4096,0.00974826713403066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,4096,3584,0.027244800329208375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,4096,3584,0.014358400305112203
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,3584,0.007994666695594788
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,3584,0.00927786628405253
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,4096,3072,0.026748800277709962
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,4096,3072,0.011040000120798747
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,3072,0.007421866556008657
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,3072,0.008829866846402485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,4096,2560,0.025923200448354083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,4096,2560,0.012645333011945089
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,2560,0.006650666892528534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,4096,2048,0.008661333719889324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,2560,0.008431999882062276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,4096,2048,0.025130667289098102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,2048,0.00600853314002355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,2048,0.007939200103282928
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,4096,1536,0.00997440020243327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,4096,1536,0.024425599972407022
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,1536,0.005196799834569296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,1536,0.007406933108965556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,4096,1024,0.022732800245285033
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,4096,1024,0.0069013332327206925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,1024,0.0045045331120491024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,1024,0.006650666892528534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,4096,768,0.02241599957148234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,4096,768,0.0067114666104316715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,768,0.004136533290147781
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,768,0.006482133269309997
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,4096,512,0.021758933862050377
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,512,0.0037471999724706015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,512,0.006157866617043813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,4096,256,0.02140480081240336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,256,0.0033333333830038703
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,256,0.0059114664793014525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,5120,3072,0.012284800410270691
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,4096,128,0.021805866559346517
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,4096,128,0.004743466774622599
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,128,0.003155199935038885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,128,0.005735466877619425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,4096,64,0.02039360006650289
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,5120,2048,0.008940800031026205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,64,0.002902399996916453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,64,0.005750399827957153
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,4096,32,0.018940800428390504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,32,0.002946133414904277
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,32,0.005737600227197012
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3584,65536,0.14061120351155598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3584,65536,0.12340479691823322
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3584,16384,0.05109440088272095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,65536,0.0991861343383789
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,16384,0.04527999957402547
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,65536,0.16449920336405438
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3584,16384,0.0408458669980367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,16384,0.02969920039176941
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3584,12288,0.044021332263946535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3584,12288,0.03146986762682597
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3584,10240,0.03474133412043254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,12288,0.02414399981498718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,12288,0.03538133303324382
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3584,10240,0.02500586708386739
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3584,8192,0.030491733551025392
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,10240,0.030747733513514203
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,10240,0.016376533110936484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,8192,0.026100265979766845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,8192,0.013306666413942972
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3584,7168,0.03292373418807983
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,7168,0.023666133483250938
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,7168,0.012422399719556172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3584,6144,0.031199999650319415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3584,6144,0.015686399737993875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,6144,0.020126932859420778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,6144,0.011506133278210958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3584,5120,0.02993280092875163
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3584,5120,0.01600320041179657
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,5120,0.011125333110491435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,5120,0.011055999994277954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3584,4096,0.028350933392842607
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3584,4096,0.012361600001653036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,4096,0.00849173367023468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,4096,0.009782399733861287
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3584,3584,0.027368533611297607
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3584,3584,0.014569600423177084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,3584,0.00792746643225352
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,3584,0.009266133109728496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3584,3072,0.026606933275858564
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3584,3072,0.011448533336321513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,3072,0.007213866710662842
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,3072,0.008843732873598735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3584,2560,0.026019199689229326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3584,2560,0.011394133170445759
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,2560,0.006428800026575724
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,2560,0.008434133728345235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,4096,8192,0.021661865711212158
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3584,2048,0.02514773408571879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3584,2048,0.00811413327852885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,2048,0.005801600217819214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,2048,0.007727999985218048
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3584,1536,0.024401066700617473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,1536,0.005070933202902476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,1536,0.007436800003051758
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3584,1024,0.022935465971628825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3584,1024,0.006351999938488007
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,1024,0.004478933413823446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,1024,0.006712533533573151
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3584,768,0.02257919907569885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3584,768,0.006402133405208588
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,768,0.0040501333773136135
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,768,0.006425599753856659
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3584,512,0.021795199314753214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3584,512,0.005692799886067709
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,512,0.003697066754102707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,512,0.006229333579540253
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3584,256,0.02116586764653524
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3584,256,0.005027199784914652
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,256,0.003357866654793421
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,256,0.0059114664793014525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3584,128,0.021811199188232423
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3584,128,0.004604800045490265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,128,0.003126399964094162
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,128,0.005746133128801982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3584,64,0.02022613286972046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,64,0.002962133288383484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,64,0.005762133498986562
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,32,0.002915200094381968
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3584,32,0.019151999553044637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,32,0.005758933226267497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3072,65536,0.1307029326756795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3072,65536,0.11382186412811279
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,65536,0.1458122730255127
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,65536,0.08843519687652587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3072,16384,0.04873493512471517
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3072,16384,0.038791465759277347
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,16384,0.04026240110397339
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,16384,0.027453867594401042
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3072,12288,0.03648639917373657
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3072,12288,0.029164799054463703
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,12288,0.031540266672770184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,12288,0.018168532848358156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3072,10240,0.03454506794611613
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3072,10240,0.024039467175801597
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,10240,0.027029333511988322
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,10240,0.015521066387494406
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3072,8192,0.030034132798512775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3072,8192,0.020219733317693077
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,8192,0.02273919979731242
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,8192,0.01325440009435018
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3072,7168,0.032798933982849124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3072,7168,0.017847466468811034
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,7168,0.01881706714630127
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,7168,0.011747200290362041
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3072,6144,0.03149333397547404
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3072,6144,0.01597119967142741
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,4096,512,0.005906133353710175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,6144,0.011990400155385335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,6144,0.011398399869600933
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,4096,256,0.005074133475621542
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3072,5120,0.029660799105962116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,5120,0.010552533467610677
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,5120,0.010563199718793232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3072,4096,0.028151466449101763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3072,4096,0.012238933642705282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,4096,0.010018133123715718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,4096,0.009343999624252319
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3072,3584,0.027619200944900512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3072,3584,0.012123733758926392
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,3584,0.009268266956011455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,3584,0.009006933371225993
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3072,3072,0.0267850657304128
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,3072,0.008410666386286418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3072,3072,0.010377599795659383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3072,2560,0.02603413263956706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,3072,0.008566400408744812
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3072,2560,0.01088853379090627
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,2560,0.006444799900054932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,2560,0.008017066617806752
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3072,2048,0.02515946626663208
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3072,2048,0.007999999821186066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,2048,0.005681066711743673
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,2048,0.007419733206431071
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3072,1536,0.024375466505686443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3072,1536,0.008901333808898926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,1536,0.005019733309745788
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,1536,0.007143466671307881
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3072,1024,0.022711465756098427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3072,1024,0.006545066833496094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,1024,0.004337066908677419
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,1024,0.006632533172766368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3072,768,0.02254826625188192
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3072,768,0.0062837332487106325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,768,0.004041599979003271
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3584,8192,0.020882133642832437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,768,0.006381866832574208
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3584,7168,0.019056000312169395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3072,512,0.022512000799179078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3072,512,0.005529599885145823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,512,0.003704533229271571
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,512,0.006189866860707601
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3072,256,0.021005866924921672
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3072,256,0.004904533425966898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,256,0.0033333333830038703
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,256,0.0059456000725428265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3072,128,0.021784534056981407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3072,128,0.004574933151404063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,128,0.0030773334205150605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,128,0.0056991999348004665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3072,64,0.02031146685282389
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,64,0.0029258665939172106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,64,0.005724800129731497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,3072,32,0.01918506622314453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,32,0.0029653333127498626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,32,0.005773866673310598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2560,65536,0.11339413324991862
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2560,65536,0.10649920304616292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,65536,0.12351146539052327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,65536,0.08095253308614095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2560,16384,0.04545493523279826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2560,16384,0.03742186625798543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,16384,0.033905065059661864
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,16384,0.02207146684328715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2560,12288,0.033581864833831784
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2560,12288,0.029458133379618327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,12288,0.027246934175491334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,12288,0.015774933497111003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2560,10240,0.03191786607106527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2560,10240,0.021831466754277547
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3584,1536,0.00921493371327718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,10240,0.023625600337982177
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,10240,0.014254933595657349
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2560,8192,0.030054400364557903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2560,8192,0.020719999074935914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,8192,0.01800640026728312
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,8192,0.012337066729863485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2560,7168,0.03289920091629028
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2560,7168,0.01779306729634603
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,7168,0.011915733416875202
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,7168,0.011620266238848369
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2560,6144,0.03096853295962016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2560,6144,0.016416000326474507
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,6144,0.011155200004577637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,6144,0.010813867052396137
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2560,5120,0.029743999242782593
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2560,5120,0.014759467045466105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,5120,0.010424533486366272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,5120,0.009992532928784688
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2560,4096,0.028350933392842607
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2560,4096,0.011480533083279927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,4096,0.00902826686700185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,4096,0.008914132912953694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2560,3584,0.02749333381652832
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2560,3584,0.013270399967829385
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,3584,0.008458667000134786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,3584,0.008575999736785888
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2560,3072,0.02676266630490621
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2560,3072,0.010754133264223736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,3072,0.007885866860548655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,3072,0.008182399968306223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2560,2560,0.026049067576726277
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,2048,0.0056533331672350565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2560,2560,0.010554666320482891
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,2560,0.006354133288065593
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,2560,0.0077567999561627705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2560,2048,0.025069866577784223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,2048,0.007459199925263722
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2560,1536,0.02390186587969462
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,1536,0.0049781332413355505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,1536,0.006923733154932658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2560,1024,0.02272640069325765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2560,1024,0.006428800026575724
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,1024,0.004314666489760081
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,1024,0.006504533191521962
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2560,768,0.02287893295288086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,768,0.004011733333269755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,768,0.006346666812896728
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2560,512,0.02172266642252604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,512,0.006116266548633576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2560,512,0.005402666827042898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,512,0.003704533229271571
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3072,5120,0.014113066593805948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2560,256,0.021134932835896812
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2560,256,0.004810666541258494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,256,0.0033290666838486993
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,256,0.00591786652803421
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2560,128,0.021715199947357176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2560,128,0.004474666714668274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,128,0.003067733347415924
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,128,0.005723733206590017
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2560,64,0.02021226684252421
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,64,0.002940800040960312
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,64,0.005785599847634633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2560,32,0.01880853374799093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,32,0.002922666569550832
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,32,0.005795200169086456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2048,65536,0.12834773063659669
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2048,65536,0.09985600312550863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,65536,0.09756480058034261
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,65536,0.07313493092854818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2048,16384,0.03868266741434733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2048,16384,0.033890132109324136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,16384,0.02801919976870219
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,16384,0.020426666736602782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2048,12288,0.03337706724802653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2048,12288,0.026744532585144042
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,12288,0.02261120080947876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,12288,0.01570026675860087
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2048,10240,0.03169493277867635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2048,10240,0.02246826688448588
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,10240,0.016778665781021117
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,10240,0.01411626636981964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2048,8192,0.030085333188374835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2048,8192,0.019195733467737834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,8192,0.010313600301742554
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,8192,0.012262399991353352
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2048,7168,0.03269866704940796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2048,7168,0.018667733669281004
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,7168,0.00965013305346171
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,7168,0.01147093375523885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2048,6144,0.031074132521947222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,6144,0.00920853316783905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2048,6144,0.01548693378766378
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,6144,0.010624000430107116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2048,5120,0.029711999495824176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,5120,0.009909333785374959
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2048,5120,0.013398399949073792
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,5120,0.008570667107899983
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2048,4096,0.02823573350906372
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2048,4096,0.011776000261306763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,4096,0.00784746656815211
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,4096,0.008821333448092144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2048,3584,0.027548799912134807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2048,3584,0.013589333494504294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,3584,0.007539199789365132
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,3584,0.008451199531555176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2048,3072,0.02632746696472168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,3072,0.007163733243942261
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,3072,0.008038400113582611
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2048,2560,0.025854933261871337
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,2560,0.0062613333264986675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,2560,0.007659733295440674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2048,2048,0.02478826642036438
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2048,2048,0.0081535999973615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,2048,0.005594666798909505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,2048,0.007201066613197327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2048,1536,0.024153600136439003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2048,1536,0.009598933657010396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,1536,0.00492799977461497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,1536,0.006793599824110668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2048,1024,0.023242666323979696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2048,1024,0.0068010667959849044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,1024,0.006413866579532623
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,1024,0.004251733422279358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2048,768,0.022282665967941283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2048,768,0.006545066833496094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,768,0.006183466811974844
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,768,0.003942399968703588
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2048,512,0.022457599639892578
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2048,512,0.005734399954477946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,512,0.00359253336985906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,512,0.006026666859785716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2048,256,0.021057067314783733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2048,256,0.004995200037956238
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,256,0.003311999887228012
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,256,0.0057781333724657696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2048,128,0.021746132771174112
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2048,128,0.00462719996770223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,128,0.003058133274316788
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,128,0.005649066468079885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2048,64,0.020317866404851278
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,64,0.0028736000259717304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,64,0.0056309332450230915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,2048,32,0.019081600507100425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,32,0.0056757330894470215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,32,0.002880000074704488
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1536,65536,0.11228799819946289
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1536,65536,0.09474666913350424
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,65536,0.07360959847768148
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,65536,0.06790293057759603
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1536,16384,0.03727146784464518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1536,16384,0.032125866413116454
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,16384,0.022909865776697794
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,16384,0.018977065881093345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1536,12288,0.03371200164159139
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1536,12288,0.02603626648585002
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,12288,0.012877866625785828
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,12288,0.01562666694323222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2560,1536,0.00836906631787618
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2560,2048,0.008066133161385854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1536,10240,0.03190293312072754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1536,10240,0.021729065974553426
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,10240,0.011218133568763732
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,10240,0.014019200205802917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,8192,0.010075733065605164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1536,8192,0.030247465769449873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1536,8192,0.018604799111684164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1536,7168,0.01655359963575999
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,8192,0.012199466427167256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1536,7168,0.03257493376731872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,7168,0.009382399916648864
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2560,768,0.006043733159701029
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,7168,0.011460266510645549
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1536,6144,0.031250133117039995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1536,6144,0.01461013356844584
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,6144,0.008753066261609394
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,6144,0.010637866457303365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1536,5120,0.029601067304611206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,5120,0.008097066481908163
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,5120,0.009869866569836934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1536,4096,0.028250666459401448
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1536,4096,0.011206400394439698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,4096,0.007423999905586243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,4096,0.008761599659919739
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1536,3584,0.02751893401145935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1536,3584,0.012735999623934426
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,3584,0.007325866818428039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,3584,0.008402132987976074
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1536,3072,0.026539733012517292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1536,3072,0.010488532980283101
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,3072,0.006972800195217133
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,3072,0.008030933141708375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1536,2560,0.025964800516764325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,2560,0.006258133550484974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,2560,0.007659733295440674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1536,2048,0.02472533384958903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1536,2048,0.007716266810894013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,2048,0.00557226687669754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,2048,0.007209600011507671
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1536,1536,0.02395626703898112
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1536,1536,0.008539733290672303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,1536,0.00491946687301
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,1536,0.0068256000677744556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1536,1024,0.022740266720453896
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1536,1024,0.006258133550484974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,1024,0.004248533149560293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,1024,0.006377600133419037
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1536,768,0.022362667322158813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1536,768,0.0061471998691558834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,768,0.0038751999537150065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,768,0.006234666705131531
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1536,512,0.021734400590260824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1536,512,0.0054058666030565895
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,512,0.0035989334185918174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,512,0.006060799956321717
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1536,256,0.02101226647694906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1536,256,0.0047765334447224935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,256,0.003234133372704188
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,256,0.005883733431498209
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1536,128,0.02157333294550578
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1536,128,0.004462933540344239
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,128,0.003018666555484136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,128,0.005743999779224396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1536,64,0.020100265741348267
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,64,0.002867199977238973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,64,0.0056650668382644655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1536,32,0.01910933256149292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,32,0.002841600030660629
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,32,0.0056096002459526065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1024,65536,0.08939306735992432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1024,65536,0.09964693387349446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,65536,0.051795200506846106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,65536,0.06586986780166626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1024,16384,0.036883199214935304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2048,3072,0.01153706709543864
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1024,16384,0.03176213304201762
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,16384,0.011242666840553283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,16384,0.018953599532445273
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1024,12288,0.033218133449554446
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1024,12288,0.024036266406377158
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,12288,0.010110933581988018
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,12288,0.015576533476511636
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1024,10240,0.03159466584523519
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1024,10240,0.02012373407681783
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2048,2560,0.012113066514333089
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,10240,0.008970666925112407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,10240,0.01404159963130951
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1024,8192,0.029767467578252153
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1024,8192,0.017847466468811034
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,8192,0.008029866715272267
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,8192,0.012239999572436015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1024,7168,0.032816000779469806
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1024,7168,0.015107199549674988
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,7168,0.00786133309205373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,6144,0.006931200126806895
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,7168,0.011408000191052755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1024,6144,0.030909866094589233
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,6144,0.010610133409500122
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1024,5120,0.02956693371136983
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1024,5120,0.012723199526468911
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,5120,0.007268266876538594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,5120,0.009902933239936828
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1024,4096,0.028321067492167156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1024,4096,0.01086293359597524
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,4096,0.006954666475454967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,4096,0.00876800020535787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1024,3584,0.027577600876490277
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1024,3584,0.010149332880973815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,3584,0.007390933235486348
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,3584,0.008362666765848795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1024,3072,0.026687999566396076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1024,3072,0.009197866916656494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,3072,0.0070592001080513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,3072,0.007977599898974102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1024,2560,0.02582719922065735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,2560,0.00625493327776591
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,2560,0.007587199906508128
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1024,2048,0.025131734212239583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,2048,0.005569066603978475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1024,2048,0.007271466652552287
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,2048,0.007161599894364675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1024,1536,0.02390506664911906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1024,1536,0.007454933226108551
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,1536,0.004884266853332519
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,1536,0.006750933329264323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1024,1024,0.022694400946299233
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,1024,0.0042250668009122215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,1024,0.006348800162474315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1024,1024,0.005623466769854228
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1024,768,0.022232532501220703
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1024,768,0.005555200080076853
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,768,0.003851733356714249
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,768,0.006181333462397257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1024,512,0.021695999304453532
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1024,512,0.005093333125114441
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,512,0.0035637333989143372
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,512,0.005986133217811584
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1024,256,0.020910932620366415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,256,0.005795200169086456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1024,256,0.004682666560014089
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,256,0.003230933348337809
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1024,128,0.021581866343816123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1024,128,0.004429866870244345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,128,0.0029685333371162414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,128,0.005596800148487091
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1024,64,0.020039467016855876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,64,0.003209600100914637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,64,0.005646933118502299
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,1024,32,0.018849066893259683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1536,5120,0.013517866532007853
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,32,0.003206400076548258
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,768,65536,0.08659413655598959
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,32,0.005692799886067709
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,768,65536,0.08669439951578775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,65536,0.039544534683227536
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,768,16384,0.03705173333485921
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,65536,0.06596693197886148
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,768,16384,0.023435733715693154
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,16384,0.01032960017522176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,16384,0.018913066387176512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,768,12288,0.03343360026677449
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,768,12288,0.020733867088953653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,12288,0.00907306671142578
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,12288,0.015585066874821982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1536,2560,0.01055680016676585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,768,10240,0.031649067004521685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,10240,0.008594133456548055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,10240,0.014097066720326743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,768,8192,0.029806933800379437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,768,8192,0.015558399756749473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,8192,0.007714133461316426
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,8192,0.01227946678797404
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,768,7168,0.03275413314501445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,7168,0.007064533233642578
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,7168,0.011357866724332174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,768,6144,0.030854399998982745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,6144,0.010573866963386535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,768,6144,0.010099200407663982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,6144,0.006862933437029521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,768,5120,0.029386667410532634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,768,5120,0.011595732967058818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,5120,0.007266133526961009
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,5120,0.009948800007502239
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,768,4096,0.027923200527826948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,768,4096,0.009315199653307597
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,4096,0.006809600194295247
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,4096,0.008874666690826417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,768,3584,0.0274944007396698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,768,3584,0.010046933094660442
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,3584,0.007346133391062419
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,3584,0.008431999882062276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,768,3072,0.008366933465003968
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,768,3072,0.02667306661605835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,3072,0.0068800002336502075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,3072,0.008035199840863545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,768,2560,0.025751467545827227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,2560,0.006230400005976359
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,2560,0.00763733337322871
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,768,2048,0.025172267357508344
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,2048,0.005564799904823304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,2048,0.007130666573842366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,768,1536,0.02419840097427368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,768,1536,0.0064746667941411335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,1536,0.004852266609668731
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,1536,0.006764799853165944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,768,1024,0.023089067141215006
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,768,1024,0.005419733126958212
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,1024,0.004217599829037985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,1024,0.006397866706053417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,768,768,0.02289386590321859
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,768,768,0.005333333214124044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,768,0.004038399954636892
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,768,0.006205866734186808
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,768,512,0.022457599639892578
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,768,512,0.005014400184154511
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,512,0.003555200000603994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,512,0.006027733286221823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,768,256,0.021272534132003786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1024,6144,0.014813866217931113
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,256,0.003385599950949351
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,256,0.00576853354771932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,768,128,0.021717333793640138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,768,128,0.004386133452256521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,128,0.003115733216206233
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,128,0.005654400090376536
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,768,64,0.020331732432047524
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,64,0.0030239999294281008
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,64,0.005724800129731497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,768,32,0.01886826753616333
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,32,0.002974933385848999
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,32,0.005645866692066193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,512,65536,0.08329813480377198
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,512,65536,0.0731765349706014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,65536,0.02927466630935669
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,65536,0.0642848014831543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,512,16384,0.037300264835357665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,512,16384,0.01932586630185445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,16384,0.008451199531555176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,16384,0.018869332472483315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,512,12288,0.03358506759007772
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,512,12288,0.01579093337059021
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,12288,0.0077354664603869125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,12288,0.015571199854214988
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1024,2560,0.00897173285484314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,512,10240,0.031804800033569336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,512,10240,0.013946666320164999
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,10240,0.007295999924341838
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,10240,0.01406719982624054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,512,8192,0.029741867383321123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,512,8192,0.012063999970753986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,8192,0.007114666700363159
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,8192,0.012152533729871113
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,512,7168,0.032356266180674234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,512,7168,0.011010133226712545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,7168,0.006974933544794719
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,7168,0.011460266510645549
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,512,6144,0.031091199318567915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,512,6144,0.010227200388908387
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,6144,0.006762666503588359
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,6144,0.010630399982134501
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,512,5120,0.029470932483673096
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,512,5120,0.009213866790135701
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,5120,0.007152000069618225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,5120,0.009763200084368389
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,512,4096,0.027867732445398967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,512,4096,0.008251733581225077
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,4096,0.006735999882221222
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,4096,0.008835200468699138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,512,3584,0.027497599522272747
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,512,3584,0.00786133309205373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,3584,0.00719893326361974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,3584,0.008379733562469483
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,512,3072,0.026589866479237872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,512,3072,0.007853866616884867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,3072,0.006861866513888042
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,3072,0.008007466793060303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,512,2560,0.025663999716440837
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,512,2560,0.007146666447321574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,2560,0.006323199967543285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,2560,0.007621333499749501
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,512,2048,0.02505279978116353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,512,2048,0.006286933521429698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,2048,0.005633066594600678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,2048,0.00713919997215271
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,512,1536,0.024029866854349772
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,512,1536,0.006305066744486491
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,1536,0.00499839981396993
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,1536,0.006750933329264323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,512,1024,0.022482132911682128
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,512,1024,0.005276800195376078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,1024,0.00432640016078949
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,1024,0.006359466910362243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,512,768,0.0223146657148997
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,512,768,0.0051360001166661584
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,768,0.004013866682847341
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,768,0.006161066889762879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,512,512,0.021652267376581828
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,512,512,0.004844800134499868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,768,10240,0.017384533087412515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,512,0.0036949334045251214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,512,0.005971199770768484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,512,256,0.02096959948539734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,512,256,0.004568533102671305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,256,0.003382399926582972
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,256,0.005682133138179779
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,512,128,0.021600000063578286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,512,128,0.004344533383846283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,128,0.003068800022204717
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,512,64,0.02031893332799276
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,768,7168,0.013199999928474426
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,128,0.0056533331672350565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,64,0.0029120000700155893
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,64,0.005691733459631602
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,32,0.00560746689637502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,512,32,0.018776534001032512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,32,0.0028927999238173166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,256,65536,0.08108373483022055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,256,65536,0.0652010679244995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,65536,0.01218239963054657
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,65536,0.05916159947713216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,256,16384,0.03733439842859904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,16384,0.006923733154932658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,16384,0.01877440015474955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,256,12288,0.033675734202067056
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,256,12288,0.015601066748301187
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,12288,0.006822399795055389
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,12288,0.015558399756749473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,256,10240,0.03147413333257039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,256,10240,0.013773866494496665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,10240,0.007106133302052816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,10240,0.01381119986375173
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,256,8192,0.029753599564234418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,256,8192,0.011963733037312825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,8192,0.0070165331164995836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,256,7168,0.010990933577219645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,8192,0.012075733145078022
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,256,7168,0.032553599278132124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,7168,0.006828799843788147
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,7168,0.011362133423487346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,256,6144,0.030750934282938642
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,768,2560,0.008220799763997396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,256,6144,0.010088533163070679
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,6144,0.006692266464233399
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,6144,0.010585600137710571
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,256,5120,0.02943466703097026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,256,5120,0.009051733215649923
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,5120,0.0070261334379514056
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,768,2048,0.006923733154932658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,5120,0.009714133540789286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,256,4096,0.028204800685246785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,256,4096,0.008150400221347808
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,4096,0.006691200037797292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,4096,0.008654933174451191
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,256,3584,0.027162667115529376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,3584,0.007234133283297221
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,3584,0.008222933113574981
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,256,3072,0.026764800151189167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,256,3072,0.007160533467928569
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,3072,0.006794666747252147
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,3072,0.007978666822115581
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,256,2560,0.025705599784851076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,256,2560,0.00665280024210612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,2560,0.006242133180300395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,2560,0.007558399935563405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,256,2048,0.025113600492477416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,256,2048,0.006217599908510844
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,2048,0.005568000177542368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,2048,0.007049599786599477
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,256,1536,0.02391466697057088
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,256,1536,0.005815466741720835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,1536,0.004875733455022176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,768,256,0.004506666461626689
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,1024,0.006217599908510844
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,1536,0.00664213349421819
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,256,1024,0.02272426684697469
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,256,1024,0.005366399884223938
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,1024,0.004231466849644979
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,256,768,0.022344533602396646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,256,768,0.0049674664934476215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,768,0.003870933254559835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,768,0.006084266801675161
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,256,512,0.021474132935206093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,256,512,0.004844800134499868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,512,0.0034976000587145484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,512,0.0057888001203536986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,256,256,0.020920532941818237
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,256,256,0.00447680006424586
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,256,0.003209600100914637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,256,0.005763199925422668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,256,128,0.021606399615605672
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,256,128,0.004279466470082601
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,128,0.003010133405526479
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,128,0.005544533332188925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,256,64,0.01991573373476664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,64,0.002809600035349528
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,64,0.005543466905752818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,256,32,0.018732800086339315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,32,0.0028575999041398365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,32,0.0055754666527112326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,128,65536,0.07925546964009603
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,128,65536,0.06568533182144165
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,65536,0.010026666522026061
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,65536,0.059204268455505374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,128,16384,0.0368938684463501
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,128,16384,0.01921066641807556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,16384,0.006685866912206013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,16384,0.01881386637687683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,128,12288,0.03328640063603719
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,128,12288,0.015754666924476624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,12288,0.006671999891599019
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,12288,0.015378133455912272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,128,10240,0.03163093328475952
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,128,10240,0.013774933417638144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,10240,0.006919466455777486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,10240,0.013790933291117349
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,128,8192,0.029656533400217695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,8192,0.006713599960009258
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,8192,0.012070399522781373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,128,7168,0.03258879979451497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,128,7168,0.010939733187357584
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,7168,0.006586666901906331
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,7168,0.011335466305414836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,128,6144,0.031035733222961426
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,128,6144,0.01002239982287089
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,6144,0.0064522668719291685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,6144,0.010514133175214132
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,128,5120,0.02942933241526286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,128,5120,0.008995200196901958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,5120,0.00680320014556249
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,5120,0.009735467036565144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,128,4096,0.008040533463160197
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,128,4096,0.028199466069539388
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,4096,0.006411733229955037
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,4096,0.008584533135096233
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,128,3584,0.027280000845591228
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,128,3584,0.007627733548482259
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,3584,0.007018666466077168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,3584,0.008239999910195668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,128,3072,0.02664426763852437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,3072,0.0065301333864529925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,3072,0.008003200093905132
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,128,2560,0.025624533494313557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,128,2560,0.006671999891599019
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,2560,0.0058559998869895935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,2560,0.007601066430409749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,128,2048,0.024628265698750814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,128,2048,0.006177066763242086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,2048,0.005241600175698599
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,2048,0.007124266525109608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,128,1536,0.023873066902160643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,128,1536,0.005815466741720835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,1536,0.0046634669105211895
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,1536,0.006718933085600535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,128,1024,0.02253119945526123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,128,1024,0.005234133203824361
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,1024,0.003994666785001755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,1024,0.006348800162474315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,128,768,0.02214399973551432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,128,768,0.004844800134499868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,768,0.0038101332883040107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,768,0.00614933321873347
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,128,512,0.021461333831151327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,128,512,0.004737066725889841
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,512,0.0034965333839257562
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,512,0.005971199770768484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,128,256,0.020805333058039346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,128,256,0.0043605332573254905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,256,0.0031925333042939507
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,256,0.005761066575845083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,128,128,0.021423999468485513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,128,128,0.004232533276081085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,128,0.0029834667841593427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,128,0.0056415999929110205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,128,64,0.01992320020993551
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,64,0.002869333326816559
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,64,0.005508266886075338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,128,32,0.019061332941055296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,32,0.0028394666810830434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,32,0.0055285334587097164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,64,65536,0.07948693434397379
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,65536,0.008076799909273784
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,65536,0.05949653387069702
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,64,16384,0.03687146504720052
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,16384,0.006498133142789205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,12288,0.006489600241184235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,16384,0.01895893414815267
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,64,12288,0.033267199993133545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,256,16384,0.019332265853881835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,12288,0.015379200379053751
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,64,10240,0.03157119949658711
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,10240,0.006797866523265838
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,10240,0.013858133554458618
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,64,8192,0.02985173265139262
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,8192,0.0065738668044408154
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,8192,0.01211199959119161
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,64,7168,0.0324458658695221
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,7168,0.006488533318042755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,7168,0.011223466197649638
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,64,6144,0.03094826738039653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,6144,0.006410666803518932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,6144,0.010431999961535137
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,5120,0.00970240036646525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,64,5120,0.029553065697352093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,5120,0.006967466572920482
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,64,4096,0.02828906575838725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,4096,0.006365866462389629
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,4096,0.008622933427492778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,64,3584,0.02749226689338684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,3584,0.006878933310508728
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,3584,0.008246399958928426
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,64,3072,0.026746666431427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,3072,0.006657066444555919
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,3072,0.007860266665617625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,256,3584,0.007642666498819987
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,64,2560,0.025627734263737996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,2560,0.005831466615200042
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,2560,0.007559466858704884
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,64,2048,0.02494186758995056
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,2048,0.005166933437188466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,2048,0.007010133564472198
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,64,1536,0.024086399873097738
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,1536,0.004650666813055674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,1024,0.006258133550484974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,1536,0.0066442668437957765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,64,1024,0.022563199202219643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,1024,0.0039893334110577905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,64,768,0.022772266467412313
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,768,0.0036789332826932275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,768,0.006106666723887126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,64,512,0.02149440050125122
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,512,0.0033503999312718712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,512,0.005965866645177205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,64,256,0.02095359961191813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,128,0.0056085333228111265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,256,0.0031829332311948144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,64,64,0.020154666900634766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,256,0.005605333546797434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,64,32,0.018806399901707967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,64,128,0.02158720095952352
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,128,0.003005866706371307
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,64,0.0028543998797734577
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,64,0.005568000177542368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,32,0.0028309332827727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,65536,0.05976959864298502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,32,0.005512533088525137
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,65536,0.008269866804281871
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,32,65536,0.07954239845275879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,32,16384,0.03707733154296875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,16384,0.006593066453933716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,16384,0.01895466645558675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,32,12288,0.03336426814397176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,12288,0.006551466882228851
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,12288,0.015397333105405173
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,32,10240,0.031549866994222006
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,10240,0.006673066814740498
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,10240,0.013951999942461648
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,32,8192,0.029743999242782593
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,8192,0.006613333523273468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,8192,0.011967999736467998
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,32,7168,0.03234346707661946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,7168,0.006512000163396199
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,7168,0.011370666821797689
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,32,6144,0.03101866642634074
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,6144,0.006393600006898243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,6144,0.010434133807818095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,32,5120,0.029605333010355634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,5120,0.006776533524195353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,5120,0.009799466530481974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,32,4096,0.028245333830515546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,4096,0.006354133288065593
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,4096,0.008614400029182434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,32,3584,0.02744106650352478
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,3584,0.006854400038719177
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,3584,0.008374399940172831
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,32,3072,0.026655999819437663
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,3072,0.0064640000462532045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,3072,0.007843199868996937
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,32,2560,0.025971200068791705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,2560,0.005764266848564148
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,2560,0.00754559983809789
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,32,2048,0.025006934007008867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,2048,0.0051594664653142296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,2048,0.006966400146484375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,32,1536,0.024053333202997844
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,1536,0.004565333326657614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,1536,0.00673173318306605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,32,1024,0.02273599902788798
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,1024,0.003994666785001755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,1024,0.006230400005976359
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,32,768,0.022373332579930624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,768,0.003651199986537298
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,768,0.006161066889762879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,32,512,0.021644800901412964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,512,0.0033386667569478357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,512,0.005832533538341522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,128,8192,0.01186240017414093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,32,128,0.0214464008808136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,32,256,0.02085439960161845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,256,0.0030613332986831666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,256,0.005657599866390228
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,128,0.002899199972550074
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,128,0.005538133283456167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,32,64,0.019900800784428914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,64,0.0027776000400384264
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,64,0.005629866818586985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,4,32,32,0.01882666746775309
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,32,0.0028234665592511495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,32,0.0054837331175804135
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,65536,16384,0.39542080561319987
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,65536,16384,0.38400745391845703
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,128,3072,0.007167999943097432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,65536,12288,0.2726069450378418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,65536,16384,0.4306922594706218
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,65536,12288,0.35190401077270506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,65536,16384,0.8124565124511719
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,65536,12288,0.30883092880249025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,65536,8192,0.18624746004740397
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,65536,12288,0.6116586685180664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,65536,10240,0.22950080235799156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,65536,10240,0.25081706047058105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,65536,8192,0.24183786710103355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,65536,10240,0.29378347396850585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,65536,10240,0.5109909375508626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,65536,8192,0.38066558837890624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,65536,8192,0.1964394728342692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,65536,7168,0.17211306889851888
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,65536,7168,0.3299178759256999
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,65536,7168,0.20577173233032225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,65536,6144,0.1478805383046468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,65536,6144,0.18849706649780273
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,65536,6144,0.14831466674804689
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,65536,6144,0.2843189239501953
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,65536,5120,0.11894506613413494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,65536,5120,0.16185280481974285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,65536,5120,0.13608320554097492
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,65536,5120,0.23598186175028482
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,65536,4096,0.0982645352681478
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,65536,4096,0.1310144027074178
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,65536,3584,0.11321173508961994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,65536,4096,0.10656533241271973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,65536,4096,0.19249173800150554
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,65536,3584,0.08817280133565267
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,65536,3584,0.16719786326090497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,65536,3072,0.10554133256276448
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,65536,3072,0.14361707369486493
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,65536,3072,0.07769386768341065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,65536,2560,0.08752000331878662
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,65536,2560,0.07213760217030843
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,65536,2560,0.06910826365152994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,65536,2048,0.07786880334218343
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,65536,2560,0.12010773022969563
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,65536,2048,0.050980265935262045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,65536,2048,0.09681173165639242
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,65536,2048,0.05770560105641683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,65536,1536,0.06537280082702637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,65536,1536,0.04092160065968831
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,65536,1536,0.07374719778696695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,65536,1536,0.04280639886856079
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,65536,1024,0.05414719978968302
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,65536,1024,0.0300053338209788
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,65536,1024,0.05014293193817139
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,65536,1024,0.030959999561309813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,65536,768,0.049746131896972655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,65536,768,0.023248000939687093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,65536,768,0.039477332433064775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,65536,768,0.025509333610534667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,65536,512,0.03906666835149129
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,65536,512,0.013611732920010885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,65536,512,0.0269269327322642
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,65536,512,0.0171615997950236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,65536,256,0.032476800680160525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,65536,256,0.009413333733876546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,65536,256,0.00775786687930425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,65536,256,0.014130133390426635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,65536,128,0.03013226588567098
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,65536,128,0.005569066603978475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,65536,128,0.011219200491905213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,65536,64,0.029009066025416058
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,65536,64,0.006550399959087372
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,65536,64,0.01122773289680481
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,65536,32,0.028812799851099653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,65536,32,0.0064650664726893115
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,65536,32,0.011218133568763732
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,16384,65536,0.4035221417744954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,65536,0.4164095878601074
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,16384,16384,0.10227519671122234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,65536,0.8001386642456054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,16384,65536,0.4697098731994629
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,16384,16384,0.14521387418111165
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,16384,0.19974400202433268
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,16384,0.11820906798044842
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,16384,12288,0.11126826604207356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,16384,12288,0.07730666796366373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,12288,0.15146133104960124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,12288,0.08918720086415609
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,16384,10240,0.09203306833902994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,16384,10240,0.06543039878209432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,10240,0.12632213433583578
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,10240,0.07096959749857584
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,16384,8192,0.07989973227183024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,16384,8192,0.053919998804728184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,8192,0.09399893283843994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,8192,0.058311466375986734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,16384,7168,0.04567573467890422
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,16384,7168,0.07729492982228597
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,7168,0.08262720108032226
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,7168,0.051376001040140784
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,16384,6144,0.03997439940770467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,16384,6144,0.07032960255940755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,6144,0.07253759702046712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,6144,0.04581226507822673
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,65536,7168,0.16113386154174805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,16384,5120,0.034983468055725095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,16384,5120,0.06357333262761435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,5120,0.060033067067464196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,5120,0.03832319974899292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,16384,4096,0.054347733656565346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,16384,4096,0.029785599311192828
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,4096,0.04889599879582723
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,16384,3584,0.05160640080769857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,4096,0.0321941335995992
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,16384,3584,0.027420800924301148
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,3584,0.04326613346735637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,3584,0.02890026569366455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,16384,3072,0.048529068628946945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,16384,3072,0.024551467100779215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,3072,0.037481598059336346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,3072,0.02717333237330119
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,65536,3584,0.08822080294291178
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,16384,2560,0.04285226662953694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,2560,0.037094398339589434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,2560,0.01984106699625651
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,65536,3072,0.07839146455128988
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,16384,2048,0.03107946713765462
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,16384,2048,0.012613333264986672
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,2048,0.027778132756551104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,2048,0.016474666198094685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,16384,1536,0.02904213269551595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,1536,0.02050559918085734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,1536,0.013822933038075766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,16384,1024,0.027037866910298664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,16384,1024,0.009385599692662557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,16384,768,0.026610134045283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,1024,0.008376533786455791
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,1024,0.011334400375684102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,16384,768,0.008022400240103405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,768,0.007153066496054332
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,768,0.010243200262387593
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,16384,512,0.025044266382853193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,512,0.005658666789531708
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,512,0.008846933643023174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,16384,256,0.024370133876800537
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,16384,256,0.005648000041643778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,256,0.004305066665013632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,256,0.007558399935563405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,16384,128,0.02285226583480835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,16384,128,0.005060266455014547
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,128,0.003640533238649368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,128,0.007233066856861115
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,16384,64,0.021602133909861244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,64,0.0033269333342711128
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,64,0.007130666573842366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,32,0.0073088000218073535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,16384,32,0.021623466412226358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,32,0.003352533280849457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,65536,128,0.007623466849327088
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,12288,16384,0.10700159867604572
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,12288,65536,0.3083072026570638
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,65536,0.305841064453125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,12288,16384,0.08586880366007486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,12288,65536,0.3649216016133626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,65536,0.5685237248738606
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,16384,0.1440597375233968
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,16384,0.09408960342407227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,12288,12288,0.08295146624247232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,12288,0.06234773397445679
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,12288,12288,0.06774293581644694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,12288,0.11928106943766277
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,12288,10240,0.07249813079833985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,12288,10240,0.0555733323097229
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,10240,0.08704319794972738
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,10240,0.05211520195007324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,12288,8192,0.0658730665842692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,12288,8192,0.04565226634343465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,8192,0.0734549363454183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,8192,0.04356480042139689
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,12288,7168,0.06015679836273193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,12288,7168,0.04372373421986898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,7168,0.06683733463287353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,7168,0.03860586484273275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,12288,6144,0.05527786811192831
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,12288,6144,0.035252265135447186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,6144,0.05471573273340861
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,6144,0.03404159943262736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,12288,5120,0.05034666856129964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,12288,5120,0.030723200241724653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,5120,0.045713067054748535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,12288,4096,0.04536960124969482
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,5120,0.029422932863235475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,4096,0.03821973403294881
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,4096,0.024935466051101685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,12288,3584,0.040151464939117434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,12288,3584,0.025225599606831867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,3584,0.034116268157958984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,3584,0.021236266692479452
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,12288,3072,0.03408000071843465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,12288,3072,0.02042133410771688
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,3072,0.029230932394663494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,3072,0.016875733931859337
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,12288,2560,0.02860693335533142
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,12288,2560,0.018148267269134523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,2560,0.024593067169189454
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,2560,0.014633599917093912
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,12288,2048,0.027157332499821978
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,12288,2048,0.013336533308029174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,2048,0.020626133680343627
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,2048,0.012731732924779257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,12288,1536,0.0116266667842865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,12288,1536,0.026495999097824095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,1536,0.011858133474985759
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,1536,0.011552000045776367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,12288,1024,0.02494293252627055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,12288,1024,0.009854933619499207
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,1024,0.008091733356316884
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,1024,0.009457066655158997
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,12288,768,0.023948800563812257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,16384,2560,0.018182400862375894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,12288,768,0.007831466694672901
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,768,0.006911999980608622
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,768,0.0086517333984375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,12288,512,0.022530132532119752
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,12288,512,0.006321066617965698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,512,0.0056309332450230915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,512,0.007709866762161255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,12288,256,0.022223999102910362
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,12288,256,0.0054624001185099285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,16384,1536,0.011571199695269267
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,128,0.0035861333211263022
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,256,0.0068693334857622785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,256,0.0042367999752362575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,12288,128,0.02071253259976705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,12288,128,0.004930133124192556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,128,0.006446933249632518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,12288,64,0.01955839991569519
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,64,0.0033781332274278007
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,64,0.006539733211199443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,12288,32,0.019182932376861573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,32,0.003382399926582972
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,32,0.006481066842873891
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,16384,512,0.006684799989064534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,10240,65536,0.24375893274943033
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,65536,0.26050453186035155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,10240,16384,0.08904319604237874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,10240,65536,0.3017002741495768
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,65536,0.4665866533915202
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,16384,0.06964053312937418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,16384,0.11778559684753417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,10240,12288,0.04920533498128255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,10240,12288,0.07298239866892496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,12288,0.1039082686106364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,12288,0.06186559995015463
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,10240,10240,0.06610453526178996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,10240,10240,0.04251413345336914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,10240,0.07529493172963461
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,10240,0.04635519981384277
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,10240,8192,0.05765120188395182
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,10240,8192,0.0344320019086202
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,8192,0.06283520062764486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,8192,0.0380351980527242
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,10240,7168,0.055048533280690516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,10240,6144,0.05063680013020834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,7168,0.053478399912516274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,7168,0.03466666539510091
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,10240,6144,0.02755413254102071
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,6144,0.04701120058695475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,6144,0.02966933250427246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,5120,0.039046398798624676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,10240,5120,0.04648213386535645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,10240,5120,0.02687573234240214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,5120,0.02629973292350769
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,10240,4096,0.039901868502298994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,10240,4096,0.018515199422836304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,4096,0.031922133763631184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,4096,0.022153600056966146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,10240,3584,0.034705066680908205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,10240,3584,0.020449066162109376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,3584,0.0283242662747701
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,3584,0.01722773313522339
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,10240,3072,0.0300053338209788
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,10240,3072,0.016849066813786825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,3072,0.02560746669769287
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,3072,0.015188266833623251
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,10240,2560,0.02832319935162862
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,10240,2560,0.014219733079274497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,2560,0.021899733940760294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,2560,0.013315199812253316
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,12288,4096,0.025500800212224322
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,10240,2048,0.0271232008934021
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,10240,2048,0.011732266346613566
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,2048,0.013724799950917563
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,2048,0.012057600418726604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,10240,1536,0.025672533114751178
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,10240,1536,0.010880000392595927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,1536,0.008291199803352356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,1536,0.01074133316675822
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,10240,1024,0.02406719923019409
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,10240,1024,0.00897173285484314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,1024,0.006483200192451477
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,1024,0.009027199943860371
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,10240,768,0.024498132864634196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,768,0.005566933254400889
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,768,0.008355200290679932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,10240,512,0.022373332579930624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,10240,512,0.0064085334539413456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,512,0.004669866462548574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,512,0.007640533149242401
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,10240,256,0.022139734029769896
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,10240,256,0.0054954667886098225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,256,0.0038453333079814913
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,256,0.006908800204594929
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,10240,128,0.020843732357025146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,10240,128,0.004986666639645894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,64,0.0031328000128269195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,64,0.006503466765085857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,128,0.00342399999499321
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,128,0.006499200065930684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,10240,64,0.019463467597961425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,10240,32,0.019272534052530925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,32,0.0031680000325044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,32,0.006559999783833821
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,8192,65536,0.20210986137390136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,8192,16384,0.07625813484191894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,65536,0.24366399447123208
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,8192,65536,0.24662399291992188
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,8192,16384,0.05298133293787638
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,65536,0.37241172790527344
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,16384,0.09222506682078044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,16384,0.06629653374354044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,8192,12288,0.04162773291269938
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,8192,12288,0.07057066758473715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,12288,0.07842559814453125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,12288,0.04776746829350789
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,8192,10240,0.057562665144602454
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,8192,10240,0.03576533397038777
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,10240,0.05888640085856119
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,10240,0.04247893492380778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,8192,8192,0.028616533676783247
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,8192,8192,0.05149439970652262
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,8192,0.047924268245697024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,8192,0.03466666539510091
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,8192,7168,0.049558401107788086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,10240,16384,0.06522560119628906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,8192,7168,0.028680533170700073
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,7168,0.042683732509613034
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,7168,0.03209386666615804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,8192,6144,0.04635733366012573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,8192,6144,0.024913066625595094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,6144,0.037264001369476316
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,6144,0.0264789342880249
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,8192,5120,0.0380021333694458
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,8192,5120,0.023092265923817953
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,5120,0.03142506678899129
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,5120,0.02195626695950826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,8192,4096,0.0299615999062856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,8192,4096,0.015396266182263692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,4096,0.025788799921671553
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,4096,0.01842666665712992
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,8192,3584,0.02829013268152873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,8192,3584,0.018518400192260743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,3584,0.022829866409301756
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,3584,0.015878400206565856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,8192,3072,0.027574400107065838
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,8192,3072,0.01585919956366221
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,10240,7168,0.03412160078684489
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,3072,0.020231467485427857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,3072,0.01402453382809957
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,8192,2560,0.026578134298324584
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,2048,0.008603733777999879
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,8192,2560,0.013304533561070761
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,2560,0.014631467064221701
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,2560,0.012753066420555115
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,8192,2048,0.02505706747372945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,8192,2048,0.010714667042096455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,2048,0.011664000153541566
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,8192,1536,0.024193066358566283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,8192,1536,0.010273067156473796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,1536,0.007426133255163829
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,1536,0.00990826686223348
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,8192,1024,0.023082667589187623
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,8192,1024,0.008413867155710856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,1024,0.0062282666563987735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,1024,0.00877333382765452
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,8192,768,0.023146667083104453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,8192,768,0.007098666826883952
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,768,0.005434666574001312
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,768,0.007979733248551685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,8192,512,0.02207039992014567
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,8192,512,0.005971199770768484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,512,0.004598399996757508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,512,0.007416533430417378
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,8192,256,0.021995733181635536
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,8192,256,0.005133866767088572
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,256,0.0037450666228930154
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,256,0.006775466601053874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,8192,128,0.020567466815312706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,64,0.0031136001149813337
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,8192,128,0.0047423998514811196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,128,0.003345066557327906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,128,0.006410666803518932
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,8192,64,0.019386667013168334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,64,0.006460799773534138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,10240,768,0.007536000013351441
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,8192,32,0.019130667050679527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,32,0.0030965333183606463
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,32,0.006548266609509785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,7168,65536,0.1790112018585205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,7168,16384,0.04904106855392456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,7168,16384,0.07001706759134928
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,65536,0.1850229263305664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,16384,0.050443732738494874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,16384,0.08435200055440267
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,7168,65536,0.20455679893493653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,65536,0.32125654220581057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,7168,12288,0.0605621337890625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,12288,0.038983468214670816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,12288,0.06997546354929605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,7168,10240,0.05375786622365316
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,10240,0.03401279846827189
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,10240,0.061127467950185144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,7168,8192,0.04801493485768636
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,7168,8192,0.027409066756566364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,8192,0.045450667540232345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,8192,0.028508800268173217
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,7168,0.04063786665598552
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,7168,7168,0.046562135219573975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,7168,0.025915733973185223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,7168,6144,0.04103999932607015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,6144,0.03604586521784465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,6144,0.023005867004394533
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,7168,5120,0.034330666065216064
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,7168,5120,0.018666666746139527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,5120,0.03075946569442749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,5120,0.01632426679134369
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,7168,4096,0.02983040014902751
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,7168,4096,0.014392532904942832
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,4096,0.023161600033442177
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,4096,0.01430506706237793
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,7168,3584,0.029074132442474365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,7168,3584,0.016687999169031777
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,3584,0.020846933126449585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,3584,0.01323199967543284
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,7168,3072,0.02770773371060689
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,7168,3072,0.013611732920010885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,3072,0.01623679995536804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,3072,0.012060800194740295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,7168,2560,0.026650667190551758
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,7168,2560,0.011738666892051696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,2560,0.009969066580136616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,2560,0.010983467102050781
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,7168,2048,0.025553067525227863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,7168,2048,0.009465600053469341
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,2048,0.008638933300971985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,2048,0.0098880002895991
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,7168,1536,0.02436479926109314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,7168,1536,0.009431466460227966
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,1536,0.007131733496983846
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,1536,0.008794666330019633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,7168,1024,0.023244800170262654
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,7168,1024,0.007777066528797149
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,1024,0.005690666536490122
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,1024,0.007776000102361043
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,7168,768,0.023229867219924927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,7168,768,0.006708266834417979
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,768,0.005046399931112925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,768,0.0072170664866765336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,7168,512,0.02184106707572937
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,7168,512,0.005795200169086456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,512,0.004295466840267182
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,512,0.006723199784755707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,7168,256,0.02168853282928467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,7168,256,0.0050570666790008545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,256,0.0036138666172822317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,256,0.0061706667145093284
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,7168,128,0.020549333095550536
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,7168,128,0.0046858668327331545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,128,0.003257599969704946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,128,0.005892266829808554
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,7168,64,0.019160532951354982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,64,0.003017599880695343
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,64,0.0059008002281188965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,7168,32,0.01882986625035604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,32,0.003048533449570338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,32,0.005851733187834421
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,6144,65536,0.1801418622334798
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,6144,65536,0.1896704037984212
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,65536,0.27962665557861327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,65536,0.1573098659515381
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,6144,16384,0.06463040113449096
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,6144,16384,0.04958933194478353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,16384,0.07446719805399576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,16384,0.04347626765569051
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,6144,12288,0.05594346523284912
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,6144,12288,0.036806400616963705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,12288,0.062300801277160645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,12288,0.03785706758499145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,6144,10240,0.04998826583226522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,10240,0.04608000119527181
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,10240,0.029998934268951415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,6144,8192,0.04538880189259847
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,6144,8192,0.025847466786702473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,8192,0.03837546507517497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,8192,0.024745599428812663
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,6144,7168,0.04368533293406169
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,6144,7168,0.028793599208196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,7168,0.03410666783650716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,7168,0.022658133506774904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,6144,6144,0.021574399868647256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,6144,6144,0.035138134161631265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,6144,0.02945173382759094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,6144,0.01813439925511678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,6144,5120,0.031052800019582112
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,6144,5120,0.018329600493113198
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,5120,0.024682666858037314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,5120,0.015497600038846334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,6144,4096,0.014105600118637086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,6144,4096,0.029205334186553956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,7168,12288,0.04044266541798909
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,4096,0.02139093279838562
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,7168,10240,0.034126933415730795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,4096,0.012964266538619994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,6144,3584,0.029242666562398274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,6144,3584,0.018167465925216675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,3584,0.017685333887736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,3584,0.01258133351802826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,6144,3072,0.0275221327940623
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,6144,3072,0.014413866400718688
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,3072,0.011168000102043153
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,3072,0.011585066715876263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,6144,2560,0.026693334182103474
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,6144,2560,0.012956800063451132
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,7168,7168,0.02834026614824931
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,2560,0.009798399607340495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,2560,0.010475732882817586
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,7168,6144,0.02318613330523173
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,6144,2048,0.024949334065119424
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,6144,2048,0.010190932949384054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,2048,0.008280533552169799
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,2048,0.009417600433031718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,6144,1536,0.02478613257408142
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,6144,1536,0.010090667009353637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,1536,0.007191466788450877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,1536,0.008534399668375652
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,6144,1024,0.022937599817911783
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,768,0.00479360024134318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,6144,1024,0.008318933347860973
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,1024,0.00557226687669754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,1024,0.007506133119265239
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,6144,768,0.023059199253718056
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,768,0.00699839989344279
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,6144,512,0.021633066733678184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,6144,512,0.0059114664793014525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,512,0.004129066566626231
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,512,0.006542933483918508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,6144,256,0.02169493238131205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,6144,256,0.005133866767088572
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,256,0.0035264000296592714
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,128,0.005834666887919108
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,256,0.0061247999469439185
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,6144,128,0.020487467447916664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,6144,128,0.004782933493455251
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,128,0.003234133372704188
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,64,0.0029909332593282064
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,6144,64,0.019059199094772338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,64,0.005901866654555003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,6144,32,0.018847999970118205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,32,0.00297173336148262
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,32,0.005769599974155426
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,5120,65536,0.15677760442097982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,5120,65536,0.1718250592549642
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,5120,16384,0.05956693490346273
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,65536,0.13062933286031086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,5120,16384,0.04496746857961019
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,65536,0.23334186871846518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,16384,0.06048426628112793
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,16384,0.03829866647720337
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,5120,12288,0.05094720125198364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,5120,10240,0.04669333299001058
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,5120,12288,0.03597653309504191
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,12288,0.04571093320846557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,12288,0.03237013419469197
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,5120,10240,0.030187733968098956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,10240,0.04069866736729939
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,10240,0.026958932479222614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,5120,8192,0.04042239983876546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,5120,8192,0.024603732426961265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,5120,7168,0.03581759929656982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,8192,0.031618134180704756
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,8192,0.019077332814534505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,5120,7168,0.02421440084775289
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,7168,0.028492800394694012
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,7168,0.0166101336479187
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,5120,6144,0.032085333267847696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,6144,0.014579199751218162
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,5120,5120,0.030364799499511718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,5120,6144,0.017612799008687337
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,6144,0.025095466772715253
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,5120,5120,0.016402133305867515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,5120,0.021522132555643718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,5120,0.013389866550763449
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,5120,4096,0.028859732548395793
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,4096,0.016872533162434897
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,4096,0.012034133076667786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,5120,3584,0.02784213423728943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,5120,3584,0.01535360018412272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,3584,0.010190932949384054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,3584,0.011239467064539592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,5120,3072,0.026774400472640993
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,5120,3072,0.01262079974015554
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,3072,0.009260800480842591
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,5120,2560,0.02581760088602702
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,6144,10240,0.03226240078608195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,3072,0.010444800059000652
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,5120,2560,0.011110400160153706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,5120,2048,0.024642133712768556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,5120,2048,0.010033067067464192
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,2560,0.008228266735871632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,2560,0.009562666217486065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,2048,0.007355733215808869
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,2048,0.008775466680526733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,5120,1536,0.02376746733983358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,5120,1536,0.009477333227793375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,1536,0.006119466821352641
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,1536,0.008162133395671844
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,5120,1024,0.02276479999224345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,5120,1024,0.00788373351097107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,1024,0.004981333514054617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,1024,0.007146666447321574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,5120,768,0.022861866156260173
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,5120,768,0.006626133124033611
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,768,0.004474666714668274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,768,0.006759466727574666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,5120,512,0.021640533208847047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,512,0.0039007999002933502
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,256,0.0035114665826161706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,512,0.006403199831644694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,5120,256,0.021627734104792275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,256,0.006043733159701029
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,5120,128,0.02044586737950643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,5120,128,0.004692266881465912
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,128,0.0032159999012947083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,128,0.005840000013510386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,5120,64,0.019103999932607016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,64,0.002963199963172277
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,64,0.005805866420269012
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,5120,32,0.018806399901707967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,32,0.002995199958483378
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,32,0.00582826683918635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,4096,65536,0.15992852846781414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,4096,65536,0.13420160611470539
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,65536,0.10845013459523518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,65536,0.1929983933766683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,4096,16384,0.05352426767349243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,16384,0.05067946513493856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,4096,16384,0.04107413291931152
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,16384,0.03237546682357788
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,4096,12288,0.04689706563949585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,6144,768,0.006854400038719177
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,4096,12288,0.03252373337745666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,12288,0.039895466963450116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,12288,0.026445867617925008
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,4096,10240,0.041595733165740965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,4096,10240,0.02752853234608968
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,10240,0.03394026756286621
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,10240,0.020778665939966835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,4096,8192,0.03243626753489177
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,4096,8192,0.021849600474039714
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,8192,0.02827306588490804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,8192,0.014497066537539164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,4096,7168,0.030162133773167926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,7168,0.0256661335627238
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,7168,0.013689600427945457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,4096,6144,0.029339732726414998
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,4096,6144,0.018474666277567546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,6144,0.02271359960238139
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,6144,0.0123690664768219
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,4096,5120,0.02804479996363322
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,4096,5120,0.01751040021578471
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,5120,0.01611306667327881
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,5120,0.011460266510645549
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,4096,4096,0.026874667406082152
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,4096,4096,0.013112533092498779
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,4096,0.010381866494814556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,4096,0.010345600048700969
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,4096,3584,0.017336533466974894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,4096,3584,0.025975465774536133
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,3584,0.008675199747085572
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,3584,0.009774933258692425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,4096,3072,0.025500800212224322
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,4096,3072,0.013864533106486002
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,3072,0.007844266792138416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,3072,0.009190400441487629
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,4096,2560,0.02510720094045003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,4096,2560,0.01179093321164449
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,2560,0.007092266778151194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,2560,0.008613333106040955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,4096,2048,0.02358186642328898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,4096,2048,0.008975999553998311
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,2048,0.0062282666563987735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,2048,0.00796693315108617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,4096,1536,0.023117866118748984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,4096,1536,0.009638399879137675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,1536,0.0052597333987553915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,1536,0.00735040009021759
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,4096,1024,0.022938666741053264
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,4096,1024,0.008117333551247915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,1024,0.004508799811204275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,1024,0.006810666620731353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,4096,768,0.02244159976641337
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,4096,768,0.006622933348019918
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,5120,4096,0.01378986636797587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,768,0.004120533416668574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,768,0.006424533327420552
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,4096,512,0.02136853337287903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,512,0.003836799909671148
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,512,0.0061589335401852924
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,4096,256,0.02149440050125122
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,4096,256,0.005008000135421753
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,256,0.003311999887228012
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,256,0.005924266576766968
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,4096,128,0.020426666736602782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,4096,128,0.004665599763393402
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,128,0.003125333289305369
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,128,0.005724800129731497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,4096,64,0.019257599115371705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,64,0.0029365333418051405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,64,0.005724800129731497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,4096,32,0.018768000602722167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,32,0.002977066735426585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,32,0.005684266487757364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3584,65536,0.14546027183532714
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3584,65536,0.1247167984644572
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,65536,0.09810986518859863
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,65536,0.17202773094177246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3584,16384,0.040115201473236085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,5120,512,0.005691733459631602
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3584,16384,0.0507914662361145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,16384,0.045201067129770914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,16384,0.029268266757329305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3584,12288,0.04482239882151286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3584,12288,0.031243733565012616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,5120,256,0.004983466863632202
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,12288,0.03537493149439494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,12288,0.02404586672782898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3584,10240,0.03439360062281291
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3584,10240,0.025554132461547852
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,10240,0.030536532402038574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,10240,0.016546133160591125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3584,8192,0.030024532477060956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3584,8192,0.021118932962417604
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,8192,0.02606186668078105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,8192,0.013303466637929282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3584,7168,0.030075732866923017
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3584,7168,0.023320533831914268
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,7168,0.02314880092938741
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,7168,0.012560000022252401
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3584,6144,0.02917120059331258
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3584,6144,0.017604267597198485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,6144,0.020141865809758505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,6144,0.011725866794586181
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3584,5120,0.02809813419977824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3584,5120,0.01590079963207245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,5120,0.011586133639017742
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,5120,0.010921600461006164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3584,4096,0.026748800277709962
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3584,4096,0.0127402663230896
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,4096,0.008738133311271667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,4096,0.009758933385213216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3584,3584,0.02640639940897624
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3584,3584,0.015449600418408713
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,3584,0.008041599889596303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,4096,7168,0.020297600825627645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,3584,0.009459199508031209
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3584,3072,0.025409066677093507
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3584,3072,0.01372160017490387
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,3072,0.007864533364772797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,3072,0.008913066983222962
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3584,2560,0.02508266568183899
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,2560,0.0065984000762303666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,2560,0.008386133114496867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3584,2048,0.023717333873113
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3584,2048,0.008381866415341695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,2048,0.005898666878541311
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,2048,0.007656533519426982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3584,1536,0.023176532983779908
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3584,1536,0.00904746651649475
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,1536,0.005211733281612396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,1536,0.007146666447321574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3584,1024,0.022252800067265828
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3584,1024,0.0075765331586201985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,1024,0.004494933287302653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,1024,0.0067221333583196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3584,768,0.02244480053583781
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,768,0.004133333265781402
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,768,0.006372266511122386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3584,512,0.021656533082326256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3584,512,0.005603200197219849
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,512,0.0037632000943024955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,512,0.006129066646099091
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3584,256,0.021588265895843506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,256,0.0033621333539485933
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,256,0.0059232001503308615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3584,128,0.020399999618530274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,128,0.003070933371782303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,128,0.005727999905745188
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3584,64,0.018938666582107543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,64,0.0029450667401154833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,64,0.005725866556167603
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3584,32,0.018976000944773357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,32,0.002940800040960312
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,32,0.005790933469931285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3072,65536,0.13149333000183105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3072,65536,0.11611733436584473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,65536,0.15025919278462727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,4096,512,0.005650133391221364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,65536,0.08793173631032308
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3072,16384,0.04827200174331665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3072,16384,0.038483198483784994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,16384,0.03999253511428833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,16384,0.02722666660944621
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3072,12288,0.037514666716257736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3072,12288,0.0287445326646169
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,12288,0.03161173264185588
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,12288,0.01814613342285156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3072,10240,0.031779199838638306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3072,10240,0.02369706630706787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,10240,0.02726400097211202
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,10240,0.015410133202870688
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3072,8192,0.02969493269920349
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3072,8192,0.019678932428359986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,8192,0.02259413401285807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,8192,0.013198933005332947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3072,7168,0.029953066507975263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3072,7168,0.01968533396720886
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,7168,0.01994133392969767
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,7168,0.012544000148773193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3072,6144,0.028317866722742717
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3072,6144,0.016190933187802632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,6144,0.01199893355369568
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,6144,0.011597866813341778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3072,5120,0.0276853342851003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3072,5120,0.014333867033322654
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,5120,0.01060693363348643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,5120,0.010838400324185688
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3072,4096,0.026371200879414875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3072,4096,0.012075733145078022
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,4096,0.010469333330790202
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,4096,0.009796266754468281
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3072,3584,0.02606613238652547
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3072,3584,0.014101333419481912
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3072,3072,0.02518613338470459
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,3584,0.009709866841634114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,3584,0.009444266557693481
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3072,3072,0.011366400122642516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,3072,0.008866133292516072
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,3072,0.008836266398429871
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3072,2560,0.0250709335009257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3072,2560,0.009986133376757304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,2560,0.006461866696675618
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,2560,0.008294400076071422
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3072,2048,0.02371306618054708
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3072,2048,0.008099199831485748
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,2048,0.005854933460553487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,2048,0.00765119989713033
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3072,1536,0.02316373387972514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3072,1536,0.008534399668375652
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,1536,0.005188266436258951
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,1536,0.007050666709740956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3072,1024,0.02219946583112081
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3072,1024,0.007269333302974701
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,1024,0.00448639988899231
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,1024,0.006569600105285645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3072,768,0.02241599957148234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3072,768,0.006226133306821187
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,768,0.004124800115823746
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,768,0.006347733239332835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3072,512,0.021383466323216756
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3072,512,0.005491200089454651
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,512,0.0038005332152048744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,512,0.006108800073464712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3072,256,0.021499733130137123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3072,256,0.004855466882387797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,256,0.0033589333295822145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,256,0.005942399799823761
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3584,2560,0.011224533120791119
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3072,128,0.02036799987157186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3072,128,0.004445866743723551
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,128,0.0030954666435718536
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,128,0.005679999788602194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3072,64,0.019271467129389444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,64,0.0028864001234372456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,64,0.005692799886067709
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,3072,32,0.018669867515563966
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,32,0.0028949332733949023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,32,0.005762133498986562
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2560,65536,0.11514773368835449
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2560,65536,0.10461013317108155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,65536,0.12302827040354411
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,65536,0.08050133387247721
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2560,16384,0.04569493134816487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2560,16384,0.038369067509969074
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3584,768,0.006309333443641663
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,16384,0.033930667241414386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,16384,0.023908267418543495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2560,12288,0.033240532875061034
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2560,12288,0.02845653295516968
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,12288,0.027926399310429888
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,12288,0.01578133304913839
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2560,10240,0.03146026730537414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,10240,0.023526400327682495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,10240,0.014109866817792258
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2560,8192,0.030158933003743487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2560,8192,0.020897066593170165
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3584,256,0.004946133494377137
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,8192,0.017163733641306557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,8192,0.012283733487129212
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3584,128,0.0045162667830785114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2560,7168,0.029925332466761274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2560,7168,0.018513067563374837
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,7168,0.01230506698290507
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,7168,0.011612799763679505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2560,6144,0.028794666131337483
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,6144,0.011464533209800721
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,6144,0.010813867052396137
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2560,5120,0.02762453357378642
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2560,5120,0.014057599504788718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,5120,0.010282666484514872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,5120,0.010050132870674133
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2560,4096,0.026763733228047686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2560,4096,0.011315199732780456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,4096,0.009170132875442504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,4096,0.008956799904505413
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2560,3584,0.026233599583307905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,3584,0.008710400263468424
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,3584,0.00855573316415151
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,3072,0.00811839997768402
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2560,3072,0.02513386607170105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2560,3072,0.010948266585667927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,3072,0.008289066453774769
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2560,2560,0.02505600055058797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,2560,0.006398933132489522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,2560,0.007760000228881836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2560,2048,0.023715200026830037
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2560,2048,0.007768533130486806
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,2048,0.005724800129731497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,2048,0.007334400216738384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2560,1536,0.023038933674494423
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2560,1536,0.00803413341442744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,1536,0.005064533154169718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,1536,0.006922666728496551
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2560,1024,0.022165334224700926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2560,1024,0.006797866523265838
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,1024,0.004522666831811269
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,1024,0.006532266736030579
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2560,768,0.022423466046651207
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2560,768,0.0059125334024429325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,768,0.00403413325548172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,768,0.006340266764163971
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2560,512,0.021409066518147786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2560,512,0.00525546669960022
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,512,0.003671466559171677
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,512,0.006122666597366333
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2560,256,0.021504000822703043
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2560,256,0.004730666677157084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,256,0.0033183999359607695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,256,0.005875200033187866
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2560,128,0.020323199033737183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2560,128,0.004413866500059763
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,128,0.003028266628583272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,128,0.005678933362166087
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2560,64,0.01899626652399699
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,64,0.002916266769170761
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,64,0.005643733342488607
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2560,32,0.018692266941070557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,32,0.0029696000119050344
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,32,0.005726933479309082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2048,65536,0.09809280236562093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2048,65536,0.13124159971872967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,65536,0.07229973475138346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,65536,0.09759573141733804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2048,16384,0.03851840098698934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,16384,0.028155734141667683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2048,16384,0.03369919856389363
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,16384,0.01949333349863688
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2048,12288,0.033504001299540204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,12288,0.022501333554585775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,12288,0.0156768004099528
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2048,10240,0.030843732754389445
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2048,10240,0.023885866006215416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,10240,0.01727573275566101
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,10240,0.01404159963130951
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2048,8192,0.029611732562383013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2048,8192,0.019474132855733236
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,8192,0.010236799716949463
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,8192,0.012174933155377706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2048,7168,0.029731200138727827
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2048,7168,0.02214933236440023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,7168,0.00993066628774007
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,7168,0.011431466539700825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2048,6144,0.028266666332880656
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2048,6144,0.016521599888801575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,6144,0.009086933732032777
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,6144,0.010629333058993022
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2048,5120,0.027616000175476073
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2048,5120,0.016500266393025716
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,5120,0.008705066641171773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,5120,0.009839999675750732
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2048,4096,0.026343466838200887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2048,4096,0.011879466970761617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,4096,0.007924266656239827
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,4096,0.008762666583061218
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2048,3584,0.025906133651733398
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2048,3584,0.016337066888809204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,3584,0.0076799998680750535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,3584,0.008386133114496867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2048,3072,0.025201066335042314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2048,3072,0.013329066832860312
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,3072,0.007342933118343354
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,3072,0.008081066608428954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2048,2560,0.024378667275110878
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2048,2560,0.010789333780606588
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,2560,0.006257066627343495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,2560,0.007573333382606506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2048,2048,0.02407146692276001
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,2048,0.005584000051021576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2048,2048,0.008973866701126099
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,2048,0.007156266768773396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2048,1536,0.02307200034459432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2048,1536,0.00951573352018992
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,1536,0.00491839994986852
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,1536,0.006754133105278015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2560,10240,0.023257599274317423
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2048,1024,0.022132267554601036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2048,1024,0.007769600053628286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,1024,0.004216533402601878
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,1024,0.006417066852251689
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2048,768,0.022374399503072104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2048,768,0.006526933113733928
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,768,0.003940266619126002
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,768,0.006226133306821187
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2048,512,0.021259733041127524
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2048,512,0.00559146652619044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,512,0.0035594666997591654
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,512,0.00598826656738917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2048,256,0.021316266059875487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2048,256,0.00491839994986852
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,256,0.003234133372704188
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,256,0.0057888001203536986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2048,128,0.020222934087117513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2048,128,0.004580266773700714
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2560,6144,0.015147733688354491
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,128,0.003011200080315272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,128,0.0056650668382644655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2048,64,0.018931200106938682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,64,0.002885333448648453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,64,0.005627733469009399
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,2048,32,0.018747733036677042
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,32,0.002880000074704488
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2560,3584,0.01337493360042572
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,32,0.0056426664193471275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,65536,0.07294080257415772
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1536,65536,0.09198079903920492
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1536,65536,0.12285439968109131
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,65536,0.06750826835632324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1536,16384,0.0365994652112325
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1536,16384,0.03218560020128886
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,16384,0.022834134101867676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,16384,0.019057067235310872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2560,2560,0.010072533289591472
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1536,12288,0.03319360017776489
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1536,12288,0.025336533784866333
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,12288,0.01344533363978068
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,12288,0.015719466408093772
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1536,10240,0.03140160044034322
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1536,10240,0.021545600891113282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,10240,0.012035199999809265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,10240,0.01406719982624054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1536,7168,0.0168938676516215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1536,8192,0.029614933331807453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1536,8192,0.018553600708643595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,8192,0.01032426655292511
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,8192,0.012268799543380737
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1536,7168,0.02991360028584798
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,7168,0.00981226662794749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,7168,0.011529599626859028
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1536,6144,0.028735999266306562
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1536,6144,0.014176000157992044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,6144,0.009250133236249288
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,6144,0.010665599505106609
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1536,5120,0.027704532941182452
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1536,5120,0.01367573340733846
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,5120,0.008467200398445129
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,5120,0.009866666793823243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1536,4096,0.026653865973154705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1536,4096,0.010990933577219645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,4096,0.007706666489442189
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,4096,0.008781866232554118
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1536,3584,0.02608213424682617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1536,3584,0.01388159990310669
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,3584,0.0074442664782206226
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,3584,0.008389332890510559
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1536,3072,0.025465599695841473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,3072,0.00714026689529419
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,3072,0.008050133287906647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1536,2560,0.024990934133529662
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1536,2560,0.009495466947555542
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,2560,0.00622080018122991
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,2560,0.007660800218582153
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1536,2048,0.023560533920923867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1536,2048,0.008246399958928426
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,2048,0.00562666654586792
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,2048,0.007089066505432129
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1536,1536,0.023065600792566934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1536,1536,0.008206933240095774
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,1536,0.004917333523432413
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,1536,0.006853333115577698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1536,1024,0.022190932432810465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2048,12288,0.02622399926185608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1536,1024,0.007181866466999054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,1024,0.004216533402601878
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,1024,0.006365866462389629
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1536,768,0.023085866371790567
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,512,0.0035562666753927866
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,768,0.003914666672547659
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,768,0.006166400015354156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1536,512,0.02132586638132731
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,512,0.0060362666845321655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1536,256,0.021496532360712688
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1536,256,0.004706133405367533
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,256,0.0032287999987602235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,256,0.005831466615200042
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1536,128,0.02036799987157186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1536,128,0.004401066899299621
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,128,0.0030570665995279947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,128,0.00559146652619044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1536,64,0.019004799922307334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,64,0.002902399996916453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,64,0.0056970665852228795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1536,32,0.018630399306615194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,32,0.0028629332780838014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,32,0.005703466633955637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1024,65536,0.0877247969309489
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1024,65536,0.09725440343221028
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,65536,0.0514847993850708
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,65536,0.06550506750742593
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1024,16384,0.03675306638081868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1024,16384,0.031506133079528806
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,16384,0.011828266580899556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,16384,0.019002666076024376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1024,12288,0.03306986689567566
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1024,12288,0.023661865790685018
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,12288,0.01055573324362437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,12288,0.015608533223470052
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1024,10240,0.031336534023284915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1024,10240,0.02016426722208659
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,10240,0.009348266323407491
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,8192,0.008410666386286418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,10240,0.01399679978688558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1024,8192,0.029578665892283123
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1024,7168,0.029678932825724286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,8192,0.012246400117874146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,7168,0.011370666821797689
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1024,7168,0.016782933473587038
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,7168,0.008138666550318401
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1024,6144,0.028649600346883138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1024,6144,0.013770666718482972
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,6144,0.007107200225194295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,6144,0.010602666934331258
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1024,5120,0.027642667293548584
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1024,5120,0.011678933103879293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,5120,0.007298133273919423
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,5120,0.009832533200581868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1024,4096,0.026537599166234334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1024,4096,0.010648533701896667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,4096,0.006980266670385997
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,4096,0.00872320036093394
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1024,3584,0.02608533302942912
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1024,3584,0.011427199840545655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,3584,0.0073173334201176955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,3584,0.008340266346931458
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1024,3072,0.025319466988245647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1024,3072,0.009500799576441447
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,3072,0.00803413341442744
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,3072,0.006977066894372304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1024,2560,0.024778666098912557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1024,2560,0.008563199639320373
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,2560,0.00625600020090739
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,2560,0.007565866907437642
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1024,2048,0.02364693284034729
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1024,2048,0.0071285332242647815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,2048,0.005555200080076853
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,2048,0.00716480016708374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1024,1536,0.023448532819747923
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1024,1536,0.007260799904664357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,1536,0.004884266853332519
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,1536,0.006759466727574666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1024,1024,0.022107734282811483
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1024,1024,0.006321066617965698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,1024,0.004212266703446706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,1024,0.006321066617965698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1024,768,0.022459733486175536
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1024,768,0.005674666663010915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,768,0.0038581334054470064
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,768,0.00622506688038508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1024,512,0.021639466285705566
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1024,512,0.004999466737111409
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,512,0.0035477332770824433
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,512,0.005987200140953064
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1024,256,0.0214848001797994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1024,256,0.004691199958324432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,256,0.0031925333042939507
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,256,0.00572266678015391
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1024,128,0.020242132743199668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1024,128,0.004386133452256521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,128,0.0029781334102153777
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,128,0.005625600119431814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1024,64,0.01918506622314453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,64,0.0031957333286603295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,64,0.005629866818586985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,1024,32,0.01889280080795288
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,32,0.003180799881617228
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,32,0.005592533449331919
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,768,65536,0.08393386999766031
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,768,65536,0.0867637316385905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,65536,0.0398911992708842
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,65536,0.06530986626942953
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,768,16384,0.036831998825073244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,768,16384,0.022204800446828207
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,16384,0.010630399982134501
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,16384,0.01880319913228353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,768,12288,0.03309973279635112
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,768,12288,0.018525866667429607
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1536,3072,0.012306132912635803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,768,10240,0.03141760031382243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,12288,0.009256533781687419
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,12288,0.015518933534622192
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,768,10240,0.01656000018119812
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,10240,0.008540800213813782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,10240,0.014043733477592468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,768,8192,0.029603199164072676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,768,8192,0.015434666474660238
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,8192,0.007771733403205872
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,8192,0.012295466661453248
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,768,7168,0.02977493405342102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,768,7168,0.015958399573961893
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,7168,0.007154133419195812
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,7168,0.011382399996121725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,768,6144,0.028967465957005816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,768,6144,0.012114133437474568
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,6144,0.006898133456707001
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,6144,0.010570666193962098
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1536,768,0.005964800218741099
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,768,5120,0.02763306697209676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,5120,0.007268266876538594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,5120,0.009844266374905904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1536,512,0.005245866874853769
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,768,4096,0.02649173339207967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,768,4096,0.009437867005666097
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,4096,0.006810666620731353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,4096,0.00878613293170929
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,768,3584,0.026128000020980834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,3584,0.007181866466999054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,3584,0.008342400193214417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,768,3072,0.0255840003490448
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,768,3072,0.009461333354314167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,3072,0.006954666475454967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,3072,0.00796693315108617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,768,2560,0.024938666820526124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,768,2560,0.007667199770609538
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,2560,0.006200533111890157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,2560,0.007586133480072021
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,768,2048,0.023534933725992836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,768,2048,0.007368533313274384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,2048,0.005553066730499268
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,2048,0.007238399982452392
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,768,1536,0.02302293380101522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,768,1536,0.006625066697597504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,1536,0.004874666531880697
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,1536,0.006854400038719177
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,768,1024,0.02221440076828003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,768,1024,0.006067200005054474
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,1024,0.004226133227348328
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,1024,0.006274133423964183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,768,768,0.022973867257436116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,768,768,0.005350400010744731
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,768,0.003836799909671148
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,768,0.006154666841030121
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,768,512,0.021637332439422608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,768,512,0.004974933465321859
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,512,0.003542399903138479
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,512,0.00594346672296524
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,768,256,0.021492266654968263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,768,256,0.004594133297602335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1024,8192,0.01859626571337382
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,256,0.003389866650104523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,256,0.005737600227197012
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,768,128,0.0203167994817098
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,768,128,0.004364799956480662
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,128,0.0031690667072931922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,128,0.0055637334783871974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,768,64,0.018962132930755615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,64,0.0030016000072161358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,64,0.005693866809209188
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,768,32,0.018588799238204955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,32,0.0029493334392706556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,32,0.005804799993832906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,512,65536,0.08223360379536947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,512,65536,0.07727680206298829
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,65536,0.028681600093841554
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,65536,0.06398933331171672
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,512,16384,0.036602667967478436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,512,16384,0.019336533546447755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,16384,0.008729599912961324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,16384,0.018905599912007652
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,512,12288,0.032843732833862306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,512,12288,0.015991466244061787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,12288,0.007742933432261149
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,12288,0.015471999843915304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,512,10240,0.03127359946568807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,512,10240,0.014387200276056925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,10240,0.007191466788450877
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,10240,0.013849600156148275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,512,8192,0.02979733347892761
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,512,8192,0.011954133709271748
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,7168,0.006917333106199901
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,8192,0.007180800040562947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,8192,0.012083199620246888
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,512,7168,0.02954240043958028
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,512,7168,0.011731200416882833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,7168,0.011364266276359558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,512,6144,0.028838400046030683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,512,6144,0.010611200332641601
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,6144,0.006712533533573151
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,6144,0.010550399621327717
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,512,5120,0.02770133415857951
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,512,5120,0.009112532933553059
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,5120,0.0071722666422526045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,5120,0.009889066219329834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,512,4096,0.026390399535497027
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,512,4096,0.008134399851163227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,4096,0.006696533163388569
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,4096,0.008709333340326945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,512,3584,0.025890133778254193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,512,3584,0.009087999661763508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,3584,0.007133866846561432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,3584,0.008298666775226593
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,512,3072,0.025059199333190917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,512,3072,0.008354133367538452
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,3072,0.0068906664848327635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,3072,0.0079925333460172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,512,2560,0.024701867500940958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,512,2560,0.007309866448243459
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,2560,0.006353066861629486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,2560,0.007562666634718577
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,512,2048,0.023897600173950196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,2048,0.005670399963855743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,2048,0.007051733136177063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,512,1536,0.02297813296318054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,512,1536,0.006253866851329804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,1536,0.005018666883309682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,1536,0.006675200164318084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,512,1024,0.022101332743962608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,512,1024,0.005655466516812643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,1024,0.004342400034268697
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,1024,0.0063733334342638654
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,512,768,0.022346667448679605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,512,768,0.005122133096059163
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,768,0.003949866692225138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,768,0.006116266548633576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,512,512,0.021575466791788737
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,512,0.003634133438269297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,512,0.00584853341182073
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,512,256,0.021314134200414024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,512,256,0.004554666578769684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,256,0.0032831999162832894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,256,0.005702400207519531
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,512,128,0.02023573319117228
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,512,128,0.004349866509437561
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,128,0.0030570665995279947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,128,0.005577600002288819
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,64,0.0056202664971351625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,512,64,0.01890346606572469
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,64,0.0029440000653266907
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,512,32,0.018693333864212035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,32,0.002919466545184453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,32,0.005655466516812643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,256,65536,0.07900479634602865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,256,65536,0.06509759823481241
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,65536,0.012469333410263062
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,65536,0.059305600325266515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,256,16384,0.0366431991259257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,256,16384,0.019336533546447755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,16384,0.00695360004901886
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,16384,0.01874026656150818
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,256,12288,0.03338453372319539
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,256,12288,0.015642666816711427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,12288,0.006810666620731353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,12288,0.015544533729553223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,256,10240,0.03141226569811503
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,256,10240,0.013745066523551942
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,768,5120,0.010230400164922078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,10240,0.007186133166154225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,10240,0.013867732882499696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,256,8192,0.029527467489242554
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,8192,0.007048533360163371
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,8192,0.012034133076667786
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,256,7168,0.02970133423805237
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,256,7168,0.010937600334485372
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,7168,0.00678719977537791
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,7168,0.011262933413187664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,256,6144,0.028833067417144774
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,768,3584,0.010327466328938802
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,256,6144,0.010016000270843506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,6144,0.006730666756629944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,6144,0.010546132922172546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,256,5120,0.02772480050722758
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,256,5120,0.00909546713034312
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,5120,0.007122133175532024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,5120,0.009679999947547913
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,256,4096,0.026625066995620728
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,256,4096,0.008065066734949748
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,4096,0.006694399813810985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,4096,0.008596266309420269
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,256,3584,0.025625600417455034
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,256,3072,0.025332266092300416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,256,3584,0.007603199779987335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,256,3072,0.007399466633796692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,3584,0.007000533243020375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,3584,0.008265600105126699
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,3072,0.006857599814732869
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,3072,0.008010666569073994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,256,2560,0.02480640014012655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,256,2560,0.00653546651204427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,2560,0.006233599781990051
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,2560,0.007506133119265239
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,256,2048,0.023961599667867026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,256,2048,0.006244266529877981
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,2048,0.005553066730499268
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,2048,0.007032533486684163
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,256,1536,0.022963200012842813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,256,1536,0.005794133245944977
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,1536,0.004881066580613455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,1536,0.0066655998428662615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,256,1024,0.021953066190083824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,1024,0.0042357335488001505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,1024,0.00624533345301946
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,256,768,0.02248106598854065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,256,768,0.004901333153247834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,768,0.003882666677236557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,768,0.006017066538333893
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,256,512,0.021294933557510377
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,256,512,0.004594133297602335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,512,0.003562666724125544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,512,0.005831466615200042
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,256,256,0.021348265806833903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,256,0.005657599866390228
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,256,256,0.004543999830881754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,256,0.0032032000521818793
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,256,128,0.02018986741701762
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,256,128,0.004260266820589701
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,128,0.002994133283694585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,128,0.005645866692066193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,256,64,0.018901334206263224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,64,0.0028746667007605235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,64,0.005514666438102722
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,256,32,0.0189194659392039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,32,0.002881066749493281
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,128,65536,0.07904533545176187
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,32,0.005554133156935374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,128,65536,0.06586453517278036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,65536,0.010020266969998677
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,65536,0.05958720048268636
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,128,16384,0.03664213418960571
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,128,16384,0.019402666886647543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,16384,0.00689279983441035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,16384,0.01876266598701477
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,128,12288,0.032985599835713704
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,128,12288,0.015787733594576518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,12288,0.006826666494210561
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,12288,0.015427199999491372
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,128,10240,0.03140906691551208
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,128,10240,0.013849600156148275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,10240,0.006920533378918965
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,10240,0.01381653348604838
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,128,8192,0.029820799827575684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,128,8192,0.011804800232251484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,8192,0.0067093332608540845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,8192,0.012077866991360982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,128,7168,0.029729066292444865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,128,7168,0.010822400450706482
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,7168,0.006640000144640605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,7168,0.011236266295115153
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,128,6144,0.028627200921376543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,128,6144,0.010082133611043294
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,6144,0.006504533191521962
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,6144,0.010406399766604107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,128,5120,0.027780266602834065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,128,5120,0.008962133526802063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,5120,0.00976639986038208
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,5120,0.006792533397674561
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,128,4096,0.026499199867248534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,128,4096,0.008080000181992848
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,4096,0.006413866579532623
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,4096,0.00862613320350647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,128,3584,0.026276266574859618
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,128,3584,0.007390933235486348
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,3584,0.006905599931875865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,3584,0.008344533046086629
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,128,3072,0.0253493328889211
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,128,3072,0.0071168000499407455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,3072,0.006690133114655812
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,512,512,0.004735999802748362
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,3072,0.007863466441631318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,128,2560,0.02448319991429647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,512,2048,0.006345599889755249
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,128,2560,0.006567466755708058
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,2560,0.005870933334032694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,2560,0.007537066439787547
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,128,2048,0.023668267329533896
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,128,2048,0.006131199995676676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,2048,0.005322666466236114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,2048,0.006995200117429097
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,128,1536,0.022856533527374268
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,1536,0.004682666560014089
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,1536,0.006716800232728322
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,128,1024,0.021944532791773476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,128,1024,0.005144533514976501
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,1024,0.003979733337958654
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,1024,0.006291200220584869
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,128,768,0.02295359969139099
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,128,768,0.004799999793370565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,768,0.0038399999340375268
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,768,0.0060576001803080235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,128,512,0.02113920052846273
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,128,512,0.004636799792448679
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,512,0.0035103999078273775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,512,0.005796266595522562
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,128,256,0.02125119964281718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,128,256,0.00443200021982193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,256,0.0031946666538715364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,256,0.005624533196290334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,128,128,0.020222934087117513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,128,128,0.004293333490689596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,128,0.0029546665648619336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,128,0.005629866818586985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,128,64,0.01882879932721456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,64,0.0028629332780838014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,64,0.005541333556175232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,128,32,0.018433066209157307
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,32,0.0028864001234372456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,32,0.005530666808287303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,65536,0.008360532919565837
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,64,65536,0.07885653177897135
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,65536,0.05948906739552816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,64,16384,0.037291733423868816
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,16384,0.006520533561706543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,16384,0.018705066045125326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,64,12288,0.03299093246459961
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,12288,0.0065098668138186145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,12288,0.01554026703039805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,64,10240,0.031285333633422854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,10240,0.006717866659164429
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,10240,0.013758933544158936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,64,8192,0.029356799523035687
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,8192,0.006887466708819072
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,8192,0.01200320025285085
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,64,7168,0.029736532767613726
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,7168,0.0064650664726893115
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,7168,0.011250133315722149
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,64,6144,0.02890346646308899
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,6144,0.006362666686375936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,6144,0.010410666465759277
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,64,5120,0.027515733242034913
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,5120,0.0067562664548556015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,256,8192,0.011830400427182515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,5120,0.009691733121871948
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,64,4096,0.02633066574732463
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,4096,0.006343466540177663
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,4096,0.008575999736785888
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,64,3584,0.025916800896326704
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,3584,0.006782933572928111
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,3584,0.00832426647345225
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,64,3072,0.025250132878621417
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,3072,0.006572799881299336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,3072,0.007793066898981731
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,64,2560,0.02434773246447245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,2560,0.005840000013510386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,2560,0.00747519979874293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,64,2048,0.023541333278020223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,2048,0.005205333232879639
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,2048,0.006994133194287618
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,64,1536,0.02339413364728292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,1536,0.00462719996770223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,1536,0.0066655998428662615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,64,1024,0.02201919953028361
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,1024,0.0039893334110577905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,1024,0.006355200211207073
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,64,768,0.022215465704600014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,768,0.0036938667297363283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,768,0.006001066664854685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,64,512,0.02122559944788615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,512,0.0033781332274278007
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,512,0.005817600091298421
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,64,256,0.021576533714930214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,256,0.003124266614516576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,256,0.0056415999929110205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,64,128,0.02031146685282389
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,128,0.0029887999097506206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,256,1024,0.0053151999910672505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,128,0.005553066730499268
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,64,64,0.018927999337514243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,64,0.0028309332827727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,64,0.005610666672388713
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,32,0.0028362666567166646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,64,32,0.01853333314259847
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,32,0.00551146666208903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,65536,0.008675199747085572
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,32,65536,0.07870720227559408
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,65536,0.05965119997660319
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,32,16384,0.036824532349904376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,16384,0.006723199784755707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,16384,0.01871466636657715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,32,12288,0.033130667606989544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,12288,0.00658240020275116
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,12288,0.015451733271280924
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,32,10240,0.03162239988644917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,10240,0.006621866424878438
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,10240,0.013909332950909934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,32,8192,0.02958400050799052
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,8192,0.0065311998128890995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,8192,0.011956266562143962
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,32,7168,0.029662932952245074
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,7168,0.00651093324025472
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,7168,0.01118933359781901
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,32,6144,0.02860586643218994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,32,4096,0.026412800947825117
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,6144,0.006376533210277558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,6144,0.010457600156466167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,32,5120,0.027434666951497395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,5120,0.00673280010620753
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,5120,0.009746133287747701
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,4096,0.00631039987007777
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,3072,0.006488533318042755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,4096,0.008543999989827473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,32,3584,0.025597866376241046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,3584,0.006821333368619282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,3584,0.008203733464082081
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,32,3072,0.025244800249735515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,3072,0.007846400141716003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,32,2560,0.024996266762415568
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,2560,0.005735466877619425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,2560,0.00749120016892751
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,32,2048,0.023449599742889404
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,2048,0.0051925331354141235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,2048,0.007032533486684163
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,32,1536,0.022703999280929567
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,1536,0.004620799918969473
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,1536,0.006614399949709575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,32,1024,0.02203413248062134
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,1024,0.003939199944337209
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,1024,0.006221866607666016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,32,768,0.02220373352368673
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,768,0.003621333340803782
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,768,0.006165333092212677
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,32,512,0.02122559944788615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,512,0.003335466732581457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,512,0.005776000022888183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,32,256,0.021313067277272543
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,256,0.0030720000465710956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,256,0.00561706672112147
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,64,0.005513600011666616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,32,128,0.020294400056203206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,128,0.0028597332537174227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,128,0.0055637334783871974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,32,64,0.018778665860493978
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,64,0.002796799937884013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,2,32,32,0.018569600582122803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,32,0.0027797333896160126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,32,0.00559146652619044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,128,1536,0.005619200070699056
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,65536,16384,0.39347413380940754
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,65536,16384,0.3615359942118327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,65536,12288,0.2703189214070638
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,65536,16384,0.42898985544840496
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,65536,12288,0.35146665573120117
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,65536,16384,0.8039093017578125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,65536,12288,0.5666133244832356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,65536,12288,0.2984522819519043
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,65536,10240,0.227347199122111
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,65536,10240,0.23179306983947753
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,65536,8192,0.18516480127970378
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,65536,10240,0.2740149180094401
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,65536,8192,0.1908394654591878
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,65536,8192,0.2478783925374349
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,65536,10240,0.5085824012756348
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,65536,8192,0.40881919860839844
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,65536,7168,0.16639466285705568
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,65536,7168,0.1659509340922038
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,65536,7168,0.2047093391418457
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,65536,7168,0.328817081451416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,65536,6144,0.14991787274678547
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,65536,6144,0.1453418731689453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,65536,6144,0.16314773559570311
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,65536,6144,0.28434346516927084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,65536,5120,0.13836906750996908
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,65536,5120,0.15087040265401203
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,65536,5120,0.13103679815928143
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,65536,5120,0.254531192779541
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,65536,4096,0.09602666695912679
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,65536,4096,0.12733973662058512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,65536,4096,0.19144320487976074
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,65536,4096,0.10410453478495281
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,65536,3584,0.0919871966044108
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,65536,3584,0.11452799638112385
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,65536,3584,0.0896405299504598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,65536,3584,0.1676106611887614
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,65536,3072,0.09728639920552572
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,65536,3072,0.07880213260650634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,65536,3072,0.14380052884419758
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,65536,2560,0.08831466833750407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,65536,2560,0.06211520036061605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,65536,2560,0.12027839819590251
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,65536,2560,0.07186133066813151
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,65536,2048,0.07752853234608968
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,65536,2048,0.09576640129089356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,65536,2048,0.05998293161392212
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,65536,1536,0.051140268643697105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,65536,1536,0.06427520116170247
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,65536,1536,0.07365226745605469
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,65536,1024,0.028735999266306562
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,65536,1024,0.05261226495107015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,65536,1024,0.0503541350364685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,65536,1024,0.039946667353312176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,65536,768,0.04923413197199504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,65536,768,0.023075199127197264
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,65536,768,0.03864746491114299
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,65536,768,0.03711893161137898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,65536,512,0.038072534402211505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,65536,512,0.013745066523551942
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,65536,512,0.02837760051091512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,65536,512,0.026104533672332765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,65536,256,0.03133119940757752
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,65536,256,0.009272533655166625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,65536,256,0.007646933197975159
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,65536,256,0.021656533082326256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,65536,128,0.028514132897059126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,65536,128,0.007490133245786031
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,65536,128,0.005497600138187409
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,65536,128,0.020156800746917725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,65536,64,0.028460800647735596
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,65536,64,0.006424533327420552
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,65536,64,0.020422399044036865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,65536,32,0.028458666801452637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,65536,32,0.006516266862551372
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,65536,32,0.020097066958745323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,16384,16384,0.14244267145792644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,16384,65536,0.4028575897216797
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,65536,0.3805930773417155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,16384,16384,0.09450559616088867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,16384,0.10875306924184162
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,16384,65536,0.4638997395833333
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,16384,0.20073919296264647
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,65536,0.8016063690185546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,16384,12288,0.0786730686823527
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,16384,12288,0.11004266738891602
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,12288,0.08034666379292807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,12288,0.15255893071492513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,16384,10240,0.0661354660987854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,16384,10240,0.09589333534240722
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,10240,0.06882133483886718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,10240,0.11588799953460693
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,16384,8192,0.05055786768595377
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,16384,8192,0.08237973054250082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,8192,0.05668799877166748
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,8192,0.10108693440755208
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,16384,7168,0.07260159651438394
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,7168,0.05080106655756632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,7168,0.08944106896718343
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,16384,6144,0.06586560010910034
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,16384,6144,0.03962026834487915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,6144,0.07235093116760254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,6144,0.04632426500320434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,16384,5120,0.058606934547424314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,16384,5120,0.03370240132013957
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,5120,0.06457279920578003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,5120,0.03874239921569824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,16384,4096,0.05247360070546468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,16384,4096,0.02972266674041748
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,4096,0.04882133404413859
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,4096,0.03322239915529887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,16384,3584,0.026318933566411334
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,16384,3584,0.05125866731007894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,3584,0.04323840141296387
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,3584,0.0300437331199646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,16384,3072,0.04794453382492066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,16384,3072,0.02370986739794413
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,3072,0.03772906859715779
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,3072,0.02837439974149068
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,16384,2560,0.04048320055007935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,2560,0.0230240007241567
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,2560,0.033986131350199386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,16384,2048,0.03139413396517436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,16384,2048,0.014473600188891092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,2048,0.027715200185775758
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,2048,0.016827734311421712
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,16384,1536,0.029521065950393676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,16384,1536,0.01171946624914805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,1536,0.02060479919115702
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,1536,0.014176000157992044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,16384,1024,0.02679786682128906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,16384,1024,0.009238400061925252
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,1024,0.008547199765841167
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,1024,0.012281599640846252
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,16384,768,0.026590933402379353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,16384,768,0.008164266745249431
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,65536,3072,0.08164160251617432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,768,0.00697813332080841
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,768,0.011322666207949321
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,16384,512,0.006579199930032094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,16384,512,0.025231999158859254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,512,0.010517332951227825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,512,0.005593599875768026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,65536,2048,0.058450134595235195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,16384,256,0.005960533519585928
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,16384,256,0.023112533489863078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,256,0.004217599829037985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,256,0.009593600034713745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,16384,128,0.021873066822687783
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,65536,1536,0.04539200067520142
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,64,0.0033088001112143196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,16384,128,0.005016533533732096
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,128,0.003642666588226954
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,128,0.009027199943860371
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,16384,64,0.021356799205144248
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,64,0.009018666545550029
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,16384,32,0.02146880030632019
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,32,0.0032874666154384612
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,32,0.008974933624267578
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,12288,16384,0.08597119649251303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,12288,65536,0.30499839782714844
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,65536,0.300599479675293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,12288,16384,0.10096960067749024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,16384,0.14317973454793292
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,16384,0.07650132973988852
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,12288,65536,0.3671882629394531
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,65536,0.5622517267862956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,12288,12288,0.09227093060811362
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,12288,0.059564801057179775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,12288,0.10612800121307372
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,12288,10240,0.05777813196182251
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,10240,0.09160640239715576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,12288,10240,0.07922773361206055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,10240,0.05065919955571493
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,12288,8192,0.045805867513020834
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,12288,8192,0.06393813292185466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,8192,0.07043946584065755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,8192,0.04899093310038249
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,12288,7168,0.0583349347114563
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,12288,7168,0.04599786599477132
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,7168,0.06198399861653646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,7168,0.03834666808446248
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,12288,6144,0.05445013443628947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,6144,0.05924266576766968
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,6144,0.03454506794611613
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,12288,5120,0.05039039850234985
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,12288,5120,0.03409813245137532
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,5120,0.04532373348871867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,5120,0.029908267656962077
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,12288,4096,0.044812798500061035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,12288,4096,0.02797973354657491
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,4096,0.03800426721572876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,4096,0.025110399723052977
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,12288,3584,0.04291839996973674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,12288,3584,0.025494400660196943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,3584,0.033633065223693845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,3584,0.02295359969139099
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,16384,7168,0.045960533618927005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,12288,3072,0.03256426652272542
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,12288,3072,0.02189226746559143
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,3072,0.02868266701698303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,3072,0.017093332608540852
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,12288,2560,0.02863360047340393
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,12288,2560,0.01709866722424825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,2560,0.025090134143829344
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,2560,0.014858667055765787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,12288,2048,0.015262933572133383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,2048,0.023437867561976113
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,12288,2048,0.02654613256454468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,2048,0.012874666849772134
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,12288,1536,0.026461867491404216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,12288,1536,0.012088533242543538
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,1536,0.01100373367468516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,1536,0.011380267143249512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,12288,1024,0.024145066738128662
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,12288,1024,0.009484799702962239
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,1024,0.008756267031033833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,1024,0.009866666793823243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,12288,768,0.02406826615333557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,12288,768,0.008218666911125183
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,768,0.0071733335653940845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,768,0.00915839970111847
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,12288,512,0.0229312002658844
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,12288,512,0.006306133170922597
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,512,0.005785599847634633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,16384,2560,0.019453867276509603
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,512,0.008449066678682964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,12288,256,0.0209279994169871
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,12288,256,0.005390933156013489
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,256,0.004277333120505015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,256,0.007796266674995422
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,12288,128,0.01948053240776062
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,12288,128,0.004935466746489207
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,128,0.003607466568549474
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,128,0.007293866574764251
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,12288,64,0.0190720001856486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,64,0.0033845332761605583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,64,0.00730560024579366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,12288,32,0.01911146640777588
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,32,0.00335359995563825
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,32,0.007294933497905731
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,10240,16384,0.06460479895273843
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,10240,65536,0.24311359723409018
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,10240,16384,0.08763840198516845
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,65536,0.2518869400024414
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,16384,0.06644906600316366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,16384,0.11666026910146077
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,10240,65536,0.2999957402547201
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,10240,12288,0.07238186995188395
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,65536,0.46561279296875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,12288,0.09324159622192382
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,12288,0.05412160158157349
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,10240,10240,0.04405119816462199
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,10240,10240,0.07145067056020102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,10240,0.0462005337079366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,10240,0.08722346623738607
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,10240,8192,0.05643946727116903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,8192,0.059988268216451014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,8192,0.03787413438161214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,10240,7168,0.052889601389567054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,10240,7168,0.034840532143910724
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,7168,0.053140266736348474
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,7168,0.03427733182907104
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,10240,6144,0.049033598105112715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,10240,6144,0.02988479932149251
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,6144,0.046529066562652585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,6144,0.03063039978345235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,10240,5120,0.044758399327596024
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,10240,5120,0.02416106661160787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,5120,0.03919466733932495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,5120,0.02698880036671956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,10240,4096,0.03919680118560791
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,12288,12288,0.06655040184656778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,10240,4096,0.019393066565195717
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,4096,0.03364053169886271
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,4096,0.023913600047429404
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,10240,3584,0.03290666739145915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,10240,3584,0.020425599813461304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,3584,0.028343466917673747
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,3584,0.017598932981491087
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,10240,3072,0.03004586696624756
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,10240,3072,0.017770665884017944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,3072,0.024836266040802003
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,3072,0.01569386621316274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,10240,2560,0.028137600421905516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,10240,2560,0.015041066209475198
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,2560,0.021155200401941934
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,2560,0.013912533720334372
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,12288,6144,0.03708693186442057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,10240,2048,0.02618559996287028
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,10240,1536,0.025517867008845015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,10240,2048,0.013489066561063131
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,2048,0.014452266693115234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,2048,0.012405332922935487
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,10240,1536,0.010964266459147136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,1536,0.008451199531555176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,1536,0.011065600315729777
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,10240,1024,0.023484800259272257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,10240,1024,0.008756267031033833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,1024,0.00666133314371109
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,1024,0.009589333335558574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,10240,768,0.02363413373629252
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,10240,768,0.007930666704972585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,768,0.0056202664971351625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,768,0.008930133779843648
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,10240,512,0.022564266125361124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,512,0.0047989333669344585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,512,0.008309333523114523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,10240,256,0.02100373307863871
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,10240,256,0.0054282665252685545
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,256,0.0038730666041374207
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,256,0.0075882668296496075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,10240,128,0.019836799303690592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,10240,128,0.004972800115744273
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,128,0.0034773332377274835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,128,0.0072202667593956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,10240,64,0.01947306593259176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,64,0.0031338666876157125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,64,0.007196799914042156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,10240,32,0.01907520095507304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,32,0.0031968000034491217
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,32,0.007267199953397115
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,8192,65536,0.19877759615580243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,65536,0.21117866834004723
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,8192,65536,0.25544212659200033
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,8192,16384,0.07529066403706869
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,65536,0.3775103886922201
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,8192,16384,0.05920106569925944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,16384,0.10960319836934407
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,16384,0.05962560176849365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,8192,12288,0.06717546780904135
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,8192,12288,0.0423253337542216
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,12288,0.07023466428120931
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,12288,0.0468778649965922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,8192,10240,0.0576416015625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,8192,10240,0.03585066795349121
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,10240,0.05899840195973714
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,10240,0.04063040018081665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,8192,8192,0.05076906681060791
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,8192,8192,0.02874666651089986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,8192,0.04783039887746175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,8192,0.03375146786371867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,8192,7168,0.04805759986241658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,8192,7168,0.03275306622187297
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,7168,0.04288533528645833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,7168,0.031177600224812824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,8192,6144,0.04511786699295044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,6144,0.03742400010426839
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,6144,0.027987200021743774
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,8192,5120,0.039989332358042404
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,5120,0.0365610678990682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,10240,12288,0.049214935302734374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,5120,0.021895466248194377
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,8192,4096,0.029795199632644653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,8192,4096,0.017802667617797852
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,4096,0.02987839976946513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,4096,0.015270400047302245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,8192,3584,0.02736426591873169
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,8192,3584,0.01853013237317403
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,3584,0.02285226583480835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,3584,0.01437333325544993
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,8192,3072,0.026250666379928587
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,8192,3072,0.01641386648019155
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,3072,0.020259199539820354
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,10240,8192,0.03403840065002441
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,2560,0.015320533514022827
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,3072,0.013194666306177775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,8192,2560,0.025898667176564534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,8192,2560,0.01432213286558787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,2560,0.01206719974676768
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,8192,2048,0.024375466505686443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,8192,2048,0.012563199798266093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,2048,0.00944640040397644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,2048,0.01087679962317149
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,8192,1536,0.02427306572596232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,1536,0.00817920019229253
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,1536,0.009969066580136616
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,8192,1024,0.02253226637840271
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,1024,0.008905599514643352
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,1024,0.006422399977842967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,8192,768,0.022975999116897582
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,8192,768,0.007579733431339264
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,768,0.005414400001366933
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,768,0.008389332890510559
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,8192,512,0.022151466210683188
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,8192,512,0.006028800209363302
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,512,0.004563199977080027
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,512,0.007948799928029378
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,8192,256,0.02093440095583598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,8192,256,0.0050911997755368555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,256,0.007657599945863088
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,256,0.003601066768169403
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,8192,128,0.01926400065422058
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,8192,128,0.00489279975493749
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,128,0.0033344000577926634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,128,0.007039999961853028
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,8192,64,0.01949119965235392
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,64,0.0070613334576288865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,64,0.0031306666632493338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,8192,32,0.019351466496785482
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,32,0.0030720000465710956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,32,0.007049599786599477
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,7168,65536,0.179640531539917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,7168,65536,0.22062719662984215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,65536,0.17273599306742352
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,65536,0.32276268005371095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,7168,16384,0.07110933462778726
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,7168,16384,0.05103360017140707
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,16384,0.049668268362681074
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,16384,0.08773866494496664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,7168,12288,0.05960533221562704
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,7168,12288,0.03853120009104411
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,10240,512,0.006341333190600078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,12288,0.06485333442687988
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,12288,0.03698346614837646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,7168,10240,0.03413333495457967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,7168,10240,0.05253119866053263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,10240,0.056849066416422525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,10240,0.03370026747385661
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,7168,8192,0.04731520016988118
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,7168,8192,0.026361600557963057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,8192,0.04501653512318929
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,8192,0.02690666715304057
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,7168,7168,0.04551253318786621
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,7168,7168,0.03046506643295288
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,7168,0.04031466643015544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,7168,0.024221867322921753
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,7168,6144,0.04150293270746867
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,7168,6144,0.02254826625188192
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,6144,0.022037333250045775
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,6144,0.03631893396377563
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,7168,5120,0.020992000897725425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,7168,5120,0.03167253335316976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,5120,0.03158506751060486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,5120,0.01390506625175476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,7168,4096,0.027873067061106364
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,7168,4096,0.014678399761517844
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,4096,0.023212800423304238
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,4096,0.011346133550008138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,7168,3584,0.028450133403142293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,7168,3584,0.01731520096460978
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,3584,0.020875734090805054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,3584,0.011060266693433126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,7168,3072,0.026055467128753663
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,7168,3072,0.015691733360290526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,3072,0.01819733381271362
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,3072,0.010245333115259807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,7168,2560,0.02584213415781657
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,7168,2560,0.013474133610725404
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,2560,0.010337066650390626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,2560,0.009399466713269551
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,7168,2048,0.024173865715662636
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,7168,2048,0.011601066589355469
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,2048,0.009075199564297993
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,8192,6144,0.02846933404604594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,2048,0.008538666367530822
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,7168,1536,0.023918932676315306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,7168,1536,0.00960213343302409
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,1536,0.007481599847475688
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,1536,0.007503999769687653
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,8192,5120,0.02309760053952535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,7168,1024,0.02281493345896403
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,7168,1024,0.007747200131416321
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,1024,0.005994666616121928
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,1024,0.00678719977537791
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,7168,768,0.022808533906936646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,7168,768,0.007145600020885467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,512,0.0043818667531013485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,768,0.005246933301289876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,768,0.0063360000650088
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,7168,512,0.021948800484339396
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,512,0.006074666480223338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,7168,256,0.02031466762224833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,7168,256,0.005013333261013031
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,8192,1536,0.010237866640090942
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,256,0.0035690667728583017
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,256,0.005705599983533224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,7168,128,0.019169066349665323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,128,0.00327360009153684
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,128,0.0053951998551686605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,7168,64,0.018835200866063436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,64,0.003035733352104823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,64,0.005400533477465311
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,7168,32,0.019042134284973145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,8192,1024,0.008307200173536937
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,32,0.0030048000315825146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,32,0.005399466554323832
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,6144,65536,0.1787157376607259
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,65536,0.15705067316691082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,6144,65536,0.18506773312886554
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,6144,16384,0.047275733947753903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,6144,16384,0.06349226633707682
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,16384,0.07623999913533529
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,65536,0.2843274752298991
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,16384,0.04673173427581787
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,6144,12288,0.03643413384755452
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,6144,12288,0.05555839935938517
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,12288,0.06113066673278809
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,12288,0.03322346607844035
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,6144,10240,0.05189013481140137
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,10240,0.0457856019337972
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,10240,0.03284693360328674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,6144,8192,0.045360000928243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,8192,0.040088534355163574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,8192,0.02467733422915141
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,6144,7168,0.04264853398005168
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,6144,7168,0.03060693343480428
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,7168,0.03288000027338664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,7168,0.022005333503087362
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,6144,6144,0.03356906572977702
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,6144,6144,0.025063467025756837
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,6144,0.031083732843399048
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,6144,0.015655466914176942
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,6144,5120,0.02962133288383484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,5120,0.025732266902923583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,5120,0.012999467055002847
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,6144,4096,0.028781867027282713
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,6144,4096,0.01687999963760376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,4096,0.021287467082341513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,4096,0.01130560040473938
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,6144,3584,0.02839573423067729
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,6144,3584,0.017998933792114258
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,3584,0.017298134167989095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,3584,0.010563199718793232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,6144,3072,0.02712213397026062
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,6144,3072,0.015997866789499916
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,3072,0.011637333035469054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,3072,0.010125866532325745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,6144,2560,0.0269269327322642
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,6144,2560,0.013970133662223817
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,2560,0.010104533036549885
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,2560,0.009070932865142822
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,6144,2048,0.025277866919835405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,6144,2048,0.012414933244387309
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,2048,0.008736000458399455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,2048,0.008308266599973042
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,6144,1536,0.02469759980837504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,6144,1536,0.010257066289583842
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,1536,0.007197866837183635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,1536,0.007527466615041096
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,6144,1024,0.02283733288447062
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,6144,1024,0.008226133386294047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,1024,0.00684799998998642
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,1024,0.0057429333527882894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,6144,768,0.022760534286499025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,6144,768,0.007422933479150136
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,768,0.004960000018278758
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,768,0.006374399860699971
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,6144,512,0.022191999355951945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,6144,512,0.005798399945100148
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,512,0.00421973317861557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,512,0.006062933305899302
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,6144,256,0.02031146685282389
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,6144,256,0.0050784001747767125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,256,0.003500800083080927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,256,0.005604266623655955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,6144,128,0.01908586621284485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,6144,128,0.004791466891765595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,128,0.005446400245030721
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,128,0.0032245332996050516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,6144,64,0.019042134284973145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,64,0.0030080000559488933
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,64,0.005388799806435903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,6144,32,0.01890986760457357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,32,0.005365333457787832
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,32,0.0029845332105954488
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,5120,65536,0.15695573488871256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,5120,65536,0.17055999437967936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,65536,0.23669333457946778
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,65536,0.13343040148417157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,5120,16384,0.05936426719029745
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,5120,16384,0.04648319880167644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,16384,0.06032853523890177
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,16384,0.03651306629180908
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,5120,12288,0.05099946657816569
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,5120,12288,0.03506453434626262
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,12288,0.04751360019048055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,12288,0.029255467653274535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,5120,10240,0.046033068497975664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,5120,10240,0.028707200288772584
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,10240,0.026562132438023883
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,10240,0.03956053256988525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,7168,512,0.005851733187834421
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,5120,8192,0.02507413427035014
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,5120,8192,0.03846720059712728
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,8192,0.03191146651903788
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,8192,0.018823466698328652
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,5120,7168,0.033321599165598556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,5120,7168,0.021651200453440347
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,7168,128,0.004681600133577982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,7168,0.02841813365618388
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,7168,0.015141333142916361
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,5120,6144,0.030728532870610552
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,5120,6144,0.01813439925511678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,6144,0.025176533063252765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,6144,0.013287466764450074
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,5120,5120,0.029332266251246138
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,5120,5120,0.0196288009484609
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,5120,0.021851734320322672
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,5120,0.012331733107566833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,5120,4096,0.027314132452011107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,5120,4096,0.014028799533843995
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,4096,0.01690453290939331
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,4096,0.010673066973686219
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,5120,3584,0.027135999997456868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,5120,3584,0.016237866878509522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,3584,0.010621866583824158
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,3584,0.010059733192125957
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,5120,3072,0.02578666607538859
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,3072,0.009522133072217305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,3072,0.009401599566141765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,5120,2560,0.025726934274037678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,5120,2560,0.012838400403658547
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,2560,0.008732799688975017
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,6144,10240,0.031642667452494305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,5120,2048,0.02402346730232239
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,2560,0.008784000078837078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,5120,2048,0.01125973363717397
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,6144,8192,0.025890133778254193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,5120,1536,0.02432960073153178
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,2048,0.007427200178305308
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,2048,0.007957333326339721
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,1536,0.0063509335120519
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,1536,0.007272533575693766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,5120,1024,0.02223466634750366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,5120,1024,0.00763626645008723
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,1024,0.0052255998055140175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,1024,0.006563200056552887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,768,0.004783999919891357
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,768,0.006198399762312571
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,6144,5120,0.020499199628829956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,5120,768,0.022533333301544188
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,5120,768,0.00699946681658427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,5120,512,0.021974400679270426
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,5120,512,0.0055978665749231975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,512,0.004013866682847341
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,512,0.005922133227189382
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,5120,256,0.020292266209920248
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,256,0.003436800092458725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,5120,128,0.004648533463478088
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,256,0.005565866827964783
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,5120,128,0.019103999932607016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,128,0.003202133377393087
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,128,0.0053941334287325535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,5120,64,0.018820265928904213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,64,0.002958933264017105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,64,0.005400533477465311
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,5120,32,0.018895999590555827
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,32,0.0029706666866938275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,32,0.005336533486843109
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,4096,65536,0.13368852933247882
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,4096,65536,0.15326399803161622
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,65536,0.10726186434427898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,4096,16384,0.052321068445841466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,65536,0.1990143934885661
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,4096,16384,0.042855465412139894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,16384,0.050602666536966955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,16384,0.03189439972241719
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,4096,12288,0.04640533526738484
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,4096,12288,0.0324618657430013
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,12288,0.03963413238525391
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,12288,0.02574613293011983
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,4096,10240,0.040550398826599124
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,4096,10240,0.02760639985402425
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,10240,0.03421013355255127
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,4096,8192,0.021896533171335855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,10240,0.020440532763799032
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,4096,8192,0.03232426643371582
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,8192,0.028225066264470418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,8192,0.012313600381215413
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,4096,7168,0.028918399413426714
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,7168,0.02566293279329936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,7168,0.011473066608111064
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,4096,6144,0.02805333336194356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,4096,6144,0.019103999932607016
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,6144,0.022758400440216063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,6144,0.010679466525713603
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,4096,5120,0.027060266335805255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,4096,5120,0.01791680057843526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,5120,0.017847466468811034
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,4096,4096,0.025620265801747637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,5120,0.009788800279299419
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,4096,4096,0.013639466961224875
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,4096,0.011426132917404175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,4096,0.008762666583061218
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,4096,3584,0.02529279987017314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,4096,3584,0.01736533244450887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,3584,0.008717866738637288
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,3584,0.008327466746171314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,4096,3072,0.02432639996210734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,4096,3072,0.015440000096956888
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,3072,0.008099199831485748
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,3072,0.007879466811815897
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,4096,2560,0.02443199952443441
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,4096,2560,0.013176533579826354
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,2560,0.007351466516653697
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,2560,0.007427200178305308
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,4096,2048,0.02307200034459432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,4096,2048,0.011748266220092774
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,2048,0.006516266862551372
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,2048,0.006911999980608622
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,4096,1536,0.02307093342145284
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,4096,1536,0.009832533200581868
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,1536,0.005510400235652924
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,1536,0.006486399968465169
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,4096,1024,0.02169493238131205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,4096,1024,0.007979733248551685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,1024,0.004583466549714406
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,1024,0.0060576001803080235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,4096,768,0.022273067633310953
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,768,0.004147200038035711
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,768,0.005804799993832906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,4096,512,0.021614933013916017
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,512,0.003756800045569738
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,4096,512,0.005628799895445505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,512,0.005618133147557576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,4096,256,0.020148267348607383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,4096,256,0.004969599843025208
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,256,0.0033610666791598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,256,0.005385600030422211
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,4096,128,0.018982400496800743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,4096,128,0.004644266764322917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,5120,3072,0.015218133727709452
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,128,0.0031146667897701263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,128,0.005275733272234599
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,4096,64,0.018680532773335777
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,64,0.0029567999144395193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,64,0.005253333350022634
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,32,0.0029333333174387617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,4096,32,0.018629332383473717
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,32,0.005250133574008942
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,5120,1536,0.009317333499590557
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3584,65536,0.12751893202463788
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,65536,0.16341013908386232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3584,16384,0.040990932782491045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3584,16384,0.04986026684443156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,65536,0.09729066689809164
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,16384,0.028861866394678755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3584,12288,0.04459199905395508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3584,12288,0.030930133660634358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,16384,0.04501333236694336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,12288,0.03623786767323812
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,12288,0.023669334252675374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3584,10240,0.03425066471099854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3584,10240,0.025158399343490602
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,10240,0.031013333797454835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,10240,0.014517333110173544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3584,8192,0.029666133721669513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,5120,256,0.005006933212280273
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3584,8192,0.02106026609738668
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,8192,0.025832533836364746
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,8192,0.012312533458073933
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3584,7168,0.028995199998219805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3584,7168,0.024472532669703166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,7168,0.02300373315811157
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,7168,0.011507200201352437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3584,6144,0.02797866662343343
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3584,6144,0.021450666586558025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,6144,0.020111999909083047
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,6144,0.01067200005054474
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3584,5120,0.026970666646957398
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3584,5120,0.018347734212875368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,5120,0.011260799566904704
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,5120,0.009718400239944459
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3584,4096,0.025099732478459674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3584,4096,0.01286079982916514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,4096,0.009084799885749817
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,4096,0.00860053300857544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3584,3584,0.025265065828959148
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3584,3584,0.01585493286450704
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,3584,0.008690133690834045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,3584,0.008331733445326488
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3584,3072,0.024219733476638795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3584,3072,0.014569600423177084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,3072,0.00783679982026418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,3072,0.007860266665617625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3584,2560,0.02426346739133199
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,2560,0.007124266525109608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3584,2048,0.010924800237019857
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,2048,0.006001066664854685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,2560,0.007462400197982788
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3584,2048,0.023702400922775268
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,4096,7168,0.022394667069117226
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,2048,0.006897066533565521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3584,1536,0.02304746707280477
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3584,1536,0.0090421328941981
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,1536,0.005545599758625031
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,1536,0.006533333162466686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3584,1024,0.02175253431002299
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,1024,0.004514133433500925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,1024,0.006044800082842508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3584,768,0.022168533007303873
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3584,768,0.006832000116507213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,768,0.00411520004272461
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,768,0.005883733431498209
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3584,512,0.021602133909861244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3584,512,0.005520000060399374
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,512,0.0037418665985266366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,512,0.005658666789531708
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3584,256,0.020105600357055664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3584,256,0.004871466755867004
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,256,0.0033429334561030067
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,256,0.005458133419354757
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3584,128,0.018974934021631876
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3584,128,0.004537599782148996
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,128,0.0030826665461063385
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,128,0.005291733145713806
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3584,64,0.01872426668802897
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,64,0.002914133419593175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,64,0.00525546669960022
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3584,32,0.018590933084487914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,32,0.0029333333174387617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,32,0.005260799825191498
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3072,65536,0.13053226470947266
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3072,65536,0.11904213428497315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,4096,768,0.007208533088366191
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,65536,0.14569600423177081
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,65536,0.08839680353800455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3072,16384,0.03847039937973022
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3072,16384,0.04748693307240804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,16384,0.026375466585159303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,16384,0.04021226565043132
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3072,12288,0.035488001505533856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3072,12288,0.0279914657274882
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,12288,0.031402667363484696
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,12288,0.015684266885121666
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3072,10240,0.031027199824651082
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3072,10240,0.02330026626586914
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,10240,0.027219200134277345
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,10240,0.013332266608874002
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3072,8192,0.029322665929794312
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3072,8192,0.020057600736618043
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,8192,0.022574933369954427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,8192,0.011642666657765706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3072,7168,0.028487465778986615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3072,7168,0.019166932503382365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,7168,0.018810667594273887
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,7168,0.011013333002726238
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3072,6144,0.02744106650352478
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3072,6144,0.01702079971631368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,6144,0.012174933155377706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,6144,0.010206933816274006
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3072,5120,0.026073600848515826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3072,5120,0.01641706625620524
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,5120,0.01114453375339508
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,4096,0.010499200224876404
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,5120,0.009734400113423665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3072,4096,0.025144533316294355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3072,4096,0.014125866691271463
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3584,65536,0.14520640373229982
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,3584,0.009831466277440389
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,4096,0.008517332871754964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3072,3584,0.025201066335042314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3072,3584,0.014282666643460593
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,3584,0.008171733220418293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3072,3072,0.012923733393351237
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3072,3072,0.024119466543197632
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,3072,0.00916373332341512
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,3072,0.007788800199826558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3072,2560,0.02494293252627055
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3072,2560,0.011637333035469054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,2560,0.006925866504510244
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,2560,0.007481599847475688
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3072,2048,0.022984532515207927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3072,2048,0.010308266679445902
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,2048,0.005949866771697998
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,2048,0.006851199766000111
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3072,1536,0.02309439977010091
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3072,1536,0.008577066659927367
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,1536,0.005179733534653981
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,1024,0.004491733511288961
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,1536,0.006455466647942861
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3072,1024,0.02195840080579122
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3072,1024,0.0071263998746871945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,1024,0.006056533257166544
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3072,768,0.022107734282811483
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3072,768,0.006584533552328746
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,768,0.00409706657131513
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,768,0.005871999760468801
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3072,512,0.0215445339679718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,512,0.003738666574160258
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,512,0.0056757330894470215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3072,256,0.020028799772262573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3072,256,0.004829866687456766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3072,128,0.004496000210444133
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,256,0.0033461332321166994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,256,0.00544213354587555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3072,128,0.018984532356262206
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,128,0.003083733220895131
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,128,0.005241600175698599
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3072,64,0.018669867515563966
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,64,0.0028927999238173166
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3584,2560,0.01220373312632243
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,64,0.005242666602134705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,3072,32,0.018832000096638997
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,32,0.0029279999434947968
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,32,0.005267199873924255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2560,65536,0.10491519769032795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2560,65536,0.11862506866455078
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,65536,0.12065920035044353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,65536,0.0806602636973063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2560,16384,0.04330026706059774
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2560,16384,0.0375765323638916
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,16384,0.03576106627782186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,16384,0.022196267048517862
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3584,1024,0.007493333518505096
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2560,12288,0.03300266663233439
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2560,12288,0.027261867125829058
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,12288,0.027437865734100342
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,12288,0.014745600024859109
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2560,10240,0.031185066699981688
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2560,10240,0.023375999927520753
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,10240,0.02370026707649231
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,10240,0.013201066851615905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2560,8192,0.029150933027267456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2560,8192,0.020915200312932335
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,8192,0.017709867159525553
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,8192,0.01156160036722819
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2560,7168,0.02855679988861084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2560,7168,0.022191999355951945
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,7168,0.012353066603342693
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,7168,0.010930132865905762
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2560,6144,0.027296000719070436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2560,6144,0.019393066565195717
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,6144,0.011847466230392456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,6144,0.01009279986222585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2560,5120,0.026065067450205488
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2560,5120,0.016169599692026772
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,5120,0.01051093339920044
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,5120,0.009452799956003826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2560,4096,0.02526613275210063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2560,4096,0.01390506625175476
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,4096,0.00925546685854594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,4096,0.008524800340334576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2560,3584,0.025356799364089966
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,3584,0.008731733759244282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2560,3584,0.013035733501116434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,3584,0.008171733220418293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2560,3072,0.024242132902145386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2560,3072,0.012330666184425354
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,3072,0.008249600231647492
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,3072,0.007825066645940144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2560,2560,0.024359466632207234
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2560,2560,0.010407466689745586
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,2560,0.006387199958165486
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,2560,0.007333333293596904
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2560,2048,0.02304853399594625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,2048,0.0056991999348004665
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,2048,0.006852266689141591
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2560,1536,0.023548799753189086
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2560,1536,0.00812906672557195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,1536,0.004994133114814758
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,1536,0.006484266618887584
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2560,1024,0.02172586719195048
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2560,1024,0.006871466835339864
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,1024,0.0043381333351135256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,1024,0.006004266440868378
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2560,768,0.022257065773010253
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2560,768,0.0063701331615448
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,768,0.00403413325548172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,768,0.006040533383687338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2560,512,0.021755733092625937
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,512,0.0036288000643253326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,512,0.005593599875768026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2560,256,0.02016213337580363
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2560,256,0.004749866823355356
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,256,0.0032618666688601174
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,128,0.0030741333961486817
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,256,0.005364266534646353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2560,128,0.018969599405924478
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,128,0.0052928000688552855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2560,64,0.01867626706759135
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,64,0.002902399996916453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,64,0.0052704001466433205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2560,32,0.0188426673412323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,32,0.0029653333127498626
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,32,0.005184000233809153
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2048,65536,0.09634772936503092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2048,65536,0.13495893478393556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,65536,0.096124800046285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,65536,0.07300373713175455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2048,16384,0.038778666655222574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2048,16384,0.033187200625737504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,16384,0.028202666838963823
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,16384,0.018143999576568603
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2048,12288,0.03282026648521423
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2048,12288,0.027694932619730633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,12288,0.02244480053583781
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,12288,0.014601600170135499
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2048,10240,0.031049599250157673
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2048,10240,0.023964800437291465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,10240,0.016454399625460307
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,10240,0.013247999548912048
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2048,8192,0.029183999697367353
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3072,512,0.005368533233801523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2048,8192,0.019630932807922365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,8192,0.010984533031781514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,8192,0.011508267124493916
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2048,7168,0.028417066733042402
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2048,6144,0.027404799064000444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,7168,0.010479999581972758
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,7168,0.010946133732795715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2048,6144,0.021337600549062093
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,6144,0.009499733646710713
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,6144,0.01014293332894643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2048,5120,0.02603306571642558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,5120,0.009070932865142822
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,5120,0.009426132837931315
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2048,4096,0.02518933415412903
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2048,4096,0.01439573367436727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,4096,0.00802346666653951
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,4096,0.008480000495910644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2048,3584,0.025302400191624958
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2048,3584,0.016668800512949625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,3584,0.007762133578459422
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,3584,0.008161066472530365
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2048,3072,0.02430079976717631
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,3072,0.007326933244864146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,3072,0.007821866869926452
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2048,2560,0.02445546587308248
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2048,2560,0.012737066547075907
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,2560,0.006284800171852112
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,2560,0.007383466760317485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2048,2048,0.022954666614532472
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2048,2048,0.011495467027028401
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,2048,0.005589333176612854
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,2048,0.0068341334660847975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2048,1536,0.022990934054056802
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,1536,0.0049002667268117275
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,1536,0.006377600133419037
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2048,1024,0.02196906606356303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2048,1024,0.007732266684373219
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,1024,0.004233600199222564
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,1024,0.006039466460545858
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2048,768,0.02234026590983073
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2048,768,0.007079466680685679
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,768,0.003909333298603693
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,768,0.005910400052865346
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2048,512,0.021705599625905354
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2048,512,0.005457066496213277
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,512,0.003572266548871994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,512,0.005598933498064677
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2048,256,0.020127999782562255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2048,256,0.004870399832725525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,256,0.0032266666491826378
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,256,0.005382399757703145
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2048,128,0.019203199942906698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2048,128,0.0046186665693918865
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,128,0.003010133405526479
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2560,2048,0.009671466549237569
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,128,0.005284266670544943
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2048,64,0.01861013372739156
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,64,0.0028864001234372456
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,64,0.005224533379077911
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,2048,32,0.01882773240407308
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,32,0.0028736000259717304
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,32,0.0052373334765434265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1536,65536,0.08963733514149984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1536,65536,0.11515946388244629
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,65536,0.0742677370707194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,65536,0.06601386864980062
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1536,16384,0.036135466893514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1536,16384,0.031754666566848756
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,16384,0.023102933168411256
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,16384,0.017785600821177163
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1536,12288,0.03282133340835571
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1536,12288,0.024576000372568765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2560,512,0.005278933544953664
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,12288,0.013685333728790283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,12288,0.014674133062362671
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1536,10240,0.03109440008799235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1536,10240,0.02114026745160421
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,10240,0.012063999970753986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,10240,0.013333333532015481
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1536,8192,0.02921066681543986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1536,8192,0.018049067258834837
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,8192,0.010453333457310993
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2560,128,0.004419200122356415
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,8192,0.011559466520945232
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1536,7168,0.028436267375946046
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1536,7168,0.018178133169809978
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,7168,0.009715200463930766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,7168,0.010853333274523418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1536,6144,0.02752000093460083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,6144,0.009105066458384197
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,6144,0.010154666503270467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1536,5120,0.02661760052045186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1536,5120,0.014749866724014283
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,5120,0.008406399687131246
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,5120,0.00941439966360728
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1536,4096,0.025153066714604693
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1536,4096,0.012869333227475485
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,4096,0.007752533257007599
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,4096,0.008504533767700195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1536,3584,0.025084799528121947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1536,3584,0.013642666737238565
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,3584,0.007429333527882893
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,3584,0.008136533200740814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1536,3072,0.024526933828989662
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1536,3072,0.012542933225631714
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,3072,0.007076266904671986
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,3072,0.00774186650911967
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1536,2560,0.024246400594711302
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1536,2560,0.010546132922172546
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,2560,0.006230400005976359
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,2560,0.007267199953397115
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1536,2048,0.023088000218073525
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2048,7168,0.02526400089263916
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1536,2048,0.009701333443323771
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,2048,0.005560533205668131
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,2048,0.00688213308652242
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1536,1536,0.023515733083089192
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,1536,0.004905599852403005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,1536,0.00639466643333435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2048,5120,0.016445866227149962
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1536,1024,0.021639466285705566
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1536,1024,0.0070250665148099255
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,1024,0.0042367999752362575
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,1024,0.006071466704209646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1536,768,0.02232960065205892
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1536,768,0.0064298664530118305
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,768,0.0038719999293486277
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,768,0.005773866673310598
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1536,512,0.021538132429122926
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2048,3072,0.015034666657447815
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,512,0.003568000098069509
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,512,0.005601066847642263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1536,256,0.02005013426144918
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1536,256,0.004756266872088114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,256,0.0032458665470282235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,256,0.005393066505591074
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1536,128,0.019172267119089762
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1536,128,0.004418133199214936
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,128,0.00525439977645874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,128,0.003028266628583272
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1536,64,0.0186463991800944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,64,0.002880000074704488
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,64,0.005218133330345154
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1536,32,0.018602667252222697
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,32,0.0029205332199732465
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,32,0.005196799834569296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1024,65536,0.09558080037434896
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1024,65536,0.08647893269856771
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,65536,0.05280213356018067
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,65536,0.06210240125656128
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2048,1536,0.009452799956003826
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1024,16384,0.036098134517669675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1024,16384,0.031169066826502483
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,16384,0.01230400005976359
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,16384,0.017633066574732462
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1024,12288,0.032679466406504314
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,12288,0.01027733286221822
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,12288,0.014684800306955972
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1024,10240,0.030832000573476154
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1024,10240,0.019953066110610963
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,10240,0.009571199615796406
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,10240,0.01323946714401245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1024,8192,0.029106134176254274
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,8192,0.011548800269762675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1024,8192,0.018345600366592406
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,8192,0.008343467116355896
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1024,7168,0.027944533030192058
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1024,7168,0.015643733739852905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,7168,0.008141866823037466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,7168,0.010853333274523418
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1024,6144,0.027101866404215497
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1024,6144,0.014274133245150247
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,6144,0.007209600011507671
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,6144,0.010102400183677673
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1024,5120,0.02616320053736369
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1024,5120,0.013385599851608277
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,5120,0.0073088000218073535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,5120,0.009434666236241658
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1024,4096,0.025172267357508344
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1024,4096,0.010170666376749675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,4096,0.006883200009663899
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,4096,0.008451199531555176
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1024,3584,0.02515946626663208
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1024,3584,0.011286399761835734
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,3584,0.0073088000218073535
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,3584,0.008090666433175405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1024,3072,0.024178133408228556
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1024,3072,0.010481066505114238
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,3072,0.006950399776299794
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,3072,0.00763626645008723
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1024,2560,0.02413439949353536
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1024,2560,0.00906773308912913
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,2560,0.006367999811967213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,2560,0.007253333429495494
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1024,2048,0.022833067178726196
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1024,2048,0.008197333415349324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,2048,0.0056757330894470215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,2048,0.0067114666104316715
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1024,1536,0.02293973366419474
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1024,1536,0.007073066631952922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,1536,0.004971733192602793
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,1536,0.0063360000650088
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1024,1024,0.021660800774892172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1024,1024,0.006085333228111267
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,1024,0.004316799839337667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,1024,0.005930666625499725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1024,768,0.022114133834838866
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1024,768,0.005905066430568695
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,768,0.003991466760635376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,768,0.005686399837334951
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1024,512,0.021692800521850585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1024,512,0.005115733544031779
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,512,0.003553066651026408
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,512,0.005606399973233541
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1024,256,0.02016213337580363
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1024,256,0.004603733122348785
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,256,0.005368533233801523
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,256,0.003205333401759466
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1024,128,0.018900267283121743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1536,6144,0.01824000080426534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1024,128,0.00446720023949941
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,128,0.003020799905061722
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,128,0.005222400029500326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1024,64,0.018655999501546224
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,64,0.003278933217128118
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,64,0.0052597333987553915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,1024,32,0.018579200903574625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,32,0.0031829332311948144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,32,0.005336533486843109
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,768,65536,0.0843178669611613
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,768,65536,0.08662400245666504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,65536,0.039901868502298994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,65536,0.06136320034662882
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,768,16384,0.036356266339619955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,768,16384,0.02464639941851298
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,16384,0.010499200224876404
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,16384,0.017870932817459106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,768,12288,0.03290026585261027
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,12288,0.0092330664396286
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,12288,0.0147189329067866
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,768,10240,0.031088000535964964
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,10240,0.013292800386746725
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,768,10240,0.015424000223477683
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,10240,0.008614400029182434
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,768,8192,0.02932693362236023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,768,8192,0.01536853313446045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1536,1536,0.008148266871770223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,8192,0.007777066528797149
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,8192,0.011485866705576579
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,768,7168,0.028387200832366944
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,768,7168,0.016753067572911583
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,7168,0.007101866602897644
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,7168,0.010889599720637005
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,768,6144,0.027061333258946735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,768,6144,0.013154133160909017
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,6144,0.006929066777229309
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,6144,0.010133333007494609
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,768,5120,0.026601600646972656
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1536,512,0.005243733525276184
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,768,5120,0.012723199526468911
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,4096,0.006868266562620799
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,5120,0.0072970668474833175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,5120,0.009439999858538311
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,768,4096,0.025098667542139692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,768,4096,0.00858133335908254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,3584,0.008208000163237254
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,4096,0.008474666873613994
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,768,3072,0.02446613311767578
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,768,3584,0.025128533442815144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,3072,0.006852266689141591
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,768,3584,0.010442666212717692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,3584,0.007147733370463054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,768,3072,0.009355733791987102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,3072,0.007826133569081625
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,768,2560,0.02422399918238322
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,768,2560,0.008292266726493835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,2560,0.006218666831652323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,2560,0.007303466896216075
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,768,2048,0.023115734259287514
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,768,2048,0.007463466624418895
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,2048,0.005573333303133646
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,2048,0.00676800012588501
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,768,1536,0.022832000255584718
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,1536,0.004891733328501383
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,1536,0.006326400240262349
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,768,1024,0.02167146603266398
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,768,1024,0.005841066439946493
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1024,12288,0.023639466365178427
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,1024,0.004197333256403605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,1024,0.005990399916966756
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,768,768,0.02229866584142049
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,768,0.0038431999584039056
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,768,0.005787733197212219
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,768,512,0.02155839999516805
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,768,512,0.004886400202910105
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,768,256,0.004553600152333578
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,512,0.003542399903138479
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,512,0.005530666808287303
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,768,256,0.020026665925979615
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,256,0.00340693344672521
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,256,0.005319466690222422
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,768,128,0.01912533243497213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,768,128,0.004316799839337667
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,128,0.0031221332649389905
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,64,0.005209599932034811
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,128,0.005228800078233084
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,768,64,0.018685867389043175
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,64,0.002951466788848241
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,768,32,0.018553600708643595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,32,0.0029567999144395193
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,512,65536,0.08110933303833008
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,32,0.005223466455936432
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,512,65536,0.07443093458811442
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,65536,0.02889813383420308
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,65536,0.05828266541163126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,512,16384,0.03626453479131063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,512,16384,0.019167999426523842
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,16384,0.008523733417193095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,16384,0.017836799224217735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,512,12288,0.03267306685447693
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,512,12288,0.01646613379319509
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,12288,0.007607466479142506
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,12288,0.01467626690864563
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,512,10240,0.030943999687830608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,512,10240,0.013790933291117349
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,10240,0.007258666555086772
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,10240,0.013250133395195008
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,512,8192,0.02912213404973348
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,512,8192,0.011852799852689107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,8192,0.007155199845631917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,8192,0.011532800396283467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,512,7168,0.027825067440668743
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,7168,0.010865066448847454
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,512,7168,0.013894400000572205
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,7168,0.006875733534495036
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,512,6144,0.027292799949645997
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,512,6144,0.010014933347702027
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,6144,0.006760533154010773
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,6144,0.010058666268984478
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,512,5120,0.026409600178400678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,512,5120,0.009316266576449076
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,5120,0.007203199962774913
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,5120,0.009412266810735067
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,512,4096,0.025015467405319215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,4096,0.008457600076993307
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,512,4096,0.008134399851163227
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,4096,0.006798933446407318
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,512,3584,0.024949334065119424
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,3584,0.0072629332542419435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,512,3584,0.008609066406885784
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,3584,0.008002133170763651
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,3072,0.007618133227030437
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,512,3072,0.024000000953674317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,512,2560,0.024807467063268026
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,512,3072,0.008463999629020691
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,3072,0.006910933554172516
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,512,2560,0.007434666653474172
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,2560,0.006347733239332835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,2560,0.007245866457621257
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,512,2048,0.022762666145960488
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,512,2048,0.00697386662165324
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,512,1536,0.00619946668545405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,2048,0.0056096002459526065
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,1536,0.0063391998410224915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,2048,0.006620799998442332
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,512,1024,0.005560533205668131
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,512,1536,0.02340266704559326
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,1536,0.004980266590913137
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,512,1024,0.021693867444992066
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,1024,0.00432640016078949
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,1024,0.0058794667323430385
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,512,768,0.022249599297841392
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,512,768,0.005108266572157542
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,768,0.0039658665657043455
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,768,0.005668266614278158
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,512,512,0.02143146594365438
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,512,512,0.004739200075467428
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,512,0.0036320000886917113
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,512,0.00556160012880961
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,512,256,0.019921066363652547
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,512,256,0.00453653335571289
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,768,12288,0.019946666558583577
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,256,0.0033141332368055976
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,256,0.005264000097910563
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,512,128,0.018889600038528444
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,512,128,0.0043712000052134195
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,128,0.0030464000999927522
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,128,0.0051818668842315676
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,512,64,0.018875734011332194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,64,0.0028778667251269023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,64,0.005218133330345154
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,512,32,0.018458666404088338
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,32,0.002916266769170761
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,32,0.005142400165398916
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,256,65536,0.07856533527374268
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,256,65536,0.06548693180084228
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,65536,0.012168533603350322
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,65536,0.05502293507258097
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,256,16384,0.03629013299942017
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,256,16384,0.01936639944712321
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,16384,0.007019733389218648
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,16384,0.01780160069465637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,256,12288,0.03266773422559102
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,256,12288,0.015658666690190635
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,12288,0.006790400048096975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,12288,0.014619732896486918
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,256,10240,0.031070933739344282
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,256,10240,0.013739732901255288
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,10240,0.007035733262697856
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,10240,0.013102933764457703
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,256,8192,0.02911679943402608
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,256,8192,0.0118367999792099
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,8192,0.006838400165239971
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,8192,0.011500799655914306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,256,7168,0.02832213242848714
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,256,7168,0.010743467013041179
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,7168,0.006794666747252147
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,768,1536,0.006684799989064534
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,7168,0.010792533556620281
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,256,6144,0.02726293404897054
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,256,6144,0.00997866690158844
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,6144,0.006733866532643636
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,6144,0.01009279986222585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,256,5120,0.026473599672317504
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,256,5120,0.008847999572753906
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,768,768,0.005525333185990652
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,5120,0.0069461335738499955
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,5120,0.00934933324654897
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,256,4096,0.024998400608698526
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,256,4096,0.008077866832415263
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,4096,0.006545066833496094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,4096,0.008393599589665731
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,256,3584,0.02495573361714681
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,256,3584,0.00746666689713796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,3584,0.007130666573842366
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,3584,0.00810346653064092
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,256,3072,0.024569600820541382
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,256,3072,0.006908800204594929
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,3072,0.0067562664548556015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,3072,0.007729066908359528
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,256,2560,0.02412266731262207
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,256,2560,0.006580266853173573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,2560,0.006287999947865804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,2560,0.007326933244864146
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,256,2048,0.022773333390553794
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,1536,0.004867200056711833
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,256,2048,0.0061749334136645
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,2048,0.0055637334783871974
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,2048,0.0067775999506314594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,256,1536,0.02291946609814962
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,256,1536,0.00561706672112147
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,1536,0.006342400113741558
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,256,1024,0.021486934026082358
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,256,1024,0.0051136001944541935
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,1024,0.004178133110205332
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,1024,0.00595413347085317
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,256,768,0.022181334098180135
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,256,768,0.004922666649023692
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,768,0.003874133278926214
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,768,0.0056970665852228795
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,256,512,0.021450666586558025
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,256,512,0.004635733366012573
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,512,0.0035125332574049628
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,512,0.005456000069777171
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,256,256,0.019989333550135293
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,256,256,0.004430933296680451
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,256,0.0032255999743938447
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,256,0.005236266553401947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,256,128,0.01882560054461161
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,256,128,0.004296533266703288
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,128,0.0029290666182835894
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,128,0.005147733290990194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,256,64,0.018707199891408285
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,64,0.0028437333802382152
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,64,0.005107200145721436
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,256,32,0.01843520005544027
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,32,0.002846933404604594
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,32,0.00521919975678126
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,128,65536,0.07881279786427817
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,128,65536,0.06573119958241781
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,65536,0.00995306670665741
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,65536,0.054929065704345706
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,128,16384,0.03633386691411336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,16384,0.006656000018119812
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,16384,0.017889066537221273
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,128,12288,0.03269760012626648
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,10240,0.006930133203665416
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,128,12288,0.01562773287296295
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,12288,0.006609066824118296
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,12288,0.014575999975204468
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,128,10240,0.03091946641604106
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,128,10240,0.013678933183352152
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,10240,0.01318186620871226
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,128,8192,0.029240532716115313
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,128,8192,0.011773866415023804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,8192,0.006753066678841908
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,8192,0.011442133784294128
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,128,7168,0.027856000264485675
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,128,7168,0.010675199826558431
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,7168,0.006622933348019918
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,7168,0.01086186667283376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,128,6144,0.02725013295809428
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,128,6144,0.009910399715105694
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,6144,0.0064074665307998655
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,6144,0.009965866804122925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,128,5120,0.02647999922434489
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,128,5120,0.008781866232554118
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,5120,0.007014399766921997
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,5120,0.0093941330909729
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,128,4096,0.025069866577784223
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,128,4096,0.00798399994770686
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,4096,0.006425599753856659
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,4096,0.008374399940172831
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,128,3584,0.02560639977455139
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,128,3584,0.007449600100517273
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,3072,0.006631466746330261
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,3584,0.006870399912198384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,3584,0.007949866851170858
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,128,3072,0.02403093377749125
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,128,3072,0.006870399912198384
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,3072,0.007670400043328603
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,128,2560,0.02489173412322998
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,128,2560,0.006478933493296306
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,2560,0.005846400062243144
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,2048,0.00673280010620753
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,2560,0.0072629332542419435
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,128,2048,0.02280319929122925
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,128,2048,0.006157866617043813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,2048,0.005283200244108836
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,128,1536,0.022963200012842813
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,128,1536,0.005568000177542368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,1536,0.004710400104522705
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,1536,0.006235733131567637
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,128,1024,0.021447465817133585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,128,1024,0.005074133475621542
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,1024,0.003999999910593033
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,1024,0.005913599828879039
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,128,768,0.02211093306541443
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,128,768,0.004745600124200186
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,768,0.0038304001092910765
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,768,0.00562666654586792
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,128,512,0.02157333294550578
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,128,512,0.004523733258247375
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,512,0.003517866631348928
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,512,0.0055071999629338585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,128,256,0.01989013353983561
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,128,256,0.0043722664316495265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,256,0.0031989333530267083
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,256,0.0053151999910672505
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,128,128,0.018863999843597413
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,128,128,0.00421013335386912
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,128,0.003020799905061722
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,128,0.005148800214131674
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,32,0.005123200019200643
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,64,0.002811733384927114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,128,64,0.018500266472498576
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,64,0.00522986650466919
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,128,32,0.01845226685206095
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,32,0.003206400076548258
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,16384,0.017723733186721803
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,65536,0.008326399823029835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,64,65536,0.07887146472930909
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,65536,0.055301332473754884
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,64,16384,0.036312532424926755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,16384,0.006632533172766368
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,64,12288,0.03276693423589071
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,12288,0.0064640000462532045
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,12288,0.014588800072669984
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,64,10240,0.03102186719576518
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,10240,0.006693333387374878
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,10240,0.013290666540463767
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,64,8192,0.02924373348553975
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,8192,0.006674133241176605
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,8192,0.011476266384124755
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,64,7168,0.02793813347816467
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,7168,0.006455466647942861
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,7168,0.010852266351381938
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,64,6144,0.027236266930898027
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,6144,0.006551466882228851
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,6144,0.01009279986222585
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,64,5120,0.025913600126902265
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,5120,0.006751999755700429
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,5120,0.009347200393676758
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,64,4096,0.02495253284772237
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,4096,0.006386133531729381
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,4096,0.008460799853007
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,64,3584,0.025085866451263428
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,3584,0.006838400165239971
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,3584,0.00809386670589447
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,64,3072,0.02405653397242228
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,3072,0.0064522668719291685
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,3072,0.007706666489442189
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,64,2560,0.024873600403467814
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,2560,0.005793066819508871
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,2560,0.007190399865309398
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,64,2048,0.022822399934132896
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,2048,0.0051925331354141235
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,2048,0.006720000008742015
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,64,1536,0.022765866915384927
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,1536,0.0046410664916038515
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,1536,0.006322133541107178
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,64,1024,0.021542400121688843
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,1024,0.003991466760635376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,1024,0.005885866781075796
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,64,768,0.021980800231297812
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,768,0.0036992001036802924
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,768,0.0057205334305763245
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,64,512,0.021489065885543824
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,512,0.0033173332611719764
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,512,0.005432533224423727
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,64,256,0.01990293264389038
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,128,0.005144533514976501
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,256,0.003160533308982849
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,256,0.005301333467165629
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,64,128,0.01885973413785299
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,128,0.0029535998900731405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,64,64,0.018771199385325114
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,32,0.005106133222579956
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,64,0.0028437333802382152
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,64,0.0051466668645540874
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,64,32,0.018784000476201376
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,32,0.0027903998891512555
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,65536,0.008726400136947633
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,32,65536,0.07908159891764323
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,65536,0.05501333475112915
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,32,16384,0.03642559846242269
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,16384,0.00651093324025472
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,16384,0.01768746574719747
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,32,12288,0.03259093364079793
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,12288,0.006586666901906331
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,12288,0.01455893317858378
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,32,10240,0.030957865715026855
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,10240,0.0066997334361076355
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,10240,0.013245866696039835
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,32,8192,0.029037866989771528
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,8192,0.006555733581384023
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,8192,0.011544533570607503
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,32,7168,0.027897600332895917
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,7168,0.006545066833496094
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,7168,0.01083626647790273
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,32,6144,0.02718720038731893
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,6144,0.006351999938488007
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,6144,0.010245333115259807
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,32,5120,0.02651626666386922
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,5120,0.006819200019041698
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,5120,0.009403733412424724
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,32,4096,0.024996266762415568
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,4096,0.006320000191529592
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,128,16384,0.019339734315872194
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,4096,0.008414933085441589
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,32,3584,0.02493120034535726
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,3584,0.006866133213043213
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,3584,0.008056533336639405
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,32,3072,0.0240831991036733
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,2048,0.005208533505598704
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,3072,0.006506666541099548
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,3072,0.007818666597207386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,32,2560,0.024229333798090617
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,2560,0.005840000013510386
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,2560,0.007246933380762736
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,32,2048,0.023528534173965453
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,2048,0.006682666639486949
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,32,1536,0.022925867636998495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,1536,0.004574933151404063
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,1536,0.0063509335120519
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,32,1024,0.021475199858347574
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,1024,0.004004266609748204
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,1024,0.005853866537412008
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,32,768,0.022216532627741495
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,768,0.0036608000596364343
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,768,0.005688533186912537
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,32,512,0.021461333831151327
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,512,0.0033962666988372804
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,512,0.005436799923578898
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,32,256,0.01991999944051107
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,256,0.0030037333567937215
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,256,0.005340800185998281
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,32,128,0.01877546707789103
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,128,0.00290133332212766
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,128,0.005134933193524678
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,32,64,0.018553600708643595
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,64,0.002825599908828735
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,64,0.005236266553401947
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,int8_wo,1,32,32,0.01851200064023336
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,32,0.0028661333024501802
SGLang,0.5.6.post2,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,32,0.005133866767088572
