framework,version,device,op_name,kernel_source,gemm_dtype,m,n,k,latency
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,65536,10240,7.437620374891494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,65536,12288,8.861886766221788
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,65536,16384,11.840764363606771
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,65536,10240,13.486104329427084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,65536,12288,16.107525295681423
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,65536,8192,6.012390984429254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,65536,8192,10.753941853841146
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,65536,16384,21.40900166829427
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,65536,7168,5.264990064832899
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,65536,7168,9.441312154134115
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,65536,6144,4.591687944200304
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,65536,6144,8.129182603624132
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,65536,5120,6.776112874348958
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,65536,5120,3.98674307929145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,65536,4096,5.455118391248916
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,65536,4096,3.1112933688693576
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,65536,12288,11.692501491970487
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,65536,3584,5.0079163445366754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,65536,3584,2.7938524881998696
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,65536,16384,15.919325086805555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,65536,8192,7.681201510959202
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,65536,3072,2.3572525448269315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,65536,6144,5.821163601345486
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,65536,3072,4.297508239746094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,65536,2560,1.9508525000678167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,65536,2560,3.644425710042318
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,65536,2048,2.9281609853108725
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,65536,2048,1.5723386340671117
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,65536,3584,3.4510116577148438
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,65536,7168,6.669915517171224
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,65536,1536,1.175747553507487
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,65536,1536,2.1437928941514754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,65536,1024,1.4993822309705946
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,65536,1024,0.7933226691351997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,65536,3072,2.928520838419596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,65536,5120,4.8606486850314665
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,65536,768,1.117038196987576
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,65536,768,0.6368826760186089
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,65536,512,0.7940550910101997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,65536,512,0.5251733462015787
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,65536,2560,2.453056971232096
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,65536,256,0.5071288744608561
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,65536,256,0.418229341506958
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,65536,1536,1.4575626585218642
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,65536,2048,2.005577723185221
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,65536,128,0.39976178275214297
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,65536,128,0.3963075478871663
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,65536,64,0.3607742256588406
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,65536,64,0.3845866786109076
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,65536,32,0.3544088999430339
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,65536,32,0.37997865676879883
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,65536,256,0.4285048908657498
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,65536,768,0.800148434109158
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,65536,512,0.590104897816976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,65536,10240,9.510037740071615
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,65536,1024,1.0096622043185763
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,65536,4096,3.8471111721462674
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,16384,16384,3.5710567898220487
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,16384,16384,5.493185679117839
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,16384,12288,4.174257914225261
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,16384,12288,2.7060943179660373
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,16384,10240,2.271624035305447
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,16384,10240,3.487750159369575
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,16384,8192,2.8049901326497397
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,16384,65536,14.405463324652779
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,16384,8192,1.8396434783935547
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,16384,7168,2.4578613705105252
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,16384,16384,4.818921830919054
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,16384,7168,1.6206578148735895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,16384,6144,2.0930008358425565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,16384,65536,22.015448676215275
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,16384,12288,3.5795360141330295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,16384,6144,1.4035519493950739
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,16384,5120,1.738775995042589
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,16384,5120,1.2591252856784396
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,16384,10240,2.996248033311632
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,16384,4096,1.3952986399332683
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,16384,4096,1.0015581978691948
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,16384,3584,1.1968960232204862
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,16384,3584,0.8984755410088433
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,16384,8192,2.402716530693902
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,16384,65536,20.51148139105903
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,16384,3072,0.7694648636711968
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,16384,3072,1.0168133841620552
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,16384,2560,0.8355724546644422
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,16384,2560,0.6362702051798502
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,16384,7168,2.1107574039035373
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,16384,2048,0.6540204683939616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,16384,2048,0.5136008792453343
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,16384,6144,1.8190587361653645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,16384,1536,0.5019537607828776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,16384,1536,0.3886008792453342
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,65536,128,0.40854043430752224
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,16384,1024,0.348582214779324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,16384,1024,0.26500709851582843
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,16384,5120,1.4745741950141058
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,16384,768,0.2860026624467638
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,16384,768,0.20699911647372773
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,16384,4096,1.1967404683430989
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,16384,512,0.21169600221845838
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,16384,3584,1.04985597398546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,16384,512,0.1786479949951172
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,16384,256,0.12120621734195286
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,16384,256,0.1316995620727539
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,16384,2560,0.7627769046359592
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,16384,3072,0.912650638156467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,16384,128,0.094577776061164
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,16384,128,0.11822577317555745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,16384,64,0.0874933335516188
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,16384,64,0.11563555399576823
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,16384,32,0.08732177813847859
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,16384,32,0.11101066403918797
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,16384,2048,0.620216899447971
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,16384,1536,0.48364798227945965
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,16384,768,0.2786746554904514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,16384,256,0.14072799682617188
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,16384,512,0.2149511045879788
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,16384,1024,0.34636354446411133
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,12288,16384,2.9708614349365234
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,12288,16384,4.019774119059245
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,12288,12288,2.2504088083902993
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,12288,12288,2.998294194539388
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,12288,10240,1.8969048394097223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,12288,10240,2.5059405432807074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,12288,65536,11.825123257107206
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,12288,8192,1.5340995788574219
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,12288,8192,2.019029405381944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,12288,65536,16.08410474989149
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,12288,7168,1.7618524763319228
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,12288,7168,1.3467164569430883
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,12288,6144,1.5066657596164281
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,12288,6144,1.1686710781521268
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,12288,16384,3.9557957119411893
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,12288,5120,1.227768898010254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,12288,5120,1.0425661934746635
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,12288,12288,2.9470337761773
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,12288,4096,0.9431591033935547
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,12288,4096,0.8399813440110949
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,12288,10240,2.4575555589463973
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,12288,3584,0.8009600109524198
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,12288,3584,0.7471662097507052
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,12288,8192,2.007042566935221
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,12288,65536,16.54251946343316
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,12288,3072,0.7171786626180013
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,12288,3072,0.6355991363525391
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,12288,2560,0.5875048637390137
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,12288,2560,0.5278773307800293
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,12288,7168,1.7315235137939453
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,12288,2048,0.47605064180162215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,12288,2048,0.4284568892584906
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,12288,6144,1.500343958536784
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,12288,1536,0.36675379011366105
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,12288,1536,0.32351912392510307
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,16384,128,0.12552711698744032
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,12288,1024,0.2533786561754015
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,12288,1024,0.22318222787645128
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,12288,5120,1.221308496263292
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,12288,768,0.214939554532369
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,12288,768,0.17623289426167807
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,12288,4096,0.9812186559041342
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,12288,512,0.15404178036583796
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,12288,512,0.1400346623526679
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,12288,3584,0.8670550982157389
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,12288,256,0.10250755813386704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,12288,256,0.11104355255762736
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,12288,2560,0.6399875746832954
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,12288,128,0.06790844599405925
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,12288,128,0.090064889854855
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,12288,3072,0.7533164554172092
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,12288,64,0.06338489055633545
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,12288,64,0.09875910811954075
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,12288,32,0.06325955523384942
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,12288,32,0.09001333183712429
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,12288,2048,0.5210480160183376
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,12288,1536,0.4092986583709717
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,12288,256,0.11703377299838597
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,12288,768,0.2345600128173828
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,12288,512,0.17921955055660674
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,10240,16384,2.66645876566569
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,12288,1024,0.2936391035715739
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,10240,16384,3.486029307047526
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,10240,12288,2.0125999450683594
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,10240,12288,2.645378748575846
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,10240,10240,1.6969377729627821
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,10240,10240,2.196686214870877
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,10240,65536,10.597063700358072
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,10240,8192,1.3696586820814345
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,10240,65536,14.056214226616753
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,10240,8192,1.784762700398763
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,10240,7168,1.2077858183119032
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,10240,7168,1.5057893329196508
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,10240,6144,1.2092488606770833
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,10240,6144,1.0505004458957248
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,10240,16384,3.5370843675401478
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,10240,5120,1.0205280515882704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,10240,5120,0.9391857782999674
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,10240,12288,2.6441733042399087
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,10240,4096,0.7820960150824653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,10240,4096,0.7413359747992622
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,10240,10240,2.2162354787190757
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,10240,65536,14.868311564127604
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,10240,8192,1.7641297446356878
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,10240,3584,0.6808773146735297
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,10240,3584,0.6641599867078993
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,10240,3072,0.5650862058003744
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,10240,3072,0.6054168807135688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,10240,7168,1.5597466362847223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,10240,2560,0.5172933472527398
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,10240,2560,0.47367021772596574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,10240,2048,0.40591732660929364
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,10240,2048,0.38350754313998753
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,10240,6144,1.3335484398735895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,12288,128,0.10061244169871013
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,10240,1536,0.3181680043538411
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,10240,1536,0.2900933424631755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,10240,1024,0.2206044461992052
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,10240,1024,0.20002310805850557
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,10240,5120,1.100768937004937
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,10240,768,0.18404711617363823
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,10240,768,0.15831822819179958
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,10240,4096,0.8864942126803927
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,10240,512,0.13157333268059626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,10240,512,0.12294222248925103
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,10240,3584,0.7848311000400119
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,10240,256,0.08725421958499485
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,10240,256,0.09583377838134766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,10240,2560,0.5802888870239258
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,10240,128,0.06148444281684028
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,10240,128,0.07950222492218018
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,10240,3072,0.6777893172370063
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,10240,64,0.05590933561325073
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,10240,64,0.07989333073298137
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,10240,32,0.056571556462181934
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,10240,32,0.07605689101748996
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,10240,1536,0.3655982282426622
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,10240,2048,0.4718480110168457
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,10240,1024,0.2661457856496175
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,10240,768,0.21152267191145155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,10240,256,0.10635199811723496
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,10240,512,0.16164355807834202
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,8192,16384,2.8811832004123263
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,8192,16384,2.3627707163492837
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,8192,12288,2.0985609690348306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,8192,65536,9.33902316623264
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,8192,12288,1.7897137535942926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,8192,10240,1.4942017661200628
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,8192,10240,1.8537591298421223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,8192,8192,1.21214665306939
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,8192,8192,1.3959173626369898
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,8192,65536,11.423791673448351
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,8192,7168,1.089421378241645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,8192,7168,1.0726239946153429
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,8192,6144,1.0364977518717449
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,8192,6144,0.9266160329182943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,8192,16384,3.1499760945638022
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,8192,5120,0.8337413469950358
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,8192,12288,2.351361804538303
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,8192,10240,1.9760665893554688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,8192,5120,0.7897751066419813
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,8192,4096,0.6158684624565972
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,8192,4096,0.6558168729146322
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,8192,3584,0.5790230962965224
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,8192,65536,13.021053738064237
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,8192,3584,0.543325318230523
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,8192,8192,1.5580355326334636
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,8192,3072,0.48575380113389754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,8192,3072,0.461042669084337
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,8192,2560,0.4091608789232042
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,8192,2560,0.398926231596205
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,8192,7168,1.3887679841783311
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,8192,2048,0.3360319932301839
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,8192,2048,0.3093831009334988
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,8192,6144,1.178360939025879
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,8192,1536,0.25969600677490234
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,8192,1536,0.23840978410508898
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,10240,128,0.09223466449313694
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,8192,1024,0.17985778384738496
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,8192,1024,0.16872710651821562
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,8192,5120,0.9830924140082465
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,8192,3584,0.7044515609741211
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,8192,768,0.14113510979546443
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,8192,768,0.13544266753726536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,8192,512,0.10277511013878717
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,8192,512,0.10609244638019139
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,8192,4096,0.7967733277214898
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,8192,256,0.0639537771542867
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,8192,256,0.07898577716615465
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,8192,2560,0.5152142312791612
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,8192,128,0.04960888955328199
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,8192,128,0.06743999984529284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,8192,3072,0.610172430674235
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,8192,64,0.045047110981411405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,8192,64,0.06588533189561632
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,8192,32,0.04617955618434482
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,8192,32,0.06338755289713542
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,8192,2048,0.42175910207960343
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,8192,1536,0.3301937845018175
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,8192,256,0.09436711337831284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,8192,128,0.08026133643256293
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,8192,512,0.14308088355594212
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,7168,65536,8.730200025770399
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,8192,1024,0.2393724388546414
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,8192,768,0.1891111135482788
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,7168,16384,2.213714599609375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,7168,16384,2.3714436425103083
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,7168,12288,1.6736542383829753
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,7168,12288,1.7925359937879775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,7168,65536,9.443179660373264
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,7168,10240,1.4883688820732965
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,7168,10240,1.4100284576416016
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,7168,8192,1.1353653801812065
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,7168,8192,1.1857946183946397
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,7168,7168,1.0010799831814237
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,7168,7168,1.0059049394395616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,7168,6144,0.8684142430623373
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,7168,6144,0.7888604270087348
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,7168,5120,0.6934737629360623
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,7168,5120,0.7421911027696398
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,7168,65536,12.165264553493925
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,7168,12288,2.185007095336914
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,7168,16384,3.0021218193901906
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,7168,4096,0.589004463619656
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,7168,4096,0.5420719782511393
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,7168,3584,0.5231431325276693
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,7168,10240,1.8287376827663844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,7168,3584,0.47729333241780597
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,7168,3072,0.4462951024373372
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,7168,3072,0.4150657653808594
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,7168,8192,1.4774853388468425
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,7168,2560,0.3526204427083333
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,7168,2560,0.37111732694837785
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,7168,7168,1.3037652969360352
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,7168,2048,0.2810746563805474
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,7168,2048,0.3029990990956624
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,7168,1536,0.21552356084187826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,7168,1536,0.22841066784328887
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,7168,6144,1.1088764402601454
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,7168,1024,0.1557662222120497
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,7168,1024,0.1638808912701077
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,7168,5120,0.9361769358317057
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,7168,768,0.12417156166500515
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,7168,768,0.1292462216483222
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,7168,3584,0.6632266574435765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,7168,512,0.09020977550082737
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,7168,512,0.09964711136288112
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,7168,2560,0.48593865500556094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,7168,4096,0.7507253222995334
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,7168,256,0.05714222457673815
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,7168,256,0.07380533218383789
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,7168,128,0.044450667169359
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,7168,128,0.06332355737686157
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,7168,1536,0.3128933376736111
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,7168,64,0.04083377785152859
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,7168,64,0.061221334669325084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,7168,32,0.041125333971447416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,7168,32,0.05886399745941162
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,7168,3072,0.5733475685119629
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,7168,2048,0.40060443348354763
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,7168,768,0.17710488372378877
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,7168,1024,0.22417243321736655
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,6144,16384,2.0459982554117837
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,6144,16384,2.0601423051622176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,6144,65536,8.098987155490452
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,7168,256,0.08804000086254543
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,6144,12288,1.3895671632554796
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,7168,512,0.1347128947575887
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,6144,12288,1.5592426723904078
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,6144,65536,8.124729580349392
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,6144,10240,1.306489732530382
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,6144,10240,1.2649449242485893
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,6144,8192,0.9772115283542209
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,6144,8192,1.055763562520345
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,6144,7168,0.9293475680881076
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,6144,7168,0.7910088962978787
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,6144,6144,0.7083093325297037
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,6144,6144,0.8064897855122884
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,6144,16384,2.748841815524631
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,6144,5120,0.5848968823750814
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,6144,5120,0.6915306515163846
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,6144,12288,2.0364702012803817
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,6144,10240,1.7019066280788844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,6144,4096,0.5448275672064887
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,6144,4096,0.46609777874416775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,6144,65536,11.160765753851997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,6144,3584,0.48449865976969403
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,6144,3584,0.41698487599690753
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,6144,8192,1.3702132966783311
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,6144,3072,0.35514222251044375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,6144,3072,0.4115600056118435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,6144,2560,0.2999173270331489
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,6144,2560,0.3445484373304579
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,6144,7168,1.201517317030165
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,6144,2048,0.24408888816833496
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,6144,2048,0.27738844023810494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,6144,6144,1.039669354756673
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,6144,1536,0.18856710857815215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,6144,1536,0.21359911229875353
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,7168,128,0.07833421892589994
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,6144,1024,0.13236622015635172
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,6144,1024,0.15193155076768664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,6144,5120,0.8719591034783257
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,6144,768,0.10565777619679768
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,6144,768,0.11938399738735622
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,6144,3584,0.6207626660664877
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,6144,512,0.07803733481301202
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,6144,512,0.09200355741712783
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,6144,2560,0.4570630921257867
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,6144,256,0.04951111144489712
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,6144,256,0.06813422176573011
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,6144,4096,0.7022364404466418
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,6144,128,0.0389120015833113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,6144,128,0.0570017761654324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,6144,3072,0.5349404546949599
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,6144,64,0.03460266523891025
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,6144,64,0.0556817783249749
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,6144,32,0.03564533260133531
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,6144,32,0.05275466707017687
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,6144,1536,0.29255734549628365
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,6144,2048,0.3742728763156467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,5120,65536,7.487236022949219
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,6144,256,0.08304888672298856
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,6144,128,0.06902133093939887
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,6144,512,0.12551821602715388
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,5120,65536,7.084486219618056
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,5120,16384,1.7620230780707464
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,6144,768,0.1645146608352661
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,5120,16384,1.9088285234239366
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,5120,12288,1.4496018091837566
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,5120,12288,1.2979670630560982
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,5120,10240,0.9439564810858833
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,5120,10240,1.2103075451321073
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,5120,8192,0.816739559173584
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,5120,8192,0.9818026224772135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,5120,7168,0.8634675343831381
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,6144,1024,0.20942933029598662
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,5120,7168,0.697439087761773
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,5120,6144,0.7529270913865832
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,5120,6144,0.5961075358920628
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,5120,5120,0.4922657542758518
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,5120,5120,0.6338186793857151
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,5120,65536,10.244716220431858
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,5120,4096,0.3971449004279242
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,5120,16384,2.516882578531901
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,5120,4096,0.505883534749349
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,5120,12288,1.9211608039008246
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,5120,3584,0.34800977177090114
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,5120,3584,0.4472159809536404
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,5120,3072,0.2971644401550293
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,5120,10240,1.5862017737494574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,5120,3072,0.37931111123826766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,5120,8192,1.2785689036051433
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,5120,2560,0.2521039909786648
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,5120,2560,0.3200968901316325
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,5120,2048,0.20893777741326225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,5120,2048,0.25828533702426487
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,5120,7168,1.124500486585829
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,5120,1536,0.1593422227435642
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,5120,1536,0.1994382275475396
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,5120,6144,0.9656302134195963
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,5120,1024,0.11358043882581924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,5120,1024,0.1403475602467855
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,5120,5120,0.8144346872965494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,5120,768,0.0937546690305074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,5120,768,0.11191999912261963
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,5120,3584,0.581224865383572
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,5120,512,0.06820888651741876
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,5120,512,0.08476889133453369
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,5120,2560,0.4289591047498915
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,5120,256,0.04551733202404446
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,5120,256,0.06245244211620755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,5120,4096,0.6568960083855523
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,5120,128,0.03319022059440613
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,5120,128,0.05119111140569051
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,5120,3072,0.5011742379930284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,5120,64,0.029751112063725788
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,5120,64,0.049280888504452176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,5120,1536,0.27227555380927193
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,5120,32,0.030986666679382324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,5120,32,0.04691911074850294
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,5120,2048,0.35163556204901797
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,5120,768,0.15461599826812744
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,4096,16384,1.3291617499457467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,5120,1024,0.1949395603603787
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,4096,16384,1.7540782292683919
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,4096,12288,0.9238737954033746
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,4096,65536,6.896140628390842
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,4096,12288,1.324618657430013
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,4096,65536,5.401946597629124
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,4096,10240,1.1096542146470811
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,4096,10240,0.8067928949991862
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,5120,128,0.06371377574072944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,5120,256,0.07654488748974271
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,5120,512,0.11710222562154134
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,4096,8192,0.6283413569132487
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,4096,8192,0.8939946492513021
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,4096,7168,0.7989564471774631
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,4096,7168,0.5289306640625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,4096,6144,0.6949902110629611
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,4096,6144,0.46395291222466367
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,4096,5120,0.3837724526723226
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,4096,5120,0.564025772942437
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,4096,16384,2.3407679663764105
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,4096,4096,0.3103475570678711
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,4096,4096,0.45281420813666445
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,4096,12288,1.763061311509874
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,4096,3584,0.27192621760898167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,4096,3584,0.40186309814453125
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,4096,65536,9.342535230848524
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,4096,3072,0.2405600017971463
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,4096,3072,0.3454684416453044
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,4096,10240,1.4685946570502386
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,4096,2560,0.19934933715396455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,4096,2560,0.29094311926099986
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,4096,7168,1.038898679945204
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,4096,8192,1.1901733610365126
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,4096,2048,0.1636026700337728
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,4096,2048,0.23517600695292154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,4096,1536,0.12641333209143743
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,4096,1536,0.1811671124564277
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,4096,6144,0.8947120242648654
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,4096,1024,0.09030489126841228
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,4096,1024,0.12783111466301814
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,4096,5120,0.754903104570177
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,4096,768,0.07271111011505127
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,4096,768,0.09972177611456977
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,4096,3584,0.5389422310723199
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,4096,512,0.054309331708484225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,4096,512,0.07636800077226427
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,4096,4096,0.6093857553270128
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,4096,256,0.03547999925083584
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,4096,256,0.05460355679194132
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,4096,2560,0.3968204392327203
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,4096,128,0.02776088813940684
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,4096,128,0.045180443260404796
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,4096,1536,0.25390932295057506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,4096,64,0.02414488957987891
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,4096,64,0.04257955484920078
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,4096,32,0.025311110748185053
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,4096,32,0.041201776928371854
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,4096,3072,0.4680150879753961
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,4096,2048,0.3253226545121935
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,4096,768,0.1439182228512234
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,4096,1024,0.18160977628495958
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3584,16384,1.1835252973768446
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3584,16384,1.6773209042019313
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3584,65536,4.765227423773871
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3584,65536,6.610593583848741
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3584,12288,0.83151732550727
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3584,12288,1.2649759716457791
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3584,10240,0.6891813278198242
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3584,10240,1.0602782567342122
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,4096,256,0.0717137787077162
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3584,8192,0.5235413445366753
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,4096,512,0.10729955302344428
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3584,8192,0.8539351357354058
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3584,7168,0.4638649092780219
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3584,7168,0.7517582045661079
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3584,6144,0.3974800109863281
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3584,6144,0.6524960199991862
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,3584,16384,2.2547181447347007
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3584,5120,0.33436176511976456
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,3584,12288,1.6910809410942926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3584,5120,0.5510026613871256
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3584,4096,0.2734195656246609
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3584,4096,0.43731021881103516
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,3584,65536,8.987400478786892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,3584,10240,1.4165137608846028
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3584,3584,0.23904000388251412
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3584,3584,0.39163910018073184
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3584,3072,0.20824710528055826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3584,3072,0.3350515630510118
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,3584,8192,1.1448649300469291
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3584,2560,0.17530755201975504
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3584,2560,0.2821377648247613
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,3584,7168,1.0010426839192708
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3584,2048,0.14409955342610678
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3584,2048,0.22876177893744576
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,3584,6144,0.8628391159905328
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3584,1536,0.1115093363655938
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3584,1536,0.17596711052788627
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,3584,3584,0.5213600264655219
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,4096,128,0.0591848889986674
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3584,1024,0.0799822211265564
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3584,1024,0.12480355633629693
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3584,768,0.06479022237989637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3584,768,0.09684444136089748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,3584,5120,0.7260284423828125
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3584,512,0.04818400078349643
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3584,512,0.07313155465655856
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,3584,2560,0.3839724328782823
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3584,256,0.031834665271970965
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3584,256,0.052410665485594005
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,3584,4096,0.5897066858079698
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3584,128,0.024812445044517517
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3584,128,0.04186311033036974
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,3584,1536,0.2455751101175944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3584,64,0.023548444112141926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,3584,3072,0.45107465320163304
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3584,64,0.03956088754865859
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3584,32,0.02382044494152069
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3584,32,0.03813866774241129
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,3584,2048,0.31357065836588544
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3072,65536,6.315656873914931
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3072,65536,4.051449669731988
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3072,16384,0.8923698001437717
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3072,16384,1.6034808688693578
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,3584,128,0.0552151103814443
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,3584,256,0.06794489092297025
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3072,12288,0.6862888866000705
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3072,12288,1.2083955340915256
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,3584,512,0.10244266854392158
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3072,10240,1.0115120146009657
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3072,10240,0.5828204684787327
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3072,8192,0.8278417587280273
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3072,8192,0.4524764484829373
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,3584,768,0.1367457840177748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3072,7168,0.7192551294962565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3072,7168,0.39978755844963926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3072,6144,0.34704889191521543
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3072,6144,0.6220302051968044
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,3584,1024,0.1738871071073744
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3072,5120,0.2898106575012207
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3072,5120,0.5234862433539497
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,3072,65536,8.598642985026041
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3072,4096,0.23643112182617188
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,3072,16384,2.159059524536133
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3072,4096,0.41939467853969997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3072,3584,0.2071653339597914
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3072,3584,0.3755546675788031
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,3072,10240,1.3573030895657008
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,3072,12288,1.6318800184461806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3072,3072,0.17992711067199707
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3072,3072,0.3226604461669922
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3072,2560,0.1529502206378513
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3072,2560,0.2716577847798665
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,3072,8192,1.0933092965020073
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3072,2048,0.12451199690500896
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3072,2048,0.21977244483100045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,3072,7168,0.9572932985093859
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3072,1536,0.09819289048512776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3072,1536,0.1702728933758206
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,3072,6144,0.8299324247572156
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3072,1024,0.0704631143146091
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3072,1024,0.11896444691552056
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,3072,5120,0.69704712761773
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3072,768,0.056530667675866016
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3072,768,0.09204177723990546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,3072,3584,0.5001040034823948
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3072,512,0.04253244400024414
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3072,512,0.06917866733339098
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,3072,2560,0.36769334475199383
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3072,256,0.02864266766442193
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,3072,4096,0.5646293428209093
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3072,256,0.04941155513127645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3072,128,0.021133333444595337
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3072,128,0.03927288783921136
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,3072,1536,0.23528711001078287
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3072,64,0.019896000623703003
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3072,64,0.03699644406636556
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,3072,32,0.020946666598320007
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,3072,3072,0.434171544180976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,3072,32,0.0352168877919515
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,3072,2048,0.30128712124294704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,3072,768,0.1304453346464369
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2560,16384,0.8396124309963651
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,3072,1024,0.16701866520775688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2560,65536,6.013385772705078
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2560,65536,3.5184569888644748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2560,16384,1.5313022401597765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2560,12288,0.5644622378879124
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2560,12288,1.1618435117933485
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2560,10240,0.48717064327663845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2560,10240,0.9689182705349393
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2560,8192,0.3871475590599908
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2560,8192,0.7862737973531088
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,3072,512,0.0978435542848375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,3072,128,0.051627556482950844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,3072,256,0.06517510943942599
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2560,7168,0.33722755644056535
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2560,7168,0.6833502451578776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2560,6144,0.2907608879937066
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2560,6144,0.5904035568237305
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2560,5120,0.24351467026604545
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2560,5120,0.5003271102905273
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,2560,16384,2.068842569986979
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2560,4096,0.19714400503370497
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,2560,12288,1.5523199505276148
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2560,4096,0.3997200065188938
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2560,3584,0.17434044679005942
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2560,3584,0.35805866453382706
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,2560,10240,1.298458629184299
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2560,3072,0.15203377935621473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,2560,65536,8.22582753499349
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2560,3072,0.3086097770267063
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2560,2560,0.12896177503797743
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2560,2560,0.2587671014997694
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,2560,8192,1.048283576965332
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2560,2048,0.1065795554055108
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2560,2048,0.2093048890431722
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,2560,7168,0.9206097920735677
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,2560,6144,0.7931289143032498
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2560,1536,0.08355466524759929
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2560,1536,0.16213866074879965
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,2560,5120,0.667985757191976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2560,1024,0.0603368878364563
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2560,1024,0.1139777766333686
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2560,768,0.048897776338789195
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2560,768,0.0869342221154107
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,2560,3584,0.47944000032212997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2560,512,0.037436445554097496
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2560,512,0.06509511338339911
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,2560,2560,0.3513440026177301
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2560,256,0.025957332717047796
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2560,256,0.046006222565968834
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,2560,4096,0.5421253310309516
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2560,128,0.019651555352740817
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2560,128,0.036129779285854764
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,2560,3072,0.4152542220221625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2560,64,0.01646399994691213
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2560,64,0.03369422091378106
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2560,32,0.01694222291310628
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2560,32,0.03194133440653483
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,2560,1536,0.2259706656138102
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,2560,2048,0.2901715437571208
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,2560,768,0.12310843997531468
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2048,16384,0.6057893435160319
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2048,16384,1.4547857708401148
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2048,65536,2.734996371799045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,2560,1024,0.15931199656592474
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2048,65536,5.721608903672959
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2048,12288,0.45141511493259007
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2048,12288,1.0929786894056532
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2048,10240,0.37722134590148926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2048,10240,0.9152124192979602
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2048,8192,0.30336088604397243
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2048,8192,0.7378373146057129
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,2560,512,0.09349777963426377
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2048,7168,0.2660204569498698
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,2560,128,0.04967911044756571
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,2560,256,0.06257777743869357
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2048,6144,0.23132533497280547
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2048,7168,0.6538577609592015
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2048,6144,0.561281787024604
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2048,5120,0.19492888450622559
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2048,5120,0.48083999421861434
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,2048,16384,1.9697394900851781
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2048,4096,0.15855822298261854
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,2048,12288,1.4836444854736328
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2048,4096,0.38083823521931964
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2048,3584,0.14038577344682482
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,2048,65536,7.811832004123264
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2048,3584,0.3393990993499756
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2048,3072,0.1220995585123698
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2048,3072,0.2913644313812256
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,2048,10240,1.242213355170356
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2048,2560,0.1030542254447937
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2048,2560,0.24567021263970268
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,2048,8192,1.001753807067871
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2048,2048,0.08468444479836358
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2048,2048,0.20011377334594727
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,2048,7168,0.8808462354871961
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2048,1536,0.06710044542948405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,2048,6144,0.7599368625217013
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2048,1536,0.15351200103759766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2048,1024,0.04900888933075798
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2048,1024,0.10685244533750747
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,2048,5120,0.6363342073228624
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2048,768,0.040051556295818753
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2048,768,0.08224711153242323
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,2048,3584,0.45822932985093856
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2048,512,0.030442665020624798
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2048,512,0.06109688679377238
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,2048,2560,0.3367866675059001
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2048,256,0.021280889709790547
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2048,256,0.042063110404544406
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,2048,4096,0.5172808965047201
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2048,128,0.015639111399650574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2048,128,0.03236888845761617
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,2048,3072,0.396994670232137
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2048,64,0.01311288856797748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2048,64,0.029880000485314265
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,2048,1536,0.21625244617462158
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,2048,32,0.014449778530332776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,2048,32,0.028020444843504164
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,2048,768,0.11863822407192653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,2048,2048,0.2765146626366509
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1536,16384,0.4537919892205133
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1536,16384,1.3774346245659723
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1536,65536,2.0085760752360025
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1536,12288,0.3494257662031386
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,2048,1024,0.15182933542463514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1536,65536,5.419305589463975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1536,12288,1.041559961107042
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1536,10240,0.28290843963623047
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1536,10240,0.870848019917806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1536,8192,0.23142311308119032
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1536,8192,0.7012533081902398
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,2048,256,0.05936888853708903
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1536,7168,0.20090222358703613
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,2048,512,0.08882666958702935
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1536,7168,0.6146871248881022
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1536,6144,0.17549067073398164
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1536,6144,0.5314968956841363
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,1536,16384,1.8839662339952257
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1536,5120,0.14697066942850748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1536,5120,0.4532622231377496
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,1536,12288,1.4179102579752605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,1536,10240,1.1892408794826932
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1536,4096,0.12041689289940728
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1536,4096,0.36461599667867023
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1536,3584,0.10633333524068196
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1536,3584,0.32459378242492676
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,1536,65536,7.437182956271702
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,1536,8192,0.9563751220703125
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1536,3072,0.09352355533175999
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1536,3072,0.27878043386671275
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1536,2560,0.07950133085250854
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1536,2560,0.2348897722032335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,1536,7168,0.8407991197374133
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1536,2048,0.06610577636294894
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1536,2048,0.18999821609920928
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,1536,6144,0.7263546519809299
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1536,1536,0.05195822318394979
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1536,1536,0.14697689480251735
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,1536,3584,0.4388471179538303
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1536,1024,0.03824444280730353
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,1536,5120,0.611019558376736
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1536,1024,0.10144444306691487
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1536,768,0.031211555004119873
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1536,768,0.07646577888064914
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,1536,2560,0.32224800851609975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1536,512,0.02436711059676276
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1536,512,0.05761155817243788
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,1536,4096,0.4963822364807129
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1536,256,0.01685688893000285
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1536,256,0.03895733422703213
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,2048,128,0.046647111574808754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1536,128,0.013037333057986366
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1536,128,0.02938400043381585
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,1536,3072,0.3800657855139838
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1536,64,0.011381333072980246
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1536,64,0.02593155536386702
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1536,32,0.011524444652928246
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1536,32,0.023823999696307715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1024,65536,1.3610977596706813
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,1536,1536,0.20741777949863008
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,1536,2048,0.2654871145884196
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1024,16384,0.2971715662214491
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1024,16384,1.3038355509440105
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1024,65536,5.154602474636502
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1024,12288,0.2251288890838623
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1024,12288,0.9825866487291125
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,1536,128,0.04378488990995619
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,1536,256,0.05759288867314657
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1024,10240,0.18955022758907744
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,1536,512,0.08558489216698541
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1024,10240,0.8214204576280383
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1024,8192,0.15304622385236952
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1024,8192,0.6617609130011665
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1024,7168,0.13529511292775473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1024,7168,0.5828453169928657
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,1536,768,0.11336178249782985
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1024,6144,0.11608621809217666
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1024,6144,0.5016266504923502
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,1536,1024,0.14475911193423802
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1024,5120,0.09768355555004543
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1024,5120,0.4273413287268744
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,1024,16384,1.7892195383707683
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1024,4096,0.07950577470991346
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1024,4096,0.3453528881072998
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,1024,12288,1.3484613630506728
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1024,3584,0.07133511039945814
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1024,3584,0.30603822072347003
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,1024,65536,7.086954752604167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1024,3072,0.06244977977540758
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1024,3072,0.26347467634412974
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,1024,10240,1.1297883987426758
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1024,2560,0.05304444498485989
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,1024,8192,0.9102293650309244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1024,2560,0.22189954916636148
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1024,2048,0.043699555926852755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1024,2048,0.18040088812510172
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,1024,7168,0.8011724683973523
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1024,1536,0.03438755538728502
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1024,1536,0.13953510920206705
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,1024,6144,0.6924115816752116
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1024,1024,0.024932444095611572
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1024,1024,0.09579200214809841
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,1024,5120,0.581766234503852
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1024,768,0.020407110452651978
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1024,768,0.07228000296486749
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,1024,3584,0.4174915684594049
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1024,512,0.015597333510716757
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1024,512,0.052926222483317055
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,1024,4096,0.47252268261379665
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1024,256,0.011025778121418424
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1024,256,0.03351733419630263
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,1024,3072,0.36359466446770566
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1024,128,0.008391111261314815
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,1024,2560,0.3070426781972249
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1024,128,0.024449777272012498
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,1024,1536,0.19553066624535453
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1024,64,0.007314666277832455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1024,64,0.021720000439220007
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,1024,32,0.007640889121426477
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,1024,32,0.020431111256281536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,1024,2048,0.2530328962537977
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,768,65536,1.0498444239298503
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,768,16384,0.2613946596781413
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,768,16384,1.3043218188815646
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,768,65536,5.141123453776042
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,768,12288,0.17404354943169487
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,1024,768,0.10747822125752766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,1024,1024,0.13726666238572863
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,768,12288,0.9817209243774414
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,768,10240,0.1451137728161282
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,768,10240,0.8218666712443033
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,768,8192,0.11782666047414143
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,768,8192,0.6630542013380263
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,1024,256,0.05085955394638909
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,768,7168,0.10285422537061904
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,1024,512,0.08042577902475993
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,768,7168,0.5816186798943414
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,768,6144,0.08979555633332993
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,768,6144,0.5012684398227268
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,768,16384,1.7450346416897242
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,768,5120,0.07592711183759901
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,768,5120,0.41435911920335555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,768,65536,6.904999203152126
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,768,10240,1.1023057301839192
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,768,4096,0.0625564455986023
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,768,12288,1.3151626586914062
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,768,4096,0.33464799986945254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,768,3584,0.05532711082034641
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,768,3584,0.2959768772125244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,768,3072,0.048271109660466514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,768,8192,0.8877840042114258
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,768,3072,0.2559013366699219
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,768,2560,0.04109777675734626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,768,2560,0.21607644028133818
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,768,7168,0.7790942192077637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,768,2048,0.03421777817938063
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,768,2048,0.1759155591328939
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,768,6144,0.6738302442762586
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,768,1536,0.02679022153218587
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,768,1536,0.1357626650068495
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,1024,128,0.038466665479871966
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,768,1024,0.019549333386951022
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,768,1024,0.09473599990208943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,768,3584,0.4071857664320204
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,768,768,0.01605333387851715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,768,2560,0.30023643705579967
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,768,768,0.07175644238789876
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,768,5120,0.567354679107666
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,768,512,0.012503999802801343
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,768,512,0.051157332128948636
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,768,256,0.008818666968080733
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,768,256,0.030973331795798406
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,768,4096,0.4607751104566786
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,768,128,0.007093333535724216
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,768,128,0.02321333355373806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,768,3072,0.3538115554385715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,768,64,0.006127999888526068
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,768,64,0.020485333270496793
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,768,32,0.006468444648716185
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,768,32,0.01907999979125129
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,768,1536,0.19042578008439806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,512,65536,0.5901618003845215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,768,2048,0.2458275424109565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,512,16384,0.15220355987548828
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,512,16384,1.2281333075629342
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,512,65536,4.836995442708333
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,512,12288,0.11537333329518636
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,512,12288,0.9229075113932291
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,768,1024,0.13415022691090903
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,512,10240,0.09719022115071614
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,512,10240,0.7726524670918783
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,768,256,0.050245334704717
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,768,128,0.0370071099864112
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,512,8192,0.07878577709197998
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,512,8192,0.6221013069152832
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,512,7168,0.06971200307210286
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,512,7168,0.5466533237033421
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,768,512,0.07854133182101779
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,512,6144,0.06036800146102905
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,512,6144,0.471176889207628
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,768,768,0.10508710808224148
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,512,5120,0.05176977647675408
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,512,5120,0.40028799904717344
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,512,16384,1.701033804151747
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,512,4096,0.042599111795425415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,512,12288,1.2837964163886177
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,512,4096,0.324257771174113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,512,65536,6.7259368896484375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,512,3584,0.03805689016977946
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,512,3584,0.29119110107421875
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,512,3072,0.03324000040690104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,512,3072,0.24769510163201225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,512,10240,1.0764488644070096
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,512,2560,0.028629332780838013
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,512,8192,0.8671058019002279
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,512,2560,0.2086320055855645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,512,2048,0.023732443650563557
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,512,2048,0.16983466678195527
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,512,7168,0.7620159784952799
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,512,1536,0.01868533260292477
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,512,1536,0.13092533747355142
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,512,6144,0.6580613454182943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,512,1024,0.013988445202509562
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,512,1024,0.09109333488676284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,512,5120,0.554366217719184
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,512,768,0.011648000114493899
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,512,768,0.06842044326994154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,512,3584,0.39755911297268337
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,512,512,0.009339555270142024
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,512,512,0.04817510975731743
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,512,4096,0.45023467805650497
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,512,256,0.007096000015735626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,512,256,0.02907466557290819
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,512,3072,0.34557512071397567
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,512,128,0.005945777727497949
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,512,128,0.021687999367713928
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,512,2560,0.2923751142289904
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,512,64,0.0052560000783867305
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,512,64,0.01905777719285753
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,512,32,0.005551110953092575
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,512,32,0.017688888642523024
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,512,1536,0.18627644909752739
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,256,65536,0.3233848942650689
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,512,2048,0.239848878648546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,256,16384,0.08276710907618205
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,256,16384,1.2233218087090385
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,256,65536,4.838024139404297
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,256,12288,0.06863466898600261
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,256,12288,0.9231075710720487
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,512,512,0.0768266651365492
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,512,768,0.10263999965455796
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,256,10240,0.05441955394215054
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,512,1024,0.13000978363884821
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,256,10240,0.7718622419569227
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,256,8192,0.044032000833087497
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,512,256,0.048560000128216214
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,256,8192,0.6216977967156304
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,256,7168,0.03943999939494663
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,512,128,0.03650044401486715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,256,6144,0.034753776258892484
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,256,7168,0.5463431146409776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,256,6144,0.47125154071384007
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,256,5120,0.03026133444574144
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,256,5120,0.3870986567603217
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,256,16384,1.6681564119127061
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,256,4096,0.025370667378107708
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,256,65536,6.588879903157552
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,256,4096,0.3129768901401096
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,256,12288,1.2570986217922635
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,256,3584,0.022781333989567224
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,256,3584,0.27693955103556317
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,256,3072,0.020251555575264823
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,256,3072,0.23944534195793998
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,256,8192,0.8497164514329699
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,256,10240,1.0522080527411568
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,256,2560,0.01608088943693373
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,256,2560,0.20263911618126762
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,256,2048,0.013077333569526672
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,256,7168,0.7460524241129557
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,256,2048,0.16525599691602919
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,256,1536,0.010703999963071613
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,256,6144,0.6441182030571831
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,256,1536,0.12767022185855442
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,256,1024,0.00833777752187517
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,256,1024,0.08812799718644883
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,256,5120,0.5420284271240234
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,256,768,0.007229333122571309
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,256,768,0.06589511368009779
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,256,3584,0.3883795473310683
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,256,4096,0.4394497871398926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,256,512,0.006079111248254776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,256,512,0.04657511247528923
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,256,256,0.004928888960017098
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,256,256,0.02755466600259145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,256,3072,0.3373066584269206
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,256,128,0.004313777718279096
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,256,128,0.02036888897418976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,256,2560,0.28522756364610463
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,256,64,0.003927111211750242
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,256,64,0.017477333545684814
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,256,32,0.004297777596447202
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,256,32,0.016012445092201233
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,256,2048,0.2331857681274414
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,256,1536,0.1813786692089505
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,128,65536,0.284954653845893
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,256,768,0.09945066769917806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,128,16384,0.0689528915617201
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,256,1024,0.12735111183590359
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,128,12288,0.05348799957169426
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,128,16384,1.2237262725830078
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,128,12288,0.9226915571424695
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,128,10240,0.045347554816140064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,128,10240,0.7722435527377658
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,128,65536,4.833932240804036
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,128,8192,0.03785689009560479
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,128,8192,0.6213671366373698
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,256,256,0.04706311225891113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,128,7168,0.0328382220533159
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,256,512,0.07454399930106269
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,256,128,0.03474400109714932
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,128,6144,0.028669334120220605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,128,7168,0.5461422072516547
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,128,6144,0.4711502393086751
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,128,5120,0.025337777203983728
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,128,5120,0.38571733898586696
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,128,8192,0.8407635688781738
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,128,4096,0.020962667134073045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,128,12288,1.2452195485432942
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,128,4096,0.31254222657945424
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,128,3584,0.01863555610179901
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,128,3584,0.27585334248012966
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,128,7168,0.7386648919847277
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,128,3072,0.015407999356587728
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,128,3072,0.2388897736867269
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,128,6144,0.6367688708835179
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,128,2560,0.010199999643696679
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,128,2560,0.20127911037868926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,128,16384,1.6496035257975261
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,128,10240,1.0421582327948677
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,128,2048,0.008348444269763099
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,128,5120,0.5364329020182291
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,128,2048,0.16416445043351915
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,128,1536,0.007149332927332983
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,128,1536,0.12636355559031168
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,128,1024,0.005955555372767978
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,128,1024,0.08705688847435845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,128,65536,6.515244377983941
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,128,768,0.005335999859703912
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,128,768,0.06567111280229357
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,128,4096,0.434728013144599
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,128,512,0.0046746668716271715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,128,3584,0.3843119939168294
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,128,512,0.046480000019073486
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,128,256,0.003959111041492886
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,128,256,0.027314666244718764
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,128,3072,0.3333591090308295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,128,128,0.003624000069167879
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,128,2560,0.2826586564381917
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,128,128,0.019873778025309246
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,128,64,0.00327822214199437
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,128,64,0.016866667403115165
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,128,32,0.0035991109907627106
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,128,32,0.015412444869677225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,128,2048,0.23088177045186362
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,64,65536,0.26850665940178764
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,128,1536,0.17984533309936523
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,64,16384,0.07144177622265287
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,64,12288,0.052766223748524986
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,128,1024,0.1259217792087131
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,64,10240,0.044179555442598134
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,64,12288,0.922885364956326
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,64,16384,1.2237813737657335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,128,768,0.0985928906334771
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,64,8192,0.037645333343082
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,64,7168,0.03288266725010342
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,64,6144,0.028985778490702312
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,64,10240,0.7720924483405219
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,64,5120,0.023881778120994568
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,64,8192,0.6217599974738227
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,64,7168,0.5467759768168131
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,64,4096,0.02074399921629164
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,64,3584,0.018780445059140522
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,64,6144,0.4709635310702854
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,64,5120,0.38716888427734375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,64,3072,0.013660444153679742
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,64,4096,0.3143217828538683
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,64,2560,0.010500444306267632
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,64,3584,0.2771600087483724
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,64,2048,0.009411555197503831
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,64,3072,0.2400240103403727
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,64,1536,0.007430222299363878
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,64,2560,0.20289599895477295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,64,1024,0.006008000009589725
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,64,2048,0.16510755485958523
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,64,768,0.0053173332578606075
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,64,1536,0.12715289327833387
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,64,1024,0.08793599738015069
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,64,768,0.06524799929724799
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,64,512,0.00453599997692638
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,64,512,0.046295109722349376
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,64,256,0.003914666672547658
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,64,256,0.027261333333121404
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,64,128,0.003488000068399641
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,64,64,0.0032293332947625052
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,64,128,0.019868445065286424
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,64,32,0.0034168888297345904
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,64,64,0.016915554801623028
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,64,32,0.015386665860811869
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,32,16384,0.0691831111907959
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,64,65536,4.835193634033203
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,32,65536,0.25888265503777397
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,32,12288,0.054021331999037
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,32,10240,0.04239999916818407
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,32,8192,0.0347306662135654
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,32,10240,0.7721101972791883
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,32,16384,1.2229475445217557
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,32,12288,0.9221270879109701
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,32,7168,0.03086844417783949
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,32,6144,0.02702666653527154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,32,8192,0.6214079856872559
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,32,5120,0.023242667317390442
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,32,7168,0.5466017723083496
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,32,4096,0.02158044444190131
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,32,6144,0.47165510389539933
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,32,3584,0.017634666628307767
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,32,5120,0.38943910598754883
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,32,3072,0.012485333614879184
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,32,4096,0.3150000042385525
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,32,2560,0.009224889179070791
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,32,3584,0.27756089634365505
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,32,2048,0.007983999947706858
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,32,3072,0.2409271134270562
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,32,2560,0.20407022370232475
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,32,1536,0.006654222392373615
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,32,2048,0.16507821612887913
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,32,1024,0.0052977779673205475
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,32,1536,0.1275751060909695
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,32,768,0.004765333400832282
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,32,1024,0.08796355459425186
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,32,512,0.0041804446114434134
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,32,768,0.06514844629499647
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,32,256,0.003543111185232798
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,32,512,0.04645955562591553
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,32,128,0.0032106666929192017
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,32,256,0.027266666293144226
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,32,64,0.0029884444342719186
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,32,128,0.019916445016860962
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8192,32,32,0.0031022222505675424
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,32,64,0.016751999656359356
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,32,32,0.015320888823933072
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8192,32,65536,4.833597395155165
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,128,256,0.047372443808449634
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,128,512,0.07328889105055067
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,65536,12288,4.617615169949002
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,65536,16384,8.062288920084635
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,65536,12288,5.782697041829427
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,65536,16384,7.647336324055989
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,65536,10240,3.9281387329101562
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,65536,10240,6.959787580702041
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,65536,12288,8.325892978244358
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,65536,8192,3.1947244008382163
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,65536,10240,4.830590142144097
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,65536,16384,11.173407660590279
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,65536,8192,3.8612471686469187
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,65536,8192,5.589322831895616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,65536,7168,2.741074668036567
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,65536,6144,2.244693332248264
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,65536,7168,3.442381329006619
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8192,128,128,0.03407111101680332
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,65536,6144,4.264151255289714
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,65536,6144,2.8543679979112415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,65536,5120,1.9502230750189886
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,65536,4096,1.5351990593804254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,65536,5120,2.317091623942057
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,65536,7168,4.8926311069064665
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,65536,5120,3.5426161024305554
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,65536,4096,2.8524400922987194
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,65536,3584,1.3922995461357965
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,65536,4096,1.880782233344184
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,65536,3584,1.6557679706149633
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,65536,3072,1.1756782531738281
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,65536,3072,1.3598249223497179
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,65536,3584,2.498509301079644
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,65536,3072,2.132474687364366
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,65536,2560,0.9732773039076063
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,65536,2560,1.8197449578179254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,65536,2560,1.1686257256401908
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,65536,2048,0.7804133097330729
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,65536,1536,0.5779866642422146
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,65536,1024,0.39688889185587567
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,65536,2048,0.932463116115994
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,65536,2048,1.4276559617784288
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,65536,1024,0.49435287051730686
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,65536,1536,1.0625662273830838
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,65536,1024,0.7015288670857748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,65536,1536,0.7063280211554633
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,65536,768,0.3303626643286811
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,65536,768,0.5505466461181641
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,65536,256,0.22231109937032065
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,65536,768,0.38983554310268825
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,65536,256,0.23345422744750977
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,65536,128,0.19867022832234701
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,65536,512,0.2692648834652371
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,65536,256,0.26251912117004395
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,65536,512,0.4036862320370144
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,65536,128,0.19835733042822945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,65536,512,0.29540978537665474
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,65536,64,0.17654311656951904
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,65536,128,0.21774400605095756
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,65536,32,0.18138310644361708
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,65536,64,0.2080986632241143
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,65536,32,0.1997413370344374
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,16384,12288,1.4784862730238173
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,16384,16384,2.2738115522596574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,16384,12288,2.148576100667318
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,16384,16384,2.3818507724338107
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,16384,12288,1.8020498487684462
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,16384,16384,2.911887062920464
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,16384,10240,1.2515413496229384
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,16384,8192,1.0292541715833876
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,16384,10240,1.4906756083170574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,16384,10240,1.8123448689778645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,16384,8192,1.4812035030788844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,16384,8192,1.1810826195610893
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,16384,7168,0.885350227355957
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,16384,7168,1.2559270858764648
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,16384,7168,1.0383066601223416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,16384,6144,1.049905776977539
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,16384,6144,0.7547626495361328
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,16384,5120,0.6329341994391547
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,16384,6144,0.8970648447672526
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,16384,5120,0.7879946496751574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,16384,4096,0.5036426650153266
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,16384,5120,0.7400471369425455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,16384,4096,0.6068968772888184
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,16384,4096,0.65829955206977
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,16384,3584,0.4528809123569065
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,16384,3584,0.5666933589511448
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,16384,3584,0.5301271014743382
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,16384,3072,0.48694668875800234
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,16384,3072,0.3854657808939616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,16384,3072,0.4589262008666992
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,16384,2560,0.32093689176771373
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,16384,2560,0.410386668311225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,16384,2560,0.3954933219485813
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,16384,2048,0.2584044403500027
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,16384,1536,0.19073688983917236
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,16384,2048,0.320287995868259
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,16384,2048,0.3320346673329671
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,16384,1536,0.2590755621592204
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,16384,1536,0.2508968777126736
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,16384,1024,0.13387110498216417
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,16384,768,0.1398586697048611
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,16384,768,0.11044977770911323
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,16384,1024,0.18517600165473092
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,16384,1024,0.1803333361943563
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,16384,512,0.10359644227557713
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,16384,768,0.14849421713087294
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,16384,256,0.06348355611165364
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,16384,512,0.09181155761082967
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,16384,512,0.11248177952236599
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,16384,256,0.07338577508926392
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,16384,128,0.04944533440801832
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,16384,256,0.08189866940180461
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,16384,128,0.0650142232577006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,16384,64,0.04553066690762838
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,16384,128,0.07442577679951985
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,16384,64,0.06624178091684978
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,16384,32,0.04645333356327481
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,16384,32,0.06366044282913208
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,16384,65536,9.65065171983507
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,16384,65536,11.340293036566841
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,12288,16384,1.8098924424913194
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,12288,16384,1.9939759572347004
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,12288,16384,2.045743942260742
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,12288,12288,1.2315235137939453
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,12288,12288,1.5101120207044814
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,12288,10240,1.0390524334377713
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,12288,12288,1.4790692859225805
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,12288,10240,1.1313787036471896
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,12288,10240,1.2374596065945096
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,16384,65536,11.425686306423612
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,12288,8192,0.8438444667392306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,12288,8192,0.9405440224541558
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,12288,8192,0.9832728703816732
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,12288,7168,0.7434115409851074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,12288,6144,0.6313778029547797
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,12288,7168,0.8716826438903809
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,12288,7168,0.8341297573513455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,12288,6144,0.7092249128553602
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,12288,6144,0.7429466777377658
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,12288,5120,0.5308808750576443
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,12288,5120,0.625274658203125
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,12288,5120,0.6014755566914877
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,12288,4096,0.420687993367513
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,12288,4096,0.4654124577840169
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,12288,4096,0.5103635787963867
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,12288,3584,0.38014311260647243
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,12288,3584,0.4496889114379883
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,12288,3072,0.3235795497894287
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,12288,3584,0.41008888350592715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,12288,3072,0.3547351095411513
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,12288,65536,8.119178771972656
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,12288,3072,0.3894684314727783
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,12288,2560,0.2693022092183431
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,12288,2560,0.2994462119208442
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,12288,2560,0.3332284556494819
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,12288,1536,0.18746044900682238
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,12288,2048,0.24363822407192656
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,12288,1536,0.1607982185151842
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,12288,65536,8.90240732828776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,12288,1536,0.21327466434902617
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,12288,2048,0.27256888813442653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,12288,2048,0.2170506715774536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,12288,1024,0.13309955596923828
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,12288,1024,0.11339022053612603
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,12288,768,0.09473599990208943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,12288,768,0.10625511407852173
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,12288,1024,0.15563289324442545
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,12288,256,0.04956444435649448
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,12288,512,0.07849688662423028
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,12288,512,0.07685333490371704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,12288,768,0.1266382270389133
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,12288,256,0.06186844242943657
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,12288,512,0.09747288624445598
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,12288,128,0.038520889149771795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,12288,128,0.05461688836415609
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,12288,256,0.07094844182332356
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,12288,64,0.03541333476702372
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,12288,128,0.0634711119863722
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,12288,64,0.05470488799942864
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,12288,32,0.03654933306905959
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,12288,32,0.05369244350327385
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,10240,16384,1.5807671017116969
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,10240,16384,1.7643778059217665
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,12288,65536,8.102139790852865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,10240,12288,1.2395226160685222
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,10240,12288,1.0969599617852104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,10240,16384,1.8087963528103297
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,10240,12288,1.3228275510999892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,10240,10240,0.9528391096327039
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,10240,10240,0.9273591571384006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,10240,10240,1.102494239807129
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,10240,8192,0.7560959921942817
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,10240,8192,0.8305902481079102
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,10240,8192,0.887939559088813
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,10240,7168,0.6608195304870605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,10240,7168,0.7017928759256998
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,10240,7168,0.7797084384494357
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,10240,6144,0.5681697527567545
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,10240,6144,0.6757626533508301
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,10240,5120,0.47426843643188477
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,10240,6144,0.5802808867560493
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,10240,5120,0.48386043972439235
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,10240,5120,0.5633653534783257
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,10240,4096,0.37823822763231063
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,10240,4096,0.39152712292141384
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,10240,4096,0.46292712953355575
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,10240,3584,0.3467893335554335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,10240,3584,0.34285688400268555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,10240,3072,0.29001956515842015
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,10240,3584,0.40737332238091367
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,10240,3072,0.35728711552090114
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,10240,65536,7.729487948947483
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,10240,2560,0.24311910735236275
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,10240,3072,0.29951379034254283
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,10240,2560,0.30314932929144967
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,10240,2560,0.2516968780093723
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,10240,2048,0.19582577546437582
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,10240,2048,0.24922222561306426
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,10240,1536,0.1610648896959093
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,10240,2048,0.20596088303460014
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,10240,1536,0.14511111047532824
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,10240,1024,0.11393688784705268
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,10240,1024,0.10325688785976833
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,10240,1536,0.1920542187160916
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,10240,768,0.09209777911504109
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,10240,1024,0.14235377311706543
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,10240,512,0.06955822308858235
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,10240,768,0.08581688668992783
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,10240,512,0.06907111406326294
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,10240,768,0.1144791046778361
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,10240,256,0.04612888892491659
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,10240,512,0.08942933215035333
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,10240,256,0.05532088875770569
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,10240,128,0.03357777661747403
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,10240,65536,7.1912443372938375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,10240,256,0.06562933656904432
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,10240,128,0.04748622245258755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,10240,64,0.03013066781891717
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,10240,128,0.059225777784983315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,10240,64,0.04846400022506714
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,10240,32,0.03113422128889296
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,10240,32,0.048807998498280845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,8192,16384,1.3633217281765406
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,8192,12288,0.9951644473605685
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,8192,16384,1.5501217312282984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,8192,16384,1.2135581970214844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,8192,12288,0.9709261788262261
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,8192,12288,1.1711191601223416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,8192,10240,0.7743271191914877
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,8192,10240,0.8167866600884331
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,8192,8192,0.6750204298231336
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,8192,8192,0.7998933262295194
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,8192,8192,0.6066213183932835
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,8192,10240,0.9821039835611979
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,8192,7168,0.5816595819261339
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,8192,7168,0.5351191096835667
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,10240,65536,7.027920193142361
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,8192,7168,0.6978915532430013
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,8192,6144,0.4599146842956543
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,8192,6144,0.49648888905843097
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,8192,5120,0.4174506664276123
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,8192,6144,0.6006008783976237
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,8192,5120,0.383208884133233
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,8192,5120,0.5109226438734267
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,8192,4096,0.3329182200961643
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,8192,65536,6.47747802734375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,8192,4096,0.3092835479312473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,8192,3584,0.30035199059380424
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,8192,4096,0.4136551221211751
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,8192,3584,0.3716124428643121
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,8192,3072,0.25567732916937935
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,8192,3072,0.23765601052178276
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,8192,3584,0.27581689092848033
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,8192,2560,0.21401333808898926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,8192,3072,0.3192373381720649
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,8192,2560,0.273796452416314
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,8192,2048,0.17331822713216147
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,8192,2560,0.1987137794494629
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,8192,65536,6.427451663547092
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,8192,2048,0.2232417795393202
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,8192,1536,0.12809422281053331
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,8192,1536,0.12791644202338323
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,8192,2048,0.1661075618531969
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,8192,1024,0.10096266534593369
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,8192,1536,0.17314044634501138
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,8192,1024,0.09086489015155369
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,8192,768,0.07295999924341838
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,8192,768,0.07574222485224406
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,8192,1024,0.1287964449988471
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,8192,512,0.054225779241985746
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,8192,512,0.06139466497633192
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,8192,256,0.035551110903422035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,8192,768,0.10616888602574666
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,8192,512,0.08084266715579562
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,8192,256,0.04758844441837735
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,8192,128,0.027878221538331773
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,8192,128,0.04191822144720289
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,8192,256,0.05988799863391452
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,8192,64,0.025109334124459162
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,8192,128,0.05268355541759067
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,8192,32,0.025920889443821375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,8192,64,0.041505777173572116
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,8192,32,0.0410497784614563
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,7168,16384,1.2637359831068251
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,8192,65536,5.408509148491754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,7168,16384,1.4618523915608723
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,7168,12288,0.8281964196099175
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,7168,16384,1.1777324676513672
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,7168,12288,0.8946630689832898
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,7168,12288,1.1062391069200304
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,7168,10240,0.7625981966654459
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,7168,10240,0.6832044389512805
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,7168,8192,0.6177466710408529
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,7168,8192,0.5466497739156088
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,7168,10240,0.9250870810614692
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,7168,8192,0.7479795349968804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,7168,7168,0.4674808714124892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,7168,7168,0.5410710970560709
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,7168,7168,0.6610097885131836
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,7168,6144,0.46283912658691406
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,7168,6144,0.5697360038757324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,7168,6144,0.4066310988532172
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,7168,5120,0.3375733428531223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,7168,5120,0.39011645317077637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,7168,4096,0.3122097651163737
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,7168,4096,0.2744533220926921
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,7168,5120,0.4803911315070258
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,7168,3584,0.27949510680304634
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,7168,4096,0.3940266768137614
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,7168,3584,0.24100799030727812
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,7168,3072,0.23838933308919272
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,7168,3584,0.34912533230251735
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,7168,3072,0.30376354853312176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,7168,2560,0.20032178031073677
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,7168,65536,5.948519812689887
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,7168,3072,0.20697600311703154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,7168,2048,0.14547022183736166
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,7168,2560,0.25869690047370064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,7168,2560,0.1761946678161621
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,7168,2048,0.1617928875817193
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,7168,1536,0.11198755105336507
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,7168,1536,0.12036000357733832
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,7168,1024,0.08099822203318278
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,7168,1024,0.08617777956856622
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,7168,1536,0.16550044218699136
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,7168,2048,0.21033688386281332
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,7168,768,0.06528533167309232
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,7168,1024,0.12096533510420059
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,7168,768,0.0721635553571913
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,7168,768,0.09989066918690999
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,7168,512,0.04887466629346212
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,7168,512,0.057328886455959745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,7168,256,0.03189333279927572
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,7168,512,0.07826044162114461
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,7168,256,0.04478489028082954
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,7168,128,0.02496799992190467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,7168,256,0.05806489123238457
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,7168,128,0.03868533174196879
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,7168,64,0.023258666197458904
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,7168,128,0.05004088746176826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,7168,32,0.024096000525686476
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,7168,64,0.038358221451441445
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,7168,65536,5.992130703396267
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,7168,32,0.03786399960517883
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,6144,16384,1.1538302103678386
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,7168,65536,4.724271138509114
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,6144,16384,0.9894079632229276
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,6144,16384,1.3674408594767253
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,6144,12288,0.8374115626017252
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,6144,10240,0.7016533215840658
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,6144,12288,0.7026871045430502
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,6144,12288,1.0303235583835177
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,6144,10240,0.5689848793877496
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,6144,8192,0.5716995663113064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,6144,10240,0.8609093560112847
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,6144,8192,0.4516977734035916
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,6144,8192,0.7007395426432291
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,6144,7168,0.5002071062723795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,6144,7168,0.40220978524949813
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,6144,7168,0.6134933365715874
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,6144,6144,0.42771466573079425
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,6144,6144,0.34180710050794816
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,6144,5120,0.3636159896850586
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,6144,6144,0.5360293388366699
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,6144,5120,0.45025867886013454
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,6144,4096,0.29179467095269096
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,6144,4096,0.23824000358581543
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,6144,5120,0.29275110032823354
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,6144,4096,0.3669573201073541
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,6144,3584,0.2615164385901557
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,6144,65536,5.315255059136285
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,6144,3584,0.2073520024617513
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,6144,3072,0.22268711196051705
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,6144,3584,0.32763467894660103
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,6144,2560,0.15232000086042616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,6144,3072,0.2830142180124919
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,6144,2560,0.18661867247687447
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,6144,2048,0.12567910883161756
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,6144,3072,0.18121333916982016
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,6144,2560,0.2421484523349338
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,6144,1536,0.09734311368730332
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,6144,2048,0.15077333980136448
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,6144,2048,0.19787289036644828
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,6144,1536,0.11242132716708714
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,6144,1024,0.06920711199442546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,6144,1024,0.08034577634599474
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,6144,768,0.056576000319586865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,6144,1536,0.15573599603441027
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,6144,768,0.06728266345130073
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,6144,1024,0.11454399426778157
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,6144,512,0.04245333207978142
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,6144,512,0.053596443600124784
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,6144,256,0.028297778632905748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,6144,768,0.09415644407272339
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,6144,256,0.0410844451851315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,6144,512,0.0737395551469591
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,6144,128,0.021198223034540813
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,6144,256,0.054614222711986966
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,6144,128,0.03582488828235202
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,6144,64,0.020025778147909377
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,6144,32,0.021300445000330608
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,6144,128,0.04730044470893013
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,6144,64,0.035391112168629967
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,6144,32,0.03481422199143304
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,6144,65536,5.517889658610026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,6144,65536,4.05372068617079
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,5120,16384,1.0716489156087239
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,5120,16384,1.270674705505371
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,5120,16384,0.7938897874620227
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,5120,12288,0.5706408818562826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,5120,12288,0.7673981984456381
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,5120,10240,0.649594677819146
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,5120,10240,0.47656530804104275
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,5120,12288,0.9621875551011828
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,5120,8192,0.38328178723653156
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,5120,8192,0.5260515742831761
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,5120,10240,0.8179795477125379
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,5120,8192,0.6567138036092123
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,5120,7168,0.460918214586046
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,5120,7168,0.3373573356204563
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,5120,6144,0.39521156416998965
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,5120,7168,0.5808017518785265
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,5120,6144,0.29135555691189235
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,5120,6144,0.49998044967651367
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,5120,5120,0.3332880073123508
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,5120,5120,0.4223804473876953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,5120,5120,0.24811911582946777
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,5120,4096,0.26694756084018284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,5120,4096,0.34558312098185223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,5120,4096,0.19900443818834093
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,5120,3584,0.24165333641899955
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,5120,3072,0.1535866657892863
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,5120,3584,0.17447022596995035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,5120,3584,0.3062551021575928
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,5120,3072,0.20504266685909697
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,5120,2560,0.13072621822357178
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,5120,65536,4.76157972547743
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,5120,3072,0.2686702145470513
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,5120,2560,0.17195733388264975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,5120,2048,0.1069271100891961
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,5120,2560,0.2275235652923584
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,5120,2048,0.14002222485012478
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,5120,1536,0.08510844575034247
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,5120,1536,0.10441511207156712
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,5120,1536,0.14628622266981336
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,5120,1024,0.07521688938140869
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,5120,1024,0.06090311209360758
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,5120,2048,0.1873475578096178
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,5120,768,0.04954044355286492
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,5120,1024,0.10774577988518609
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,5120,768,0.06245866749021742
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,5120,512,0.037462221251593694
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,5120,512,0.05005422234535217
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,5120,768,0.08806222014957005
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,5120,256,0.025958221819665697
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,5120,128,0.019460444649060566
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,5120,512,0.06990577777226765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,5120,65536,5.0386373731825085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,5120,256,0.03773066732618544
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,5120,256,0.0528657767507765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,5120,128,0.0328746669822269
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,5120,64,0.016622222132152982
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,5120,128,0.04452711012628344
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,5120,32,0.016866667403115165
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,5120,64,0.03223911258909438
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,5120,32,0.031515555249320135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,5120,65536,3.5647180345323353
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,4096,16384,0.9412142435709635
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,4096,16384,0.629276434580485
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,4096,12288,0.4501804245842828
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,4096,12288,0.6911724408467611
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,4096,16384,1.1814532809787326
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,4096,10240,0.5778657595316569
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,4096,12288,0.8886231316460503
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,4096,10240,0.37733689943949383
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,4096,10240,0.7532595528496636
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,4096,8192,0.3019875685373942
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,4096,8192,0.47420623567369247
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,4096,7168,0.40877511766221786
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,4096,7168,0.2692791091071235
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,4096,8192,0.6057324409484863
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,4096,7168,0.5354035695393881
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,4096,6144,0.3480195469326443
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,4096,5120,0.2904062271118164
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,4096,6144,0.4660951296488444
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,4096,6144,0.23367643356323242
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,4096,5120,0.193359997537401
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,4096,4096,0.23454133669535318
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,4096,4096,0.1575040022532145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,4096,5120,0.3938853210873074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,4096,3584,0.13927911387549505
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,4096,4096,0.3225626680586073
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,4096,3584,0.2072568866941664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,4096,3072,0.12211467160118951
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,4096,65536,3.8672167460123696
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,4096,2560,0.10405777560340033
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,4096,3584,0.28509423467848033
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,4096,2560,0.15151111284891763
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,4096,3072,0.17963200145297578
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,4096,2048,0.08558133575651382
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,4096,1536,0.06697955396440294
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,4096,2560,0.21204622586568198
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,4096,2048,0.1227475537194146
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,4096,1536,0.09309866693284775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,4096,2048,0.1748880015479194
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,4096,1024,0.04849955439567566
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,4096,1536,0.13565954897138807
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,4096,3072,0.24813689125908744
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,4096,1024,0.06819644239213732
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,4096,768,0.03962666789690653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,4096,1024,0.10050044457117717
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,4096,768,0.05632444222768148
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,4096,512,0.030223111311594646
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,4096,768,0.08319022258122762
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,4096,256,0.020915556285116408
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,4096,512,0.04543377624617683
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,4096,512,0.06620888577567206
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,4096,256,0.03378933336999681
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,4096,128,0.015712888704405892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,4096,256,0.04914755622545878
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,4096,65536,4.729894426133898
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,4096,128,0.029020445214377508
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,4096,32,0.014366222752465142
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,4096,64,0.013403555585278405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,4096,64,0.028315554062525432
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,4096,32,0.02721600068940057
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,4096,128,0.04130666785769992
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3584,16384,0.8891981972588433
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,4096,65536,2.67885250515408
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3584,12288,0.4237031141916911
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3584,16384,0.6193857722812229
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3584,12288,0.6557333204481337
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3584,10240,0.5512168672349718
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3584,10240,0.358463101916843
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3584,8192,0.45224089092678493
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3584,8192,0.2898888852861193
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3584,7168,0.25413865513271755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3584,65536,3.6268026563856335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3584,7168,0.3897955682542589
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3584,6144,0.21997511386871338
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3584,6144,0.3337599966261122
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3584,65536,2.6395929124620223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3584,5120,0.18651021851433647
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3584,5120,0.28064711888631183
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,3584,16384,1.1824568642510307
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3584,4096,0.1507528887854682
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3584,4096,0.22692889637417266
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,3584,12288,0.8958639568752713
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3584,3584,0.13315021991729736
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,3584,10240,0.7543759875827365
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3584,3584,0.20132177405887178
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3584,3072,0.11587022410498725
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,3584,8192,0.6096586651272243
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3584,3072,0.17426133155822754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3584,2560,0.09870311286714341
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3584,2560,0.14738133218553331
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,3584,7168,0.5379795498318142
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3584,2048,0.08138755295011732
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3584,2048,0.11832266383700901
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,3584,6144,0.46769242816501194
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3584,1536,0.06420889165666369
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,3584,65536,4.6450682746039496
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3584,1536,0.09074578020307754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3584,1024,0.04705066813362969
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3584,1024,0.06607821914884779
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,3584,5120,0.39380444420708555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3584,768,0.037905779149797224
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3584,768,0.05467999974886576
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,3584,3584,0.28737688064575195
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3584,512,0.022980444961124
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3584,512,0.04390044344796074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,3584,2560,0.2136088874604967
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3584,256,0.016728000508414376
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3584,256,0.032433778047561646
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,3584,4096,0.32197777430216473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3584,128,0.014386667145623101
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3584,128,0.027259555127885606
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,3584,3072,0.2488631142510308
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3584,64,0.012863111164834766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3584,64,0.02529511186811659
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3584,32,0.014003554979960123
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3584,32,0.024816888901922438
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,3584,1536,0.136062224706014
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,3584,2048,0.17425421873728433
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,3584,1024,0.10289333264033
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,3584,768,0.0832586685816447
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3072,65536,3.3940764533148875
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3072,16384,0.4456933339436849
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3072,16384,0.8475937843322754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,3072,16384,1.0875430636935763
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3072,65536,2.0216452280680337
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3072,12288,0.6285164621141222
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3072,12288,0.33956178029378253
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,3072,65536,4.2898915608723955
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3072,10240,0.5272275606791178
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,3072,12288,0.8225742446051704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,3072,10240,0.6985546747843424
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3072,8192,0.23016354772779676
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3072,10240,0.28504978285895455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3072,8192,0.4305111037360297
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3072,7168,0.3734924528333876
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,3072,8192,0.5607306692335341
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3072,7168,0.20275021923912895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,3072,7168,0.4975866741604275
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3072,6144,0.17416089110904268
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3072,5120,0.14820088280571833
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3072,6144,0.32042932510375977
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,3072,6144,0.4292417897118463
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3072,4096,0.1202968888812595
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3072,5120,0.26969867282443577
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3072,4096,0.21735644340515137
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,3072,5120,0.36321600278218585
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3072,3584,0.10648444626066421
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3072,3584,0.1944577826393975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3072,3072,0.09238488806618585
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,3072,4096,0.2971502145131429
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,3072,3584,0.26476089159647626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3072,3072,0.16735466321309408
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3072,2560,0.07857955826653375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3072,2048,0.06631466415193346
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,3072,3072,0.2311626540289985
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3072,2560,0.14173689153459337
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,3072,2560,0.1973279979493883
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3072,2048,0.11292000611623128
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3072,1536,0.052143110169304736
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3072,1536,0.08595289124382867
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3072,1024,0.03791822327507867
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,3072,2048,0.16239999400244817
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,3072,1536,0.12575377358330622
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3072,1024,0.0634933312733968
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3072,768,0.030847999784681533
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3072,768,0.052469333012898765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,3072,1024,0.09340622027715047
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3072,512,0.024193776978386775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,3072,768,0.07740088966157702
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3072,512,0.041722665230433144
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3072,256,0.016708445217874315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,3072,512,0.06225866741604275
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3072,128,0.012895111408498553
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3072,256,0.030765334765116375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3072,128,0.02477866742346022
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3072,64,0.010994666980372535
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,3072,256,0.045956446064843066
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,3072,32,0.011519999967681037
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,3072,128,0.037059555451075234
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3072,64,0.024248000648286607
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,3072,32,0.02313066687848833
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,3584,256,0.0498560004764133
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,3584,128,0.04191644324196709
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,3584,512,0.06655289067162408
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2560,16384,0.3770737648010254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2560,16384,0.8008969095018176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2560,12288,0.5933253500196669
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2560,12288,0.28260710504319936
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2560,10240,0.4990222189161513
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2560,10240,0.23891377449035645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2560,8192,0.4072737693786621
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2560,8192,0.19121155473921034
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2560,7168,0.16825332906511095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2560,65536,3.1982159084743924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2560,7168,0.35288532574971515
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2560,6144,0.14519643783569336
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2560,65536,1.8090008629692926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2560,6144,0.3025591108534071
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2560,5120,0.122505784034729
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2560,5120,0.25531111823187935
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,2560,16384,1.0953351126776802
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,2560,12288,0.8289475440979004
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2560,4096,0.10080977943208481
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2560,4096,0.20639999707539877
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2560,3584,0.0889964434835646
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2560,3584,0.18384267224205864
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,2560,10240,0.6965351104736328
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,2560,8192,0.565208911895752
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2560,3072,0.07686489158206515
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2560,3072,0.1585137711630927
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2560,2560,0.06612533330917358
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2560,2560,0.13456177711486816
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,2560,7168,0.49976444244384766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2560,2048,0.05439200003941854
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,2560,6144,0.4322151078118219
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2560,2048,0.10815733008914524
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2560,1536,0.04256622327698601
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2560,1536,0.0816391110420227
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,2560,65536,4.274013519287109
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,2560,5120,0.36584088537428117
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2560,1024,0.03086844417783949
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2560,1024,0.060416890515221484
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2560,768,0.024841777152485315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2560,768,0.04991733365588718
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,2560,3584,0.26574399736192494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2560,512,0.01926666663752662
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2560,512,0.039615111218558416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,2560,2560,0.1979102161195543
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2560,256,0.013175110850069257
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2560,256,0.02700533303949568
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,2560,4096,0.2992008792029487
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2560,128,0.010045333041085137
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2560,128,0.02289155622323354
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,2560,1536,0.1269813378651937
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2560,64,0.008624888956546783
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2560,64,0.021799999806616042
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2560,32,0.00904888908068339
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,2560,3072,0.23060443666246203
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2560,32,0.021791110436121624
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,2560,768,0.07911466889911227
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,2560,2048,0.1620159943898519
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2048,16384,0.29651467005411786
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2048,16384,0.7668702337476941
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2048,65536,3.071607165866428
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2048,65536,1.1932391060723198
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2048,12288,0.5660035345289443
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2048,12288,0.2244159910413954
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,2560,1024,0.0952133337656657
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,2048,16384,1.000716421339247
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,2048,65536,3.9264373779296875
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2048,10240,0.4741413328382704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2048,10240,0.18833511405520967
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,2048,12288,0.7608737945556641
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2048,8192,0.15182933542463514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,2048,10240,0.6359413464864095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2048,7168,0.13583111763000488
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2048,8192,0.3882693449656169
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2048,6144,0.11737422148386638
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,2048,8192,0.5186497900221083
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2048,7168,0.3369768990410699
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2048,6144,0.2887786759270562
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2048,5120,0.09839466545316909
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,2048,7168,0.4553057882520888
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2048,4096,0.07949066824383207
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2048,5120,0.24379287825690377
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,2048,6144,0.3971742259131537
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2048,4096,0.19708532757229277
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,2048,5120,0.334716452492608
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2048,3584,0.07083821958965726
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,2048,4096,0.27413511276245117
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2048,3584,0.17548978328704834
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2048,3072,0.061866667535569936
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2048,3072,0.1508880058924357
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,2048,3584,0.24283999866909453
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2048,2560,0.0532622238000234
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2048,2560,0.1282622284359402
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,2048,3072,0.2126248942481147
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2048,2048,0.043396443128585815
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2048,2048,0.10316889153586493
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,2048,2560,0.18181333276960585
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2048,1536,0.034246222840415105
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2048,1536,0.07739466428756714
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,2048,2048,0.14879111448923746
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2048,1024,0.025077333052953083
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,2048,1536,0.11592088805304633
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2048,1024,0.05585155884424845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2048,768,0.02036355601416694
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2048,768,0.04738222228156196
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,2048,1024,0.08590222067303127
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2048,512,0.015468445089128284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,2048,768,0.0718782212999132
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2048,512,0.03503288825352987
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2048,256,0.010844444235165914
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,2048,512,0.054049776660071484
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2048,256,0.025324443976084392
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2048,128,0.008314666648705801
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,2048,256,0.039270222187042236
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2048,128,0.02131822208563487
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2048,64,0.00720088928937912
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,2048,128,0.03352355625894334
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2048,64,0.020238222347365487
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,2048,32,0.007619555625650618
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,2048,32,0.019680000013775296
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1536,16384,0.22769067022535536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1536,16384,0.7248711056179471
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1536,12288,0.17175911532508004
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1536,65536,1.0543271170722115
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1536,12288,0.5384373135036892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1536,65536,2.8257448408338757
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1536,10240,0.14417866865793863
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1536,10240,0.45369423760308164
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1536,8192,0.11681421597798665
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,2560,256,0.043567111094792686
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,2560,128,0.03590933481852213
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1536,7168,0.10356089141633774
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,2560,512,0.06282489167319404
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1536,8192,0.36920976638793945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1536,6144,0.08891377846399943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1536,7168,0.3217617670694987
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1536,6144,0.2766835424635145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1536,5120,0.07521600193447538
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1536,5120,0.2329795625474718
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,1536,16384,0.9516151216295031
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1536,4096,0.06210933128992716
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1536,4096,0.1890337732103136
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,1536,12288,0.7216746542188855
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,1536,10240,0.6073013411627876
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1536,3584,0.054901334974500865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1536,3584,0.16801867220136854
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1536,3072,0.047799110412597656
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,1536,65536,3.7389068603515625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,1536,7168,0.43588532341851133
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1536,3072,0.145487109820048
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1536,2560,0.04084800018204583
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1536,2560,0.12254488468170166
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1536,2048,0.03362577822473314
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,1536,8192,0.4927440219455295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1536,2048,0.1007502211464776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,1536,6144,0.37841778331332737
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1536,1536,0.02684533264901903
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1536,1536,0.07586577865812513
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1536,1024,0.019839111301634047
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1536,1024,0.05396711164050632
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,1536,5120,0.3195528984069824
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1536,768,0.015923556354310777
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1536,768,0.04349066813786825
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,1536,768,0.06827021969689263
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1536,512,0.012431999875439538
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1536,512,0.03329066766632928
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,1536,3584,0.23237421777513292
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1536,256,0.008763555851247575
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1536,256,0.02383377816942003
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,1536,2560,0.17383821805318198
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1536,128,0.007095110913117726
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1536,128,0.019951999187469482
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,1536,1536,0.11139644516838922
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1536,64,0.00620177760720253
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1536,64,0.01901333365175459
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1536,32,0.0064426664676931165
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1536,32,0.01871200071440803
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,1536,4096,0.2617786725362142
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,1536,3072,0.20339732699924043
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,1536,2048,0.14225067032708064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1024,16384,0.1513777838812934
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1024,65536,0.6109297540452745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,1536,1024,0.08243022362391154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1024,12288,0.11489510536193848
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1024,16384,0.6769795417785645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1024,65536,2.6618870629204645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,1024,16384,0.9151200188530816
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1024,12288,0.5090426868862575
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1024,10240,0.09615466992060344
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,1024,12288,0.6904728677537707
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1024,8192,0.07839466465844049
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1024,10240,0.42793599764506024
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,1024,10240,0.5814240243699815
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1024,8192,0.34646932284037274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,1024,8192,0.4714248975118001
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,1024,65536,3.5571920606825085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1024,7168,0.06933422221077813
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1024,6144,0.06071644359164768
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1024,7168,0.30477955606248647
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1024,5120,0.05102577805519104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1024,6144,0.26409512095981175
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,1024,7168,0.4163164562649197
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,1024,6144,0.3632951047685411
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1024,4096,0.04188622368706597
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1024,5120,0.22159466478559706
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1024,3584,0.03720000055101182
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,1024,5120,0.3074951171875
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1024,4096,0.18083288934495714
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,1024,4096,0.2505564424726698
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1024,3072,0.0326186650329166
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1024,3584,0.15998400582207575
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1024,2560,0.027838223510318335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1024,3072,0.1395511097378201
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,1024,3584,0.22304889890882704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,1024,3072,0.1947235531277127
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1024,2048,0.02327644493844774
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1024,2560,0.11746133698357476
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,1024,2560,0.1664444473054674
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1024,2048,0.09583022197087605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1024,1536,0.018629334039158292
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,1024,2048,0.13611111376020643
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1024,1024,0.013808888693650564
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1024,1536,0.07211644119686551
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,1024,1536,0.10679733090930515
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1024,768,0.011384888655609555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1024,1024,0.051092445850372314
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,1024,1024,0.07964889208475749
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1024,512,0.009169778062237633
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1024,768,0.04128177629576789
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,1024,768,0.0658586687511868
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1024,256,0.00686488880051507
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1024,512,0.031219555271996394
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1024,256,0.022341334157519873
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1024,128,0.005724444571468566
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,1024,512,0.052236444420284696
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1024,64,0.005085333353943295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,1024,256,0.03708711266517639
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1024,128,0.018535999788178336
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,1024,32,0.005370666583379109
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,1024,128,0.031150221824645996
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1024,64,0.01740444368786282
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,1024,32,0.016950221525298226
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,768,16384,0.1152542167239719
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,768,16384,0.6705697907341851
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,768,12288,0.08865511417388916
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,768,16384,0.8861458036634656
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,768,65536,0.4495120048522949
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,768,12288,0.504902203877767
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,768,10240,0.07471822367774116
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,768,12288,0.6723262468973795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,768,10240,0.42408000098334414
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,768,8192,0.060347557067871094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,768,10240,0.56568357679579
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,768,8192,0.3447271188100179
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,768,7168,0.053695112466812134
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,768,65536,2.6206044091118708
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,768,8192,0.4595502217610677
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,768,7168,0.30267466439141166
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,768,6144,0.047078222036361694
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,768,7168,0.4063599904378255
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,768,5120,0.040101332796944514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,768,6144,0.261208004421658
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,768,4096,0.03261422117551168
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,768,5120,0.22042844030592176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,768,6144,0.3531484338972304
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,768,5120,0.2989119953579373
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,768,3584,0.029121776421864826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,768,4096,0.179185774591234
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,768,65536,3.464376025729709
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,768,4096,0.24411021338568795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,768,3072,0.025557332568698462
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,768,3584,0.15928267108069524
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,768,2560,0.022135999467637803
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,768,3584,0.21730044153001574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,768,3072,0.13866489463382298
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,768,2560,0.11688000626034206
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,768,2048,0.018414222531848483
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,768,3072,0.18958400355445015
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,768,2560,0.1616951094733344
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,768,1536,0.0147706667582194
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,768,2048,0.09522310892740886
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,768,1024,0.011005333728260465
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,768,1536,0.07149955299165514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,768,2048,0.13294577598571777
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,768,1536,0.10394222206539577
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,768,768,0.009369778136412302
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,768,1024,0.05029066734843784
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,768,1024,0.07759466436174181
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,768,512,0.0075662218862109715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,768,768,0.04085866610209147
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,768,768,0.06455644634034899
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,768,256,0.0058444444504049085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,768,512,0.031011554929945204
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,768,512,0.049351112710105054
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,768,256,0.02235377828280131
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,768,128,0.00498933345079422
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,768,256,0.036455111371146307
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,768,64,0.004464000049564573
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,768,128,0.018580444984965853
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,768,32,0.0047280001971456735
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,768,128,0.030294222964180842
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,768,64,0.01737600068251292
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,768,32,0.016812443733215332
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,512,16384,0.07801422145631579
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,512,16384,0.6329991022745768
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,512,65536,0.3080693350897895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,512,12288,0.06170577473110623
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,1536,128,0.03184088733461168
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,512,12288,0.4767635663350423
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,512,10240,0.050916443268458046
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,512,10240,0.40004976590474445
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,512,65536,2.4662693871392145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,1536,256,0.03807288739416335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,512,8192,0.04212177793184916
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,512,7168,0.037074665228525795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,512,8192,0.3252320024702284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,512,7168,0.28616266780429417
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,512,6144,0.03228622343805101
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,512,6144,0.24764977561102974
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,1536,512,0.05372800098525154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,512,5120,0.027552000350422327
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,512,5120,0.20879111025068495
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,512,12288,0.6584595574273003
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,512,16384,0.8683662414550781
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,512,4096,0.02308799988693661
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,512,4096,0.1705564392937554
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,512,3584,0.020450666546821594
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,512,3584,0.1516773303349813
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,512,10240,0.5535590913560655
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,512,3072,0.017815111411942374
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,512,3072,0.13087822331322566
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,512,65536,3.3893138037787547
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,512,8192,0.45082844628228086
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,512,2560,0.01571822166442871
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,512,7168,0.3970506721072727
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,512,2560,0.11078311337365045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,512,2048,0.013106666505336761
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,512,2048,0.09075022406048244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,512,1536,0.010660444696744284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,512,6144,0.345778677198622
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,512,1536,0.06738666693369548
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,512,1024,0.008369777765538957
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,512,1024,0.04793688986036512
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,512,5120,0.29204710324605304
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,512,768,0.007175111108356052
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,512,768,0.03864533371395535
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,512,4096,0.23941511578030059
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,512,512,0.006088888893524806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,512,512,0.028692444165547688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,512,3584,0.21286933951907686
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,512,256,0.004978666702906291
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,512,256,0.02107111116250356
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,512,2560,0.15819556183285183
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,512,3072,0.1860711177190145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,512,128,0.00433955548538102
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,512,128,0.01721155477894677
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,512,64,0.003933333274390963
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,512,64,0.01586666703224182
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,512,32,0.0041573333243529005
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,512,2048,0.12961422072516546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,512,32,0.015494222442309061
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,512,1536,0.10164799955156113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,256,65536,0.16939377784729004
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,512,1024,0.0759448872672187
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,256,16384,0.051551110214657254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,256,16384,0.620831118689643
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,256,12288,0.0421324438518948
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,256,65536,2.42756716410319
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,256,12288,0.4704462157355414
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,256,10240,0.038350221183564924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,256,10240,0.39539111985100645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,512,768,0.06355200211207072
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,256,8192,0.034853332572513156
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,256,8192,0.31990400950113934
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,512,512,0.04944444364971585
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,512,256,0.035599999957614474
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,256,7168,0.028650667932298448
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,512,128,0.02955289019478692
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,256,7168,0.2826426559024387
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,256,6144,0.023904000719388325
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,256,6144,0.244741333855523
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,256,5120,0.0219217770629459
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,256,5120,0.2068693372938368
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,256,16384,0.852567990620931
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,256,4096,0.018618666463428073
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,256,4096,0.16839110851287842
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,256,12288,0.6464799775017632
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,256,3584,0.015955555770132277
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,256,65536,3.3312916225857205
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,256,3584,0.14893866909874812
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,256,3072,0.013562666873137156
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,256,3072,0.12949600484636095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,256,10240,0.5444382031758627
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,256,8192,0.44166845745510525
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,256,2560,0.012085333466529846
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,256,2560,0.10986577802234226
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,256,2048,0.009561777942710454
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,256,2048,0.08972533543904622
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,256,6144,0.33877333005269367
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,256,7168,0.390608893500434
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,256,1536,0.008371555142932469
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,256,1536,0.06674489047792223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,256,1024,0.006322666588756774
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,256,1024,0.0476382209195031
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,256,5120,0.28759288787841797
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,256,768,0.0058088890380329555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,256,768,0.03807199994723002
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,256,4096,0.2347431182861328
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,256,512,0.004805333498451444
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,256,512,0.028008888165156048
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,256,3584,0.20953422122531465
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,256,256,0.004383999854326248
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,256,256,0.020411555965741474
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,256,3072,0.18366311656104195
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,256,128,0.003637333297067218
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,256,128,0.016696888539526198
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,256,2560,0.15572444597880045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,256,64,0.0033937777496046494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,256,64,0.015303111738628812
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,256,32,0.003552888830502828
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,256,32,0.01475644442770216
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,128,65536,0.16698933972252739
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,256,2048,0.12783288955688477
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,128,65536,2.4262025621202254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,128,16384,0.04637066523234049
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,128,16384,0.6211271286010742
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,256,1536,0.09982755449083115
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,128,12288,0.03422310948371887
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,256,768,0.06202933523390028
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,256,1024,0.07462222046322294
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,128,12288,0.4703422122531467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,128,10240,0.030032889710532293
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,128,10240,0.39536354276869035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,128,8192,0.022914666268560622
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,128,8192,0.3199404345618354
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,256,256,0.03489244315359328
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,128,7168,0.019306666321224637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,256,128,0.028652442826165095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,256,512,0.04663911130693224
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,128,6144,0.016590222716331482
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,128,7168,0.2826320065392388
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,128,6144,0.2443466716342502
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,128,5120,0.011797332929240333
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,128,5120,0.20643288559383818
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,128,65536,3.314079920450846
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,128,4096,0.010118222071064843
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,128,4096,0.16840710904863146
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,128,16384,0.8490079773796929
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,128,3584,0.008903111020723978
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,128,3584,0.148890667491489
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,128,12288,0.643139574262831
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,128,3072,0.00852177788813909
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,128,10240,0.5411218007405599
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,128,3072,0.12947021590338811
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,128,2560,0.0076044441925154785
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,128,2560,0.10993422402275933
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,128,8192,0.43878756629096133
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,128,2048,0.006565333240562015
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,128,2048,0.08914489216274685
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,128,7168,0.38813066482543945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,128,6144,0.3363617791069879
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,128,1536,0.0058648888435628675
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,128,1536,0.06618666648864746
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,128,1024,0.004871110949251386
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,128,1024,0.04719466633266873
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,128,5120,0.2849057780371772
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,128,768,0.004486222234037188
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,128,768,0.037815110550986394
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,128,4096,0.23301598760816786
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,128,512,0.003996444245179494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,128,3584,0.20747999350229898
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,128,512,0.027862221002578735
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,128,256,0.003567999849716822
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,128,256,0.02010577751530541
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,128,3072,0.18097510602739122
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,128,128,0.00332266671790017
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,128,128,0.016531555189026725
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,128,2560,0.15453067090776232
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,128,64,0.003112000102798144
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,128,64,0.01499555508295695
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,128,32,0.0034017778105205963
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,128,2048,0.12655733691321477
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,128,32,0.014446222119861178
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,64,65536,0.14684266514248318
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,64,16384,0.04151022103097703
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,64,16384,0.6204426553514268
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,64,12288,0.032835556401146784
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,64,12288,0.4703013102213542
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,64,10240,0.028675556182861328
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,64,65536,2.4276373121473527
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,64,8192,0.025216889050271776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,64,10240,0.395069334242079
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,64,7168,0.018418666389253404
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,64,8192,0.3201857672797309
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,128,1536,0.0987573332256741
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,128,1024,0.07377066877153185
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,64,6144,0.014304889572991265
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,64,5120,0.010605333579911126
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,128,768,0.06116888920466105
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,64,7168,0.2825617790222168
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,64,4096,0.008676444490750631
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,64,6144,0.24460887908935547
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,64,3584,0.008059555457697975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,64,5120,0.20680533515082467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,64,4096,0.16850845019022623
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,64,3072,0.007188444336255391
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,64,3584,0.14895999431610107
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,64,2560,0.0071928890215026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,64,3072,0.12923644648657903
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,64,2048,0.006205333189831839
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,64,2560,0.10965866512722439
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,64,1536,0.005603555589914322
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,64,2048,0.08869866530100505
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,64,1024,0.00462666650613149
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,64,1536,0.06626044379340278
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,64,768,0.004295111116435793
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,64,1024,0.04749600092569987
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,64,512,0.003839111162556542
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,64,768,0.037280887365341187
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,64,256,0.0034088889757792154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,64,512,0.02751022246148851
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,64,128,0.0030951110853089225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,64,256,0.020101333657900494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,64,64,0.002959111084540685
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,64,128,0.01650399963061015
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,64,64,0.01482311056719886
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,64,32,0.003044444446762403
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,64,32,0.01442399952146742
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,32,65536,0.13485066095987955
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,32,12288,0.030121776792738173
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,32,16384,0.037759111987219915
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,32,10240,0.02626577847533756
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,32,16384,0.6211475796169704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,32,12288,0.47065067291259766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,32,8192,0.023050665855407715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,32,10240,0.3949884573618571
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,32,7168,0.020283554991086323
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,32,8192,0.3199128839704725
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,32,6144,0.016688000824716356
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,32,5120,0.011668444507651858
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,32,7168,0.28217776616414386
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,32,4096,0.010565333068370819
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,32,6144,0.2446506553226047
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,32,5120,0.20661155382792154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,32,3584,0.009736888938479954
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,32,4096,0.16826844215393066
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,32,3072,0.008972444468074376
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,32,3584,0.14872977468702528
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,32,2560,0.006236444330877728
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,32,2048,0.005525333185990651
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,32,3072,0.1288426717122396
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,32,2560,0.1095306674639384
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,32,1536,0.004923555586073133
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,32,2048,0.08892622258928086
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,32,65536,2.4296934339735246
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,32,1024,0.004302222281694412
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,32,1536,0.0660604437192281
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,32,768,0.004018666843573253
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,32,1024,0.047284444173177086
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,32,512,0.003642666671011183
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,32,768,0.037905779149797224
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,32,256,0.003253333270549774
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,32,512,0.02754844393995073
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,32,128,0.0029760001020299066
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,32,256,0.020224000016848247
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,32,128,0.016454221473799813
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,32,64,0.0027946666296985415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4096,32,32,0.003018666679660479
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,32,64,0.014942222171359591
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4096,32,32,0.014507555299335055
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,128,256,0.034309334225124784
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,128,512,0.04604444570011563
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,65536,12288,2.217024061414931
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,65536,16384,3.2635519239637585
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,65536,10240,3.521201875474718
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,65536,10240,1.8436781565348308
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,65536,12288,4.170090569390191
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,65536,8192,1.5369804170396593
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,65536,16384,5.597735087076823
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,65536,8192,2.8240631951226125
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4096,128,128,0.028459555572933618
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,65536,7168,1.3246515062120225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,65536,12288,2.717719184027778
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,65536,7168,2.572725296020508
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,65536,16384,3.8068453470865884
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,65536,6144,1.1184853447808158
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,65536,5120,0.9638542599148221
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,65536,6144,2.105259577433268
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,65536,5120,1.6626702414618597
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,65536,10240,2.3181510501437717
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,65536,4096,0.7277164459228516
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,65536,4096,1.3420879575941298
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,65536,3584,1.0906204647488065
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,65536,3584,0.6809377670288086
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,65536,8192,1.8956498040093317
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,65536,3072,0.561606248219808
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,65536,3072,1.0172622468736436
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,65536,7168,1.611238267686632
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,65536,2560,0.45941776699490017
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,65536,2560,0.8007680045233833
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,65536,6144,1.360352940029568
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,65536,2048,0.3758400016360813
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,65536,5120,1.0981039471096463
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,65536,2048,0.6671866840786405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,65536,1536,0.49358309639824766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,65536,1536,0.2784302234649658
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,65536,1024,0.20783821741739908
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,65536,1024,0.34620266490512425
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,65536,768,0.2741875648498535
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,65536,768,0.1750568946202596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,65536,4096,0.9069990581936307
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,65536,3584,0.7926008966233996
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,65536,512,0.20453066296047637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,65536,512,0.14738933245340982
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,65536,256,0.13407199912601048
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,65536,256,0.12047199408213298
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,65536,2560,0.5833377838134766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,65536,128,0.10384177499347264
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,65536,128,0.11347999837663438
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,65536,3072,0.6761697663201226
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,65536,64,0.09560444619920518
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,65536,64,0.11728710598415798
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,65536,32,0.09424355294969346
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,65536,32,0.11192533704969619
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,65536,1536,0.366866668065389
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,65536,2048,0.4650835461086697
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,65536,256,0.13005956013997397
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,65536,512,0.16857777701483834
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,65536,768,0.21960711479187012
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,16384,16384,1.0136195288764107
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,65536,1024,0.26721689436170787
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,16384,16384,1.3355431026882594
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,16384,16384,1.1723581949869792
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,16384,12288,0.7339039908515083
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,65536,128,0.1256879965464274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,16384,12288,0.9957990646362305
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,16384,12288,0.8687351014879016
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,16384,10240,0.6215448909335667
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,16384,10240,0.7422080039978027
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,16384,10240,0.7474933200412326
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,16384,8192,0.5171359909905328
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,16384,8192,0.5932781961229112
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,16384,8192,0.6021004252963597
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,16384,7168,0.43582932154337567
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,16384,7168,0.5188097953796387
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,16384,7168,0.534005324045817
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,16384,6144,0.36566487948099774
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,16384,6144,0.45578575134277344
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,16384,6144,0.4599875344170465
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,16384,5120,0.31123020913865834
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,16384,5120,0.38712888293796116
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,16384,5120,0.3821537759568956
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,16384,4096,0.24251021279229057
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,16384,4096,0.3105164368947347
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,16384,4096,0.31274578306410045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,16384,3584,0.22043022844526503
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,16384,65536,4.86918470594618
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,16384,65536,4.765872955322266
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,16384,3584,0.2769662274254693
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,16384,3584,0.2754400041368273
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,16384,3072,0.18481600284576416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,16384,2560,0.15508088800642225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,16384,3072,0.240955564710829
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,16384,3072,0.2378071149190267
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,16384,2560,0.2068399985631307
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,16384,2048,0.12516889307234022
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,16384,2560,0.20127644803788927
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,16384,1536,0.10065778096516927
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,16384,2048,0.16502843962775335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,16384,1536,0.1397386656867133
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,16384,2048,0.1735964483684964
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,16384,1024,0.07479733228683472
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,16384,1536,0.1269706620110406
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,16384,768,0.07298577494091459
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,16384,1024,0.09045600228839451
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,16384,1024,0.10539199908574422
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,16384,768,0.0644435551431444
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,16384,512,0.054487999942567616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,16384,768,0.08760888708962335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,16384,512,0.05288088983959622
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,16384,256,0.03549333413441976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,16384,256,0.05613955524232653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,16384,512,0.07019378079308404
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,16384,256,0.04474666714668274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,16384,128,0.028143111202451918
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,16384,128,0.04078755444950528
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,16384,64,0.02510133385658264
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,16384,128,0.05060266786151462
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,16384,64,0.04089422358406915
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,16384,32,0.02625688910484314
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,16384,32,0.040989332728915744
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,12288,16384,0.838928911421034
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,12288,16384,0.9837217330932617
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,12288,12288,0.7302035755581326
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,12288,16384,0.9683467017279731
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,12288,12288,0.6178542243109809
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,12288,12288,0.7430355813768176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,12288,10240,0.5164648691813151
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,12288,10240,0.5741848945617676
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,12288,8192,0.4246479935116238
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,12288,10240,0.6263004408942329
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,12288,8192,0.4493733512030707
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,12288,8192,0.5009075800577799
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,12288,7168,0.36374489466349286
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,12288,7168,0.3996551036834717
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,12288,7168,0.4424115551842584
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,16384,65536,5.398924509684245
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,12288,65536,3.7341172960069446
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,12288,6144,0.3080684343973796
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,12288,6144,0.3811946709950765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,12288,6144,0.34365778499179417
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,12288,5120,0.2649831242031521
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,12288,5120,0.32477956347995335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,12288,4096,0.2039946715037028
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,12288,4096,0.2330231136745877
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,12288,5120,0.2913840081956652
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,12288,4096,0.26428090201483834
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,12288,3584,0.18461422125498453
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,12288,3072,0.1553573343488905
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,12288,3584,0.23314044210645887
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,12288,3072,0.2046133279800415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,12288,2560,0.1298853291405572
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,12288,3072,0.1787102222442627
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,12288,65536,3.994199964735243
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,12288,3584,0.20742489231957328
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,12288,2560,0.17568533950381807
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,12288,2048,0.1068826648924086
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,12288,2560,0.15269688765207926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,12288,1536,0.08460799853006999
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,12288,2048,0.14905867311689588
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,12288,1024,0.07033777899212308
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,12288,1024,0.06444088617960612
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,12288,1536,0.11979111035664876
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,12288,768,0.055925336149003774
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,12288,1024,0.09091999795701768
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,12288,2048,0.12473866674635146
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,12288,768,0.05567644702063667
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,12288,1536,0.09945511155658299
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,12288,512,0.04267377654711405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,12288,768,0.07693688737021552
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,12288,512,0.04531022244029575
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,12288,256,0.02792444494035509
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,12288,256,0.0380711125002967
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,12288,512,0.06259555286831327
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,12288,128,0.021627555290857952
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,12288,128,0.03435111045837402
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,12288,256,0.05141689048873054
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,12288,64,0.02030666669209798
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,12288,128,0.04514133267932468
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,12288,64,0.03453600075509813
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,12288,32,0.02219555609756046
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,12288,32,0.03407466742727492
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,10240,16384,0.7435759968227811
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,10240,12288,0.5775386492411295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,10240,16384,0.8840302361382378
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,10240,16384,0.7719848950703939
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,10240,12288,0.5444924566480849
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,10240,10240,0.4570142428080241
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,10240,12288,0.6646480030483669
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,10240,10240,0.5597350862291124
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,10240,10240,0.47629690170288086
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,10240,8192,0.381515555911594
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,10240,8192,0.3823395570119222
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,10240,7168,0.32843200365702313
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,10240,8192,0.455367989010281
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,10240,6144,0.27625777986314565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,10240,7168,0.4032284418741862
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,12288,65536,4.011803521050347
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,10240,7168,0.33782490094502765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,10240,65536,3.4246586693657766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,10240,6144,0.29155556360880536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,10240,5120,0.23811999956766763
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,10240,6144,0.348582214779324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,10240,4096,0.18407377931806776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,10240,5120,0.2962995635138618
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,10240,5120,0.24457067913479277
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,10240,4096,0.1990435520807902
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,10240,3584,0.16804533534579805
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,10240,3072,0.13908178276485866
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,10240,4096,0.2421466774410672
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,10240,65536,3.5325599246554904
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,10240,3584,0.1745377779006958
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,10240,3072,0.1521679957707723
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,10240,3584,0.2139599985546536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,10240,3072,0.18553688791063097
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,10240,2560,0.11727733082241482
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,10240,2048,0.09611555602815415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,10240,1536,0.08403555552164714
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,10240,2560,0.16460088888804117
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,10240,1536,0.07706933551364474
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,10240,2048,0.13630755742390951
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,10240,1024,0.060757332377963595
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,10240,2560,0.13028977976904974
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,10240,1024,0.059826665454440646
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,10240,1536,0.1109244426091512
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,10240,2048,0.1064079999923706
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,10240,768,0.049229333798090615
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,10240,768,0.050345778465270996
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,10240,1024,0.08417777882681952
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,10240,512,0.03762844536039565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,10240,256,0.025998221503363714
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,10240,768,0.07165422042210896
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,10240,512,0.04156622290611267
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,10240,512,0.058725330564710826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,10240,256,0.0348364445898268
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,10240,128,0.019627556204795837
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,10240,256,0.04763377706209818
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,10240,128,0.031439112292395696
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,10240,64,0.031096001466115315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,10240,64,0.016748444901572335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,10240,32,0.017651556266678702
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,10240,128,0.04274310999446445
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,10240,32,0.03129333257675171
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,8192,16384,0.6526062223646376
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,8192,16384,0.7962933116488986
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,8192,16384,0.6252480083041722
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,10240,65536,3.5198631286621094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,8192,12288,0.47846047083536786
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,8192,12288,0.45981423060099286
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,8192,12288,0.6022799809773763
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,8192,10240,0.39771199226379395
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,8192,10240,0.5016169018215603
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,8192,10240,0.38038844532436794
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,8192,8192,0.3316471046871609
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,8192,8192,0.4111502170562744
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,8192,8192,0.30465600225660533
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,8192,7168,0.28878045082092285
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,8192,7168,0.36548177401224774
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,8192,7168,0.2698604530758328
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,8192,6144,0.24208799997965494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,8192,6144,0.3145848910013835
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,8192,6144,0.23175467385186088
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,8192,5120,0.2056302229563395
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,8192,5120,0.19461688730451795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,8192,5120,0.2686684396531847
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,8192,4096,0.1623964442147149
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,8192,4096,0.16020354959699842
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,8192,3584,0.14490400420294866
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,8192,4096,0.21638755003611246
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,8192,3584,0.1938471131854587
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,8192,3072,0.12344088819291855
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,8192,3584,0.13959822389814588
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,8192,3072,0.16781866550445557
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,8192,2560,0.1039502223332723
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,8192,3072,0.12260977427164714
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,8192,2560,0.14657155672709146
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,8192,2048,0.08542399936252171
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,8192,2560,0.10449600219726562
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,8192,2048,0.12333688471052383
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,8192,1536,0.06690933307011922
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,8192,2048,0.08456533484988743
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,8192,65536,2.827346589830187
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,8192,1536,0.06887733274035983
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,8192,1024,0.049192888869179614
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,8192,1536,0.10080799791547988
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,8192,1024,0.05269955595334371
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,8192,768,0.039897776312298246
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,8192,768,0.045759999089770846
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,8192,1024,0.07644177807701959
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,8192,512,0.03045155604680379
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,8192,768,0.06631733311547174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,8192,512,0.03739022215207418
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,8192,256,0.021132444341977436
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,8192,512,0.05504177676306831
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,8192,256,0.030765334765116375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,8192,128,0.016226665841208566
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,8192,256,0.04384088847372267
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,8192,128,0.02791022260983785
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,8192,64,0.013820444544156393
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,8192,128,0.03861244519551595
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,8192,32,0.014526221487257214
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,8192,64,0.02709688908523983
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,8192,32,0.026523555318514507
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,8192,65536,3.1590222252739797
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,7168,16384,0.6083040237426758
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,8192,65536,2.6971910264756946
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,7168,16384,0.5189938015407985
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,7168,12288,0.4435546663072374
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,7168,12288,0.3859937720828586
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,7168,10240,0.3761884371439616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,7168,10240,0.32424354553222656
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,7168,8192,0.31199200948079425
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,7168,8192,0.2586142222086589
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,7168,7168,0.22596089045206705
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,7168,65536,2.5926168229844833
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,7168,7168,0.2661031087239583
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,7168,6144,0.19460177421569824
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,7168,6144,0.22838576634724936
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,7168,65536,2.2872045305040145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,7168,5120,0.16494399971432155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,7168,5120,0.1921306716071235
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,7168,16384,0.7877395417955187
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,7168,4096,0.13325688574049208
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,7168,12288,0.6014613575405544
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,7168,4096,0.1526764498816596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,7168,10240,0.5073644320170084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,7168,3584,0.11680266592237686
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,7168,3584,0.1358213292227851
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,7168,8192,0.41115644243028426
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,7168,3072,0.11647910541958278
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,7168,3072,0.10097155306074355
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,7168,2560,0.08517333534028794
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,7168,2560,0.09775466389126247
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,7168,65536,3.145899454752604
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,7168,2048,0.07337333096398248
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,7168,2048,0.08179644743601482
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,7168,7168,0.3653875456915961
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,7168,1536,0.05436977744102478
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,7168,1536,0.06555466519461738
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,7168,6144,0.3156079981062147
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,7168,1024,0.03816533419820998
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,7168,1024,0.05018666717741224
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,7168,5120,0.2704968982272678
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,7168,768,0.030579557021458943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,7168,768,0.043848001294665866
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,7168,3584,0.19615288575490317
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,7168,512,0.022691556149058874
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,7168,512,0.03558222121662564
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,7168,4096,0.21960977713267008
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,7168,256,0.015564444992277356
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,7168,256,0.028973334365420874
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,7168,2560,0.14848000473446316
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,7168,128,0.013023999830087027
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,7168,128,0.026013332936498854
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,7168,3072,0.17025244235992432
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,7168,64,0.012522666818565793
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,7168,64,0.024736000431908503
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,7168,32,0.013409777647919126
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,7168,32,0.024663999676704407
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,7168,1536,0.10101511081059773
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,7168,2048,0.12441689438290066
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,7168,768,0.0676177806324429
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,7168,1024,0.07854133182101779
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,6144,16384,0.45452531178792316
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,6144,65536,2.3418763478597007
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,6144,16384,0.5640302234225804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,6144,16384,0.7008346451653374
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,6144,65536,2.7445786794026694
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,6144,12288,0.41374844974941677
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,6144,12288,0.33965065744188094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,6144,10240,0.3493359883626302
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,6144,65536,2.0855831570095487
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,6144,12288,0.5330479939778646
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,6144,10240,0.2849457793765598
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,6144,8192,0.28901423348320854
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,6144,10240,0.4487066798739963
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,6144,8192,0.22780089908175996
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,6144,8192,0.365938663482666
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,6144,7168,0.20333066251542833
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,7168,128,0.040213333235846624
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,6144,7168,0.24886488914489746
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,6144,6144,0.21120444933573404
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,6144,5120,0.18172266748216417
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,7168,256,0.044455111026763916
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,6144,6144,0.17376532819535997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,6144,7168,0.324219544728597
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,6144,4096,0.14268978436787924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,7168,512,0.05609333515167236
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,6144,6144,0.281056006749471
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,6144,5120,0.2388239966498481
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,6144,5120,0.14673244953155518
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,6144,4096,0.1930248869789971
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,6144,4096,0.12133333418104385
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,6144,3584,0.12719999419318304
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,6144,3072,0.10788444677988689
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,6144,3584,0.17304444313049316
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,6144,2560,0.07859200239181519
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,6144,2560,0.09249333540598552
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,6144,3072,0.09272533655166626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,6144,2048,0.06551822026570638
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,6144,3072,0.15114311377207437
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,6144,3584,0.10684355762269761
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,6144,2048,0.07607199748357137
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,6144,1536,0.05243200063705444
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,6144,2560,0.1317840019861857
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,6144,2048,0.11015022463268703
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,6144,1024,0.038332445753945246
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,6144,1536,0.09101421965493096
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,6144,1536,0.06150311231613159
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,6144,768,0.0313022236029307
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,6144,1024,0.04751822352409363
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,6144,1024,0.07045688894059923
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,6144,768,0.041217777464124895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,6144,512,0.023988443944189284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,6144,768,0.060637334982554116
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,6144,512,0.03330399923854404
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,6144,256,0.017229333519935608
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,6144,128,0.02407288882467482
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,6144,512,0.050252444214291044
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,6144,256,0.027063111464182537
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,6144,128,0.013198222551080914
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,6144,256,0.04050222370359633
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,6144,64,0.011176889141400656
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,6144,64,0.02269866731431749
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,6144,32,0.011483555866612328
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,6144,128,0.034952001439200506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,6144,32,0.02280355493227641
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,5120,16384,0.5159777535332574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,5120,12288,0.2840675512949626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,5120,16384,0.3779733445909288
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,5120,12288,0.3790160020192464
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,5120,10240,0.24145865440368652
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,5120,10240,0.32039909892612034
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,5120,8192,0.19440978103213838
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,5120,8192,0.2650364504920112
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,5120,65536,2.104452345106337
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,5120,7168,0.22877245479159883
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,5120,7168,0.16871466901567247
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,5120,6144,0.14557955000135633
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,5120,6144,0.19574577278561064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,5120,65536,1.6353271272447374
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,5120,5120,0.1266382270389133
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,5120,5120,0.16706044144100615
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,5120,16384,0.6984151204427084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,5120,4096,0.09961422284444173
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,5120,4096,0.1324168841044108
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,5120,12288,0.5334453582763672
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,5120,3584,0.0891031093067593
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,5120,3584,0.11724710464477539
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,5120,10240,0.45026667912801105
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,5120,3072,0.07703289058473375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,5120,3072,0.10023911131752862
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,5120,8192,0.36475110054016113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,5120,2560,0.06583999925189547
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,5120,2560,0.08459466695785522
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,5120,65536,2.696664810180664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,5120,7168,0.3244711028205024
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,5120,2048,0.0549759997261895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,5120,2048,0.07088000244564481
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,5120,1536,0.04309689005215963
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,5120,1536,0.057479110029008657
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,5120,6144,0.2816630999247233
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,5120,1024,0.031007111072540283
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,5120,1024,0.044758223825030856
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,5120,5120,0.23982932832505968
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,5120,768,0.025314667158656653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,5120,768,0.03800355394681295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,5120,3584,0.174599117702908
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,5120,512,0.01924888955222236
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,5120,512,0.03083911206987169
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,5120,4096,0.19405244456397164
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,5120,256,0.013540444274743399
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,5120,2560,0.13234133190578884
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,5120,256,0.02443733314673106
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,5120,128,0.009970666633711921
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,5120,128,0.02145244512293074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,5120,3072,0.15241332848866782
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,5120,64,0.008542222281297049
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,5120,64,0.021086222595638696
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,5120,1536,0.09071555402543809
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,5120,32,0.00888355573018392
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,5120,32,0.021199999584092036
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,5120,2048,0.11072622405158149
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,5120,768,0.0611706640985277
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,4096,65536,1.8990071614583333
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,5120,1024,0.07190755340788099
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,4096,16384,0.4716657532585992
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,4096,16384,0.29932355880737305
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,4096,16384,0.6047022077772353
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,4096,65536,1.367529763115777
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,4096,12288,0.22598844104342988
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,4096,12288,0.34946489334106445
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,4096,65536,2.3240542941623263
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,4096,10240,0.2936675548553467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,4096,12288,0.46076446109347874
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,4096,10240,0.19099999798668754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,4096,8192,0.24052800072564018
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,4096,8192,0.1519671148724026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,4096,10240,0.3922737704383002
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,4096,7168,0.13513510757022434
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,4096,8192,0.3184373378753662
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,4096,7168,0.20988444487253824
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,4096,7168,0.28236354721917045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,4096,6144,0.18067377143436006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,4096,6144,0.24617510371738008
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,4096,5120,0.15288533104790583
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,4096,5120,0.09824089209238689
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,4096,5120,0.21099466747707793
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,4096,6144,0.11772444513108994
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,4096,4096,0.1224000056584676
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,4096,3584,0.07168533404668172
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,4096,4096,0.17143377992841932
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,4096,3584,0.10745422045389812
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,4096,4096,0.07959200276268853
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,4096,3072,0.06269777483410306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,4096,3584,0.15199022822909886
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,4096,3072,0.09166755278905232
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,4096,2560,0.052743110391828746
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,4096,3072,0.13293066289689806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,4096,2560,0.0778764420085483
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,4096,2048,0.043783999151653714
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,4096,2048,0.06474400228924222
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,4096,2560,0.11512800057729085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,4096,1536,0.03452800048722161
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,4096,2048,0.0972835554016961
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,4096,1536,0.053447110785378345
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,4096,1024,0.0251226673523585
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,4096,1536,0.08121422264311048
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,4096,1024,0.03968800107638041
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,4096,768,0.020554666717847187
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,4096,1024,0.06104355388217502
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,4096,768,0.035537779331207275
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,4096,512,0.01591199967596266
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,4096,512,0.02772533396879832
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,4096,768,0.053977777560551964
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,4096,256,0.010971555279360877
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,4096,512,0.04524088899294535
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,4096,128,0.008382221890820397
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,4096,256,0.022408000297016565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,4096,128,0.02026844521363576
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,4096,64,0.007207999626795451
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,4096,256,0.034556445148256086
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,4096,128,0.031572444571389094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,4096,32,0.007629333270920648
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,4096,64,0.019751111666361492
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,4096,32,0.019369777705934312
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,5120,512,0.050791111257341176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,5120,128,0.03628888726234436
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,5120,256,0.04079466727044847
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3584,16384,0.298638211356269
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3584,16384,0.4672231144375271
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3584,12288,0.35012266370985246
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3584,12288,0.2265226576063368
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,3584,16384,0.606279108259413
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3584,10240,0.2945680088467068
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,3584,12288,0.4661448796590169
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3584,10240,0.18870311313205293
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,3584,10240,0.39119556215074325
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3584,8192,0.24118577109442818
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3584,8192,0.15155733956231013
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,3584,8192,0.3222800095876058
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3584,7168,0.20958044793870714
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3584,65536,1.8507217831081813
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3584,7168,0.1351191070344713
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,3584,7168,0.2832924524943034
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3584,6144,0.18084800243377686
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3584,6144,0.11539288361867268
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,3584,6144,0.24730934037102592
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3584,5120,0.15271466308169895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3584,4096,0.0805528892411126
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,3584,5120,0.20961333645714653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3584,4096,0.12191822793748643
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3584,5120,0.09815733300315009
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3584,3584,0.07121866941452026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3584,3584,0.10766400231255426
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,3584,4096,0.1719528834025065
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3584,3072,0.061362663904825844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3584,3072,0.09305510918299358
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,3584,3584,0.1533013317320082
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3584,2560,0.05319199959437052
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3584,2560,0.07929955588446723
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,3584,3072,0.1341093381245931
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3584,2048,0.043544001049465604
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3584,2048,0.06444888644748263
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,3584,2560,0.11535378297170003
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3584,1536,0.03413510984844632
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,3584,2048,0.09804977973302205
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3584,1536,0.05344444513320923
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3584,1024,0.024894222617149353
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,3584,1536,0.08044977982838948
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3584,1024,0.03972533345222473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3584,768,0.020486222373114694
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,3584,1024,0.06287733051511976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3584,768,0.03501244386037191
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3584,512,0.015604444675975375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,3584,768,0.05494844582345751
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,3584,512,0.045455111397637256
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3584,512,0.027642667293548584
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3584,256,0.010687111152542962
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3584,256,0.022293332550260756
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,3584,65536,2.344543033175998
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3584,128,0.008325333396593729
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3584,65536,1.1769306394788954
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,3584,256,0.03510488735304938
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3584,128,0.019723556107944913
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3584,64,0.006993778049945831
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,3584,128,0.0324248903327518
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3584,64,0.01955200069480472
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3584,32,0.007705777883529663
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3584,32,0.01945600079165565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3072,12288,0.31876977284749347
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3072,16384,0.42853689193725586
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3072,16384,0.22884265581766763
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3072,12288,0.1717982159720527
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3072,10240,0.2686515649159749
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3072,10240,0.14412355422973633
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3072,8192,0.22234400113423666
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3072,8192,0.1185528834660848
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3072,7168,0.10424711306889851
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3072,65536,1.6773484547932942
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3072,7168,0.1930302249060737
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3072,6144,0.0904231137699551
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3072,6144,0.16668355464935303
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3072,65536,1.0438675350613065
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3072,5120,0.0759377810690138
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3072,5120,0.1411093340979682
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,3072,16384,0.559982246822781
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3072,4096,0.061595552497439914
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3072,4096,0.1132426659266154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,3072,12288,0.4266951084136963
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3072,3584,0.05552088883188036
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3072,3584,0.1001431081030104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,3072,10240,0.360534217622545
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3072,3072,0.04809777935345968
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3072,3072,0.08552889029184978
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,3072,8192,0.294123543633355
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3072,2560,0.0407244430647956
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3072,2560,0.07314222388797335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,3072,7168,0.2619226773579915
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3072,2048,0.03379022081693014
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3072,2048,0.05982044670316908
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,3072,6144,0.22899111111958823
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3072,1536,0.026904000176323786
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,3072,65536,2.149828381008572
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3072,1536,0.049081779188579984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3072,1024,0.019855111837387085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3072,1024,0.03727911247147454
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,3072,5120,0.1948311063978407
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3072,768,0.016400888562202454
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3072,768,0.03134222163094415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,3072,3584,0.14159822463989258
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3072,512,0.012776888906955719
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3072,512,0.026649778087933857
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,3072,4096,0.1598115497165256
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3072,256,0.0087333329849773
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3072,256,0.020735111501481798
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,3072,2560,0.10691555341084798
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3072,128,0.006974221931563483
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,3072,768,0.050548444191614784
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3072,128,0.018459555175569322
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3072,64,0.006083555519580841
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3072,64,0.018389332625601027
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,3072,32,0.0064088888466358185
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,3072,32,0.018332444959216647
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,3072,3072,0.12332710954878066
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,3072,1536,0.0751182238260905
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,3072,2048,0.09102488888634576
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2560,16384,0.19043555524614122
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2560,16384,0.4231288962894016
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2560,65536,0.828863090938992
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2560,65536,1.6566427018907335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2560,12288,0.3169608910878499
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2560,12288,0.14749511082967123
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,3072,1024,0.05830844243367513
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2560,10240,0.1223191155327691
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2560,10240,0.2679048909081353
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2560,8192,0.09790133105383979
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2560,8192,0.21997866365644667
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,3072,128,0.03064889046880934
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,3072,256,0.03352444370587667
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2560,7168,0.08673422204123603
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2560,7168,0.18922844198015001
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2560,6144,0.07540888918770684
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2560,6144,0.16346044010586208
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,3072,512,0.042097777128219604
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2560,5120,0.06444799900054932
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2560,5120,0.1405831045574612
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,2560,12288,0.4121697743733724
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2560,4096,0.05160355567932129
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2560,4096,0.11353511280483669
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,2560,16384,0.5395529005262587
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,2560,65536,2.069951163397895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2560,3584,0.046147555112838745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2560,3584,0.10055110851923625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2560,3072,0.040580444865756564
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2560,3072,0.08598755465613471
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,2560,8192,0.2860106627146403
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,2560,10240,0.3493671152326796
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2560,2560,0.0343413319852617
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2560,2560,0.07274578015009563
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2560,2048,0.028444445795483057
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2560,2048,0.05988533629311455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,2560,7168,0.2539173232184516
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2560,1536,0.022853333089086745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2560,1536,0.04884355598025852
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,2560,6144,0.22226399845547148
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2560,1024,0.017105778058369953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2560,1024,0.03737599982155694
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,2560,5120,0.1888951063156128
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2560,768,0.014230221509933472
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2560,768,0.03069066670205858
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,2560,3584,0.1381031142340766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2560,512,0.011131555669837527
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2560,512,0.02520177761713664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,2560,2560,0.10538844267527263
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2560,256,0.00796888851457172
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2560,256,0.020637333393096924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,2560,4096,0.15482132964664036
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2560,128,0.006539555473460092
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2560,128,0.018601778480741713
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,2560,3072,0.12078400452931722
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2560,64,0.005802666561471091
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2560,64,0.018352889352374606
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,2560,2048,0.088755554623074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2560,32,0.006034666465388403
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2560,32,0.018221333622932434
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,2560,768,0.05004088746176826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2048,65536,1.515250629848904
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2048,16384,0.15008889304267034
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2048,65536,0.5887813568115234
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,2560,1536,0.07393955522113375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2048,16384,0.3890177673763699
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,2048,65536,1.9634009467230902
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2048,12288,0.28719109959072536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2048,12288,0.11494310696919759
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,2048,16384,0.5161937607659234
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,2048,12288,0.3934924337599013
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2048,10240,0.24235288302103677
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2048,10240,0.09647999869452582
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,2048,10240,0.3335404396057129
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2048,8192,0.07780000236299303
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2048,8192,0.19864888985951742
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2048,7168,0.06909955872429742
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,2048,8192,0.2724631097581652
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2048,7168,0.17450755172305635
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2048,6144,0.06000355879465739
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,2048,7168,0.2427013450198703
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2048,6144,0.15069244967566595
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2048,5120,0.05090399914317661
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,2048,6144,0.211517333984375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2048,4096,0.04173422190878126
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2048,5120,0.12746755282084146
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,2048,5120,0.1792924404144287
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2048,4096,0.10321955548392402
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2048,3584,0.037117332220077515
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,2048,4096,0.14745155970255533
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2048,3584,0.09059288766649033
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2048,3072,0.032525334093305804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,2048,3584,0.13052799966600206
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2048,3072,0.07817511426077949
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2048,2560,0.027980443504121568
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,2048,3072,0.1140320036146376
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2048,2560,0.066520889600118
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2048,2048,0.023377777801619634
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,2048,2560,0.09970666964848836
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2048,2048,0.05499733487764994
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2048,1536,0.018792889184421964
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,2048,2048,0.08492977751625909
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2048,1024,0.014027555783589682
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2048,1536,0.044221334987216525
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2048,1024,0.03340444299909804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,2048,1536,0.07005511389838324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2048,768,0.011735111474990845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,2048,1024,0.05454311105940077
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2048,768,0.028076443407270644
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2048,512,0.009197333620654212
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,2048,768,0.04599555664592319
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2048,256,0.006909333169460297
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2048,512,0.023082666926913794
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2048,256,0.019056000643306308
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,2048,512,0.03832355472776625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2048,128,0.005716444303592046
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,2048,256,0.03211022085613675
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2048,64,0.005032888717121548
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2048,128,0.016865778300497267
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,2048,32,0.0053084443012873335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2048,64,0.016743999388482835
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,2048,128,0.029487109846538965
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,2560,1024,0.057589332262674965
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,2048,32,0.016591999265882704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1536,16384,0.3850942187839084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1536,16384,0.1164533297220866
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,1536,16384,0.4916231367323134
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1536,12288,0.08864000108506943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1536,12288,0.2899297873179118
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,1536,12288,0.3776853349473741
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1536,10240,0.07373599873648749
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1536,10240,0.24428534507751465
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,1536,10240,0.31860444280836314
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1536,8192,0.06079555882347954
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1536,65536,0.44918224546644425
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1536,65536,1.4857955508761937
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1536,7168,0.0621946652730306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1536,8192,0.19959378242492676
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1536,7168,0.175491558180915
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,1536,8192,0.261806223127577
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1536,6144,0.04613599843449063
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,1536,65536,1.8725164201524522
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,1536,7168,0.23577777544657388
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,2560,128,0.03049422303835551
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1536,5120,0.03916710946294997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1536,6144,0.15173510710398355
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1536,4096,0.032495998673968844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,1536,6144,0.20327911112043592
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1536,5120,0.12869600454966226
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,1536,5120,0.173962672551473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1536,4096,0.10406755738788181
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1536,3584,0.028938667641745672
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1536,3584,0.09100533194012111
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,1536,4096,0.14164533880021837
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1536,3072,0.025794666674402025
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,1536,3584,0.12510844071706137
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1536,3072,0.0783804456392924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,1536,3072,0.10962311426798503
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1536,2560,0.022152000003390845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1536,2560,0.06670133272806804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1536,2048,0.018584888842370775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,1536,2560,0.09657688935597737
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1536,1536,0.01478488908873664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1536,2048,0.05499733487764994
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,1536,2048,0.08177066511578031
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1536,1536,0.044256889157825045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1536,1024,0.011158222125636207
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,1536,1536,0.06769155793719821
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,2560,256,0.032701333363850914
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1536,1024,0.03216266632080078
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1536,768,0.009347555538018545
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,1536,1024,0.052741332186592944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1536,768,0.029069334268569946
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1536,512,0.007518221934636434
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,1536,768,0.04680888851483663
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1536,512,0.02344977855682373
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1536,256,0.0058764442801475525
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,1536,512,0.03761511047681173
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1536,256,0.01924622224436866
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1536,128,0.00499911109606425
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,1536,256,0.031128889984554712
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1536,128,0.017120889491505094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1536,64,0.004440000073777305
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,1536,128,0.028844445943832397
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1536,32,0.0046986668474144405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1536,64,0.01680177781316969
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1536,32,0.016491555505328707
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1024,16384,0.34809599982367617
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1024,16384,0.07834844456778632
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1024,12288,0.060079110993279346
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1024,12288,0.26295555962456596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1024,10240,0.05047022302945455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1024,10240,0.22215822007921007
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1024,8192,0.041231112347708807
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1024,65536,1.341615147060818
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1024,8192,0.18179644478691948
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1024,7168,0.036728888750076294
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1024,65536,0.3039013279808892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1024,7168,0.16021511289808485
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1024,6144,0.03202844328350491
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1024,6144,0.13879556126064724
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,2560,512,0.041215999258889094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1024,5120,0.02735822233888838
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1024,5120,0.1175502273771498
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,1024,16384,0.47222667270236546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1024,4096,0.022522666388087805
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,1024,12288,0.3623315493265788
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1024,4096,0.09577333264880711
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1024,3584,0.02037333283159468
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1024,3584,0.08406133121914333
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,1024,65536,1.805665757921007
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,1024,10240,0.3065066602494982
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,1024,8192,0.25138844384087455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1024,3072,0.01800622211562263
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,1024,7168,0.22294399473402235
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1024,2560,0.015564444992277356
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1024,3072,0.07189955314000447
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1024,2560,0.06042933464050293
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1024,2048,0.013212444053755866
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1024,2048,0.05057955450481839
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1024,1536,0.01069777790043089
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1024,1536,0.040933334165149264
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,1024,6144,0.19419999917348227
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1024,1024,0.008249777886602614
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1024,1024,0.030039111773173015
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,1024,5120,0.16588000456492105
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1024,768,0.0070977773931291364
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1024,768,0.026068444053332012
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,1024,4096,0.13557777139875624
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,1024,3584,0.12027466297149658
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1024,512,0.006038222048017714
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1024,512,0.02142222225666046
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1024,256,0.004866666677925321
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1024,256,0.017702221870422363
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,1024,2560,0.09223644600974189
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1024,128,0.004315555509593752
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,1024,3072,0.10570133394665188
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1024,128,0.015666666958067153
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1024,64,0.0038773334688610504
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1024,64,0.015255999233987598
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,1024,2048,0.07845511039098103
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,1024,32,0.004104888687531154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,1024,1536,0.06599822309282091
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,1024,32,0.015173332558737861
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,768,65536,0.23205600844489205
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,768,16384,0.05937066343095568
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,1024,1024,0.049633777803844877
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,768,65536,1.2954319847954643
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,768,16384,0.33924267027113175
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,768,12288,0.04575644599066841
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,768,12288,0.25891555680169
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,768,10240,0.03929600119590759
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,768,10240,0.21880711449517143
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,1024,768,0.04328711165322197
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,768,8192,0.031874666611353554
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,768,8192,0.1776986651950412
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,1024,512,0.036361777120166354
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,1024,256,0.030124445756276447
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,1024,128,0.028037332826190527
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,768,7168,0.02821422285503811
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,768,7168,0.1575360033247206
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,768,6144,0.02569688856601715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,768,6144,0.13669955730438232
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,768,5120,0.02141155633661482
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,768,5120,0.11604444185892741
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,768,12288,0.3599813249376085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,768,4096,0.017628444565667045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,768,4096,0.09350844224294026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,768,16384,0.4703973134358724
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,768,3584,0.01607733302646213
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,768,3584,0.08194577693939209
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,768,65536,1.7899350060356989
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,768,3072,0.014315555493036905
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,768,10240,0.30567823515997994
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,768,8192,0.24953688515557182
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,768,3072,0.07077333662245008
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,768,2560,0.012492444780137805
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,768,2560,0.05971466832690769
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,768,2048,0.010399111443095738
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,768,2048,0.049878223074807056
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,768,7168,0.22184356053670248
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,768,1536,0.008804444637563493
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,768,1536,0.039898667070600725
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,768,6144,0.19446221987406412
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,768,1024,0.0069368887278768755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,768,1024,0.029288000530666772
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,768,5120,0.1651511059867011
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,768,768,0.006112888869312074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,768,768,0.02693333393997616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,768,3584,0.12045866913265652
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,768,512,0.005225777626037598
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,768,2560,0.09239644474453396
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,768,512,0.021224000387721594
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,768,256,0.004377777791685528
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,768,256,0.017473777135213215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,768,4096,0.1351146697998047
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,768,128,0.003947555604908201
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,768,3072,0.10559022426605225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,768,128,0.015508444772826301
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,768,64,0.003601777884695265
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,768,64,0.015089778436554803
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,768,32,0.0037635556525654266
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,768,32,0.014876445134480795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,768,2048,0.07895200120078193
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,512,65536,0.15907822714911565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,512,65536,1.2950542238023546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,512,16384,0.04778933193948534
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,512,16384,0.33759911855061847
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,768,768,0.044870220952563815
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,512,12288,0.03878844446606106
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,512,12288,0.2584773434533013
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,768,1536,0.06601333618164062
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,512,10240,0.03245688809288873
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,512,10240,0.21791466077168783
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,768,1024,0.049309333165486656
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,512,8192,0.03124799993303087
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,768,256,0.029858665333853826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,768,128,0.027661333481470745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,512,8192,0.1778728829489814
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,512,7168,0.024864888853496973
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,512,7168,0.15707377592722574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,512,6144,0.02183733383814494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,512,6144,0.1362533304426405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,768,512,0.035762667655944824
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,512,5120,0.019169777631759644
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,512,5120,0.11566489272647434
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,512,65536,1.7406222025553386
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,512,4096,0.016925333274735343
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,512,4096,0.09368355406655206
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,512,16384,0.45629156960381406
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,512,3584,0.014372444815105863
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,512,3584,0.0815804468260871
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,512,12288,0.3506293296813965
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,512,3072,0.01147555559873581
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,512,3072,0.07015999820497301
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,512,10240,0.29676977793375653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,512,2560,0.010456889039940305
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,512,2560,0.059742218918270536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,512,8192,0.24310043123033312
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,512,2048,0.00850399997499254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,512,7168,0.2158515585793389
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,512,2048,0.0499582224422031
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,512,1536,0.007476444045702617
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,512,1536,0.03996088769700792
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,512,6144,0.18910133838653564
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,512,1024,0.005993777679072485
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,512,1024,0.028885334730148315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,512,5120,0.16070844067467585
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,512,768,0.0053102220926019884
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,512,768,0.025156444973415796
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,512,4096,0.13130311171213785
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,512,512,0.004651555584536659
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,512,3584,0.1167715522978041
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,512,512,0.021036444438828364
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,512,256,0.004024888906213972
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,512,256,0.017291555802027386
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,512,3072,0.10271377695931329
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,512,2560,0.09014666742748684
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,512,128,0.003647110942337248
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,512,128,0.015252444479200574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,512,64,0.0034017778105205963
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,512,64,0.014904888139830695
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,512,32,0.003495111233658261
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,512,32,0.01477599971824222
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,256,65536,0.09636177619298299
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,256,65536,1.2954897350735135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,512,2048,0.0769875579410129
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,256,16384,0.031211555004119873
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,512,1536,0.06411644485261705
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,256,16384,0.3380284574296739
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,256,12288,0.0273297776778539
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,256,12288,0.25820888413323295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,512,1024,0.04799377918243408
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,256,10240,0.024533333049880132
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,256,10240,0.21819467014736602
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,512,256,0.029488888051774766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,256,8192,0.020120888948440552
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,256,8192,0.17686666382683647
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,512,512,0.03533600105179681
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,512,768,0.041328890456093684
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,256,7168,0.014914666612943014
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,256,7168,0.15659911102718777
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,256,6144,0.013059555656380124
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,256,6144,0.136043561829461
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,512,128,0.027414222558339436
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,256,5120,0.011721777419249216
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,256,5120,0.11523377895355225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,256,65536,1.7045297622680664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,256,4096,0.010266666611035665
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,256,4096,0.09385155306922065
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,256,16384,0.4503466818067763
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,256,3584,0.009085333181752099
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,256,3584,0.08140355348587036
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,256,12288,0.3437093363867865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,256,3072,0.008090666598743862
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,256,3072,0.06971910927030775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,256,10240,0.2914302349090576
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,256,2560,0.007531555162535773
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,256,2560,0.05941333373387655
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,256,8192,0.2385235627492269
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,256,2048,0.0064426664676931165
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,256,2048,0.04957688848177592
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,256,7168,0.21150399578942192
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,256,1536,0.005727111051479976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,256,1536,0.0399528874291314
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,256,6144,0.18436977598402235
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,256,1024,0.004816888935036129
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,256,1024,0.028986665937635634
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,256,5120,0.15802400641971165
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,256,768,0.004534222185611725
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,256,768,0.024832000335057575
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,256,4096,0.12895911269717747
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,256,512,0.004031110968854693
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,256,512,0.020841777324676514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,256,3584,0.11478043927086724
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,256,256,0.003531555748648114
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,256,3072,0.10035200251473321
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,256,256,0.016939555605252583
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,256,128,0.0032400000426504347
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,256,128,0.014912888407707214
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,256,2560,0.08809066481060451
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,256,64,0.003104000041882197
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,256,64,0.014410666293568082
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,256,32,0.0033617777129014335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,256,32,0.01425066590309143
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,128,65536,0.07927022377649943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,256,2048,0.07483733362621732
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,128,65536,1.2966409259372287
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,128,16384,0.025791111919615004
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,256,1536,0.0618835555182563
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,128,16384,0.3382408883836534
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,128,12288,0.020868443780475195
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,256,1024,0.04750577939881218
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,128,12288,0.25814665688408744
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,128,10240,0.016389333539538913
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,128,10240,0.21814044316609701
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,256,768,0.04029688901371426
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,128,8192,0.013380444712109037
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,128,8192,0.17699200577206084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,256,512,0.03455466694302029
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,128,7168,0.012042666474978128
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,128,7168,0.15701954894595677
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,256,256,0.029003553920321997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,128,6144,0.010993777877754636
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,256,128,0.027319111757808264
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,128,6144,0.13606399959988064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,128,5120,0.010125333236323463
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,128,5120,0.11530400647057427
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,128,16384,0.4424862331814236
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,128,4096,0.010371555884679159
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,128,4096,0.09324089023802017
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,128,12288,0.34027822812398273
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,128,3584,0.009373333719041612
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,128,65536,1.681656837463379
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,128,3584,0.08167288700739543
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,128,3072,0.00813333359029558
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,128,10240,0.2877599928114149
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,128,3072,0.06976177957322863
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,128,2560,0.007121777368916406
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,128,2560,0.0592177775171068
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,128,8192,0.23579110039605033
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,128,2048,0.0064239998658498125
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,128,2048,0.049359109666612416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,128,7168,0.2093591160244412
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,128,1536,0.005625777774386936
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,128,1536,0.03969688879119025
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,128,6144,0.18251821729871961
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,128,1024,0.0046773333516385825
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,128,5120,0.15578222274780273
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,128,1024,0.028934223784340754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,128,768,0.004275555411974589
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,128,768,0.024646222591400146
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,128,4096,0.1275377803378635
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,128,512,0.0038133333954546186
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,128,512,0.020503110355801053
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,128,3584,0.11323466565873887
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,128,256,0.0033973333322339584
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,128,256,0.01682577696111467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,128,3072,0.09930133157306248
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,128,128,0.0030951110853089225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,128,128,0.014780445231331719
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,128,2560,0.08661422464582656
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,128,64,0.0029644444584846497
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,128,64,0.014215999179416232
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,128,32,0.0030462222380770575
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,128,32,0.014039999908871122
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,64,65536,0.07261421945359972
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,128,2048,0.07436711258358426
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,64,16384,0.02360533343421088
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,64,16384,0.3384444448682997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,64,12288,0.018034666776657104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,64,65536,1.2959733539157443
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,64,10240,0.013606222139464485
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,64,12288,0.2581128809187147
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,64,10240,0.2179493374294705
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,64,8192,0.011491555306646558
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,64,7168,0.010587555666764578
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,64,8192,0.1773271163304647
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,64,6144,0.009873777627944946
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,64,7168,0.15690844588809544
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,64,5120,0.009196444518036312
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,64,6144,0.13600533538394505
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,64,4096,0.008343111309740279
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,64,5120,0.11562844117482503
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,64,4096,0.09355911281373765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,64,3584,0.007696888513035244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,64,3584,0.08151733213000827
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,64,3072,0.007313778003056844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,64,3072,0.0695564415719774
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,64,2560,0.006613333192136552
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,64,2048,0.005882666756709416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,64,2560,0.059280885590447314
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,64,1536,0.005122666557629903
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,64,2048,0.04955110947291056
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,64,1536,0.039501332574420504
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,64,1024,0.004378666480382283
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,64,1024,0.02906755606333415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,64,768,0.004021333323584663
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,64,768,0.02464177707831065
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,64,512,0.003644444462325838
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,64,512,0.020524443851576913
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,64,256,0.0033271111961868075
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,64,256,0.016736888223224215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,64,128,0.00300444434914324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,64,128,0.014689778288205465
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,64,64,0.0028577778074476454
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,64,32,0.002899555489420891
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,64,64,0.01421688828203413
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,64,32,0.014109333356221518
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,32,65536,0.07023555702633329
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,32,16384,0.022407111194398668
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,128,1536,0.061591108640034996
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,32,12288,0.016588444511095684
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,32,16384,0.33818401230706113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,32,10240,0.011552888486120435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,32,12288,0.25839821497599286
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,32,10240,0.21768800417582193
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,32,8192,0.010551999840471478
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,32,7168,0.009871111147933537
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,32,65536,1.295765346950955
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,32,8192,0.17722845077514648
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,32,7168,0.15683644347720677
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,32,6144,0.009174221919642555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,128,1024,0.04636889033847385
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,32,5120,0.008287111090289222
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,32,6144,0.13587110572391087
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,32,4096,0.007599111232492659
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,32,5120,0.11537155840131973
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,32,3584,0.007376000285148621
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,32,4096,0.09372444285286798
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,32,3584,0.08118400308820936
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,32,3072,0.007000889215204451
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,32,2560,0.006180444525347815
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,32,3072,0.06953155332141452
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,32,2560,0.05887555413775974
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,32,2048,0.005580444302823808
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,32,1536,0.004962666581074397
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,32,2048,0.049154665735032826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,32,1024,0.004315555509593752
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,32,1536,0.03906933466593424
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,32,768,0.003930666794379552
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,32,1024,0.02970577942000495
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,32,512,0.003616888903909259
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,32,768,0.024540444215138752
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,32,256,0.0032239999208185407
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,32,512,0.020613332589467365
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,32,128,0.002938666691382726
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,32,256,0.016744888491100736
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,32,64,0.002776888923512565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,32,128,0.014742222097184924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2048,32,32,0.002970666728085942
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,32,64,0.014204444156752693
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2048,32,32,0.014126222994592456
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,128,768,0.03993155558904012
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,128,512,0.03443200058407254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,128,256,0.02866577770974901
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,65536,12288,1.0845733218722873
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,65536,16384,1.5521351496378581
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,65536,12288,1.3821306228637695
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,65536,12288,2.1262800428602433
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,65536,10240,0.9475271436903211
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,65536,16384,1.8503333197699652
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,65536,10240,1.7941697438557942
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,65536,10240,1.1109191046820748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,65536,16384,2.8549156188964844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,65536,8192,0.7651919788784451
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,65536,8192,1.4148178100585938
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,65536,8192,0.8763635423448352
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,65536,7168,0.6635866694980198
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,65536,7168,0.8227421972486707
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,65536,6144,0.5450791252983941
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2048,128,128,0.027105778455734253
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,65536,6144,0.6543866793314616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,65536,7168,1.2324417961968315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,65536,5120,0.45744090610080296
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,65536,6144,1.0005679660373263
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,65536,5120,0.8118675549825033
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,65536,5120,0.5694417953491211
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,65536,4096,0.35341954231262207
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,65536,4096,0.4482924673292372
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,65536,4096,0.6560693316989475
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,65536,3072,0.27441954612731934
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,65536,3584,0.3312293423546685
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,65536,3584,0.4079128901163737
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,65536,3072,0.35069865650600857
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,65536,2560,0.22871910201178658
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,65536,2048,0.18574133184221056
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,65536,3072,0.4992195765177409
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,65536,3584,0.563653310139974
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,65536,2560,0.41396088070339626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,65536,2560,0.30386845270792645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,65536,2048,0.25154132313198513
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,65536,1536,0.2569608953264025
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,65536,1536,0.14794488747914633
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,65536,1024,0.11084444655312432
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,65536,1024,0.15057599544525146
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,65536,1536,0.2031857834921943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,65536,768,0.09402577744589911
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,65536,2048,0.33214489618937176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,65536,768,0.12337422370910645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,65536,768,0.14461510711246067
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,65536,512,0.10383911265267266
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,65536,1024,0.18539288308885363
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,65536,512,0.07987111144595675
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,65536,256,0.06859999895095825
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,65536,128,0.054197334580951266
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,65536,256,0.07991200023227267
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,65536,256,0.06435377730263604
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,65536,128,0.06364711125691731
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,65536,64,0.0484444432788425
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,65536,512,0.10001599788665771
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,65536,64,0.06428088744481404
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,65536,32,0.04976000057326423
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,65536,128,0.07550311088562012
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,65536,32,0.062236441506279834
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,16384,12288,0.35757867495218915
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,16384,16384,0.5927235815260146
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,16384,16384,0.5046150949266222
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,16384,12288,0.4534248775906033
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,16384,12288,0.4504302342732747
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,16384,16384,0.6064017613728842
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,16384,10240,0.30676266882154674
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,16384,8192,0.2538391219245063
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,16384,10240,0.37712444199456113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,16384,10240,0.38243288464016384
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,16384,8192,0.30814043680826825
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,16384,7168,0.20960622363620332
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,16384,8192,0.3045733239915636
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,16384,7168,0.2750186655256483
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,16384,7168,0.2659013271331787
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,16384,6144,0.17861333158281115
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,16384,6144,0.23734845055474174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,16384,5120,0.1548808945549859
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,16384,5120,0.1941199964947171
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,16384,5120,0.2031182183159722
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,16384,6144,0.23029867808024088
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,16384,65536,2.2217644585503473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,16384,4096,0.12257156107160781
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,16384,4096,0.1694906685087416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,16384,3584,0.11178844504886204
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,16384,4096,0.15874489148457846
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,16384,3584,0.1402613321940104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,16384,3584,0.15204621685875788
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,16384,2560,0.08331911431418525
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,16384,3072,0.09632711278067695
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,16384,3072,0.12142666180928548
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,16384,3072,0.1347751087612576
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,16384,2560,0.10396889183256362
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,16384,2048,0.06936444176567926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,16384,2560,0.11678489049275716
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,16384,1536,0.05738844474156698
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,16384,1536,0.06726221905814277
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,16384,2048,0.10010755724377102
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,16384,1024,0.04549333453178406
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,16384,2048,0.08540266752243042
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,16384,768,0.03979644510481093
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,16384,1536,0.08378489149941339
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,16384,1024,0.06617066595289442
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,16384,1024,0.04846133457289802
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,16384,768,0.03917510973082648
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,16384,512,0.03087200058831109
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,16384,256,0.021187555458810594
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,16384,768,0.05714311202367147
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,16384,512,0.03437688946723938
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,16384,512,0.04944800006018745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,16384,256,0.029530667596393164
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,16384,128,0.015776000089115567
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,16384,65536,2.40029419793023
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,16384,256,0.041943109697765775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,16384,128,0.02719822194841173
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,16384,64,0.01347555551264021
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,16384,128,0.038571556409200035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,16384,64,0.027051554785834417
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,16384,32,0.01405244403415256
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,16384,32,0.031445334355036415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,12288,16384,0.42349865701463485
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,12288,16384,0.4990551206800673
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,12288,12288,0.33712177806430393
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,12288,12288,0.30048712094624835
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,12288,16384,0.4496720102098253
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,12288,12288,0.3820524480607774
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,12288,10240,0.25238490104675293
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,12288,10240,0.3226239946153429
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,12288,10240,0.2858817842271593
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,12288,8192,0.2104053364859687
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,16384,65536,2.6880355411105685
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,12288,8192,0.22808800803290474
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,12288,8192,0.26188890139261883
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,12288,7168,0.17616711722479927
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,12288,7168,0.20206756061977812
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,12288,7168,0.23481334580315483
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,12288,6144,0.15106666088104248
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,12288,6144,0.17373955249786377
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,12288,5120,0.1269164482752482
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,12288,6144,0.20340177747938368
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,12288,5120,0.14806666639116076
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,12288,5120,0.1731039947933621
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,12288,4096,0.10363555616802639
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,12288,3584,0.09515200058619182
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,12288,4096,0.1456728908750746
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,12288,4096,0.11928710672590469
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,12288,3584,0.10662133163876003
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,12288,3584,0.13066044118669298
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,12288,65536,1.9588194953070746
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,12288,3072,0.08269688818189833
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,12288,3072,0.09348533550898235
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,12288,3072,0.11554133892059326
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,12288,65536,1.6966355641682942
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,12288,2560,0.0800959997706943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,12288,2560,0.07200533151626587
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,12288,2048,0.06041510899861654
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,12288,2560,0.10455555386013454
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,12288,1536,0.05028266708056132
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,12288,2048,0.08695911036597358
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,12288,1024,0.038542220989863075
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,12288,1536,0.07389511002434625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,12288,1024,0.03955022162861294
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,12288,2048,0.06594844659169515
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,12288,768,0.03118844495879279
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,12288,1024,0.05884799692365858
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,12288,768,0.051366223229302295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,12288,768,0.034671111239327326
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,12288,512,0.02420799930890401
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,12288,1536,0.05196266704135471
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,12288,512,0.0303422212600708
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,12288,256,0.017312000195185345
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,12288,128,0.013386666774749756
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,12288,512,0.0450524455971188
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,12288,256,0.025415110919210646
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,12288,256,0.038776887787712946
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,12288,64,0.011188444164064197
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,12288,128,0.023756444454193115
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,12288,128,0.03583644496070014
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,12288,32,0.011755555868148804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,12288,64,0.02256800068749322
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,12288,32,0.022987556126382615
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,10240,16384,0.372217787636651
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,10240,12288,0.2855902247958713
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,10240,16384,0.37540356318155926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,10240,12288,0.26832087834676105
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,10240,10240,0.22582933637830946
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,10240,10240,0.23946311738755968
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,10240,65536,1.530331505669488
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,12288,65536,2.057199902004666
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,10240,8192,0.19452355967627633
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,10240,8192,0.18905156188540986
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,10240,7168,0.15777688556247288
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,10240,7168,0.17027644316355386
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,10240,6144,0.14744444688161215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,10240,6144,0.13491555054982504
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,10240,65536,1.6673715379503038
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,10240,5120,0.12614133622911242
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,10240,5120,0.11494933234320746
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,10240,16384,0.4999102486504449
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,10240,4096,0.10087288750542535
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,10240,4096,0.09370488590664333
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,10240,12288,0.3851457701789008
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,10240,3584,0.0889964434835646
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,10240,3584,0.08573689063390096
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,10240,10240,0.32295555538601345
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,10240,3072,0.0770791106753879
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,10240,3072,0.07474222448137072
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,10240,8192,0.26525155703226727
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,10240,2560,0.06664266851213244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,10240,7168,0.23543288972642687
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,10240,2560,0.06595910920037164
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,10240,65536,1.9461296929253473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,10240,2048,0.054972443315717906
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,10240,2048,0.056870222091674805
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,10240,1536,0.043819556633631386
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,10240,1536,0.046832889318466187
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,10240,6144,0.2034355534447564
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,10240,1024,0.031567109955681696
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,10240,1024,0.03604266709751553
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,10240,5120,0.17442133691575792
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,10240,768,0.025418667329682246
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,10240,768,0.031677332189348005
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,10240,4096,0.1454204453362359
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,10240,512,0.019367999500698514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,10240,512,0.027441778116756018
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,10240,3584,0.13312355677286783
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,10240,256,0.013609777722093793
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,10240,256,0.023481777972645227
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,10240,2560,0.10303378105163574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,10240,3072,0.11523732874128555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,10240,128,0.010453333457310995
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,10240,128,0.0207022229830424
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,10240,64,0.008611555728647444
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,10240,64,0.020759999752044678
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,10240,32,0.009019555317031013
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,10240,32,0.02128977742460039
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,10240,1536,0.07418844434950086
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,10240,2048,0.08771999677022298
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,10240,1024,0.05980977747175428
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,10240,768,0.05197422371970283
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,8192,16384,0.3307644526163737
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,8192,65536,1.27982849544949
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,8192,16384,0.2994035614861382
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,8192,16384,0.406384891933865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,8192,12288,0.2354524400499132
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,10240,256,0.0390035543176863
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,10240,512,0.04564977685610453
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,8192,10240,0.19925688372717965
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,8192,12288,0.2250524361928304
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,8192,12288,0.31228799290127224
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,8192,65536,1.1937369240654838
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,8192,65536,1.569128884209527
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,10240,128,0.03532800078392029
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,8192,10240,0.2664319939083523
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,8192,10240,0.18993422720167372
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,8192,8192,0.21631199783749053
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,8192,7168,0.1407866610421075
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,8192,7168,0.19282844331529406
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,8192,8192,0.16500266393025717
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,8192,8192,0.15290755695766872
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,8192,7168,0.13602044847276476
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,8192,6144,0.1195875538720025
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,8192,5120,0.1008604433801439
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,8192,6144,0.11705511146121556
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,8192,6144,0.16661600271860758
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,8192,4096,0.08443733056386311
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,8192,3584,0.07648177941640218
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,8192,4096,0.07987378040949504
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,8192,5120,0.14417066839006212
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,8192,4096,0.12062488661872016
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,8192,3584,0.1085582243071662
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,8192,5120,0.09981599781248306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,8192,3072,0.06746221913231744
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,8192,3584,0.07136444250742595
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,8192,2560,0.05957688887914022
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,8192,3072,0.09797155857086182
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,8192,3072,0.0625493327776591
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,8192,2048,0.04877510997984144
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,8192,2560,0.08608444531758626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,8192,1536,0.0348364445898268
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,8192,2048,0.07368533478842841
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,8192,1536,0.042295111550225147
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,8192,1024,0.02536711096763611
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,8192,2048,0.045016888115141124
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,8192,2560,0.05278222097290886
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,8192,1024,0.032291554742389254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,8192,1024,0.05105511016315884
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,8192,1536,0.06430044439103869
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,8192,768,0.020427554845809937
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,8192,512,0.02551733288500044
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,8192,768,0.029081778393851385
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,8192,512,0.015847999188635085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,8192,768,0.04502133197254605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,8192,256,0.011381333072980246
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,8192,512,0.03990222348107232
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,8192,256,0.021591110361946955
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,8192,128,0.008650666309727563
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,8192,256,0.034817778401904635
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,8192,128,0.01941511034965515
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,8192,64,0.007231999602582719
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,8192,128,0.03127022253142463
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,8192,64,0.019541333119074505
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,8192,32,0.007686222592989604
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,8192,32,0.01939022209909227
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,7168,16384,0.3238133324517144
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,7168,16384,0.4125662114885118
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,7168,12288,0.22622044881184897
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,7168,12288,0.2363119920094808
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,7168,12288,0.3155013455284966
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,7168,16384,0.2988284428914388
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,7168,10240,0.19893422391679552
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,7168,10240,0.18823733594682482
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,7168,10240,0.2688159942626953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,7168,8192,0.16544355286492243
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,7168,8192,0.2184008889728122
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,7168,8192,0.15334310796525744
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,7168,7168,0.13977332909901938
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,7168,7168,0.13465600543551975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,7168,6144,0.12063288688659668
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,7168,7168,0.19447644551595053
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,7168,6144,0.11643732918633355
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,7168,6144,0.16897688971625435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,7168,4096,0.08303378025690715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,7168,5120,0.09860799709955852
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,7168,5120,0.10147733158535427
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,7168,5120,0.14467555946773952
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,7168,4096,0.12180266115400527
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,7168,3584,0.07639644543329875
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,7168,4096,0.08026755518383451
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,7168,3072,0.06740978029039171
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,7168,3584,0.11052888631820679
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,7168,3584,0.07142133182949491
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,7168,3072,0.09809333086013794
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,7168,2560,0.05954755677117241
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,7168,3072,0.06196710798475477
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,7168,2048,0.04390933447413974
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,7168,2560,0.08759199910693699
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,7168,2048,0.04895377821392483
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,7168,1536,0.03441866570048862
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,7168,2560,0.05345244540108574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,7168,65536,1.2739458084106445
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,7168,2048,0.07433511151207818
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,7168,1536,0.041914665036731295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,7168,65536,1.560872925652398
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,7168,768,0.0203813330994712
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,7168,1024,0.024866667058732774
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,7168,1024,0.051924443907207914
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,7168,512,0.015574221809705099
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,7168,1536,0.0641546646753947
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,7168,1024,0.03243555625279745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,7168,768,0.02889066603448656
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,7168,768,0.046015110280778676
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,7168,256,0.011144888897736868
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,7168,512,0.025388444463411968
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,7168,256,0.020597333709398907
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,7168,512,0.040896889236238264
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,7168,128,0.008203555312421586
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,7168,256,0.035678221119774714
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,7168,128,0.01940711173746321
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,7168,64,0.007199111084143321
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,7168,128,0.03135822216669718
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,7168,64,0.019167999426523846
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,7168,32,0.007520889242490132
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,7168,32,0.019310222731696237
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,6144,16384,0.284858677122328
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,6144,12288,0.17326133781009248
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,7168,65536,1.2230346467759874
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,6144,16384,0.22703467475043404
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,6144,12288,0.20722311072879365
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,6144,10240,0.1755297713809543
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,6144,10240,0.1469848950703939
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,6144,8192,0.11727289358774821
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,6144,8192,0.1461884445614285
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,6144,65536,1.1155306498209636
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,6144,7168,0.1037599974208408
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,6144,7168,0.12432711654239231
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,6144,6144,0.08922488821877374
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,6144,6144,0.10597688621944851
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,6144,65536,0.9752648671468099
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,6144,5120,0.07739555835723877
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,6144,5120,0.09163377682367961
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,6144,16384,0.3614968988630507
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,6144,12288,0.2796951134999593
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,6144,4096,0.06200622187720405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,6144,4096,0.07374488645129733
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,6144,10240,0.23776443799336752
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,6144,3584,0.055099556843439736
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,6144,3584,0.06730400191413032
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,6144,3072,0.04840355449252658
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,6144,3072,0.05950399902131823
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,6144,65536,1.3678649266560872
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,6144,2560,0.041843556695514254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,6144,8192,0.19411910904778373
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,6144,2560,0.05369777811898125
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,6144,7168,0.1711262199613783
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,6144,2048,0.03436355458365546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,6144,2048,0.043988443083233304
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,6144,1536,0.027368000812000696
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,6144,1536,0.03717599974738227
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,6144,6144,0.14842132727305093
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,6144,1024,0.019878221882714164
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,6144,1024,0.03000977635383606
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,6144,5120,0.1293813387552897
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,6144,768,0.016516443755891588
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,6144,768,0.026527111728986103
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,6144,3584,0.09854044516881307
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,6144,512,0.012996444271670448
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,6144,512,0.022618666291236877
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,6144,4096,0.10864444573720296
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,6144,256,0.008799110849698385
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,6144,256,0.01907111042075687
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,6144,2560,0.07881689071655273
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,6144,128,0.007141333487298753
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,6144,128,0.017921777235137094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,6144,1536,0.05710133579042223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,6144,64,0.00611111107799742
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,6144,64,0.0181351105372111
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,6144,32,0.006462222172154321
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,6144,32,0.018073777357737224
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,6144,768,0.04264977905485365
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,6144,3072,0.08777955505583022
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,6144,2048,0.0671644475724962
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,5120,16384,0.19073866473303902
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,6144,1024,0.04711911082267761
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,5120,16384,0.27726221084594727
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,5120,65536,1.088388442993164
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,5120,12288,0.14447466532389322
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,5120,12288,0.20787199338277182
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,5120,10240,0.12163733111487494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,5120,65536,0.8470222155253092
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,5120,10240,0.17521511183844674
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,5120,8192,0.0985484454366896
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,5120,8192,0.14342933230929905
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,6144,128,0.029839111698998347
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,6144,256,0.03274844421280755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,5120,7168,0.08612800306744045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,5120,7168,0.1226302252875434
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,5120,6144,0.07513422436184354
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,5120,6144,0.10536622338824803
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,6144,512,0.038387556870778404
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,5120,5120,0.06379022200902303
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,5120,5120,0.08975378010008071
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,5120,16384,0.343064890967475
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,5120,4096,0.052015999952952065
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,5120,4096,0.07340266969468859
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,5120,65536,1.2886853747897677
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,5120,12288,0.2647706667582194
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,5120,3584,0.0672524438963996
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,5120,3584,0.04651022288534376
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,5120,3072,0.040656887822681
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,5120,3072,0.05982933441797892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,5120,10240,0.22454489601982963
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,5120,2560,0.03525333272086249
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,5120,2560,0.05240799983342489
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,5120,8192,0.18383467197418213
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,5120,2048,0.02895911203490363
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,5120,2048,0.04376266731156243
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,5120,7168,0.16339911354912653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,5120,1536,0.022882666852739122
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,5120,6144,0.14335466755761042
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,5120,1536,0.036723554134368896
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,5120,1024,0.017264000243610807
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,5120,1024,0.02972444560792711
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,5120,5120,0.12361599339379205
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,5120,768,0.014327999618318347
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,5120,768,0.02623288830121358
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,5120,3584,0.09484355317221747
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,5120,512,0.011316444310877057
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,5120,512,0.021920889616012573
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,5120,4096,0.10342578093210857
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,5120,256,0.008086221913496653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,5120,256,0.0192275560564465
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,5120,3072,0.08453066481484307
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,5120,128,0.006632888896597757
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,5120,128,0.018008000320858426
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,5120,2560,0.07601422071456909
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,5120,64,0.005752888818581899
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,5120,64,0.0181404451529185
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,5120,32,0.006025777922736273
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,5120,32,0.01800622211562263
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,5120,768,0.041843556695514254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,5120,2048,0.0648017790582445
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,5120,1536,0.05531644490030077
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,4096,65536,0.9261350631713867
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,4096,65536,0.5932933489481608
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,4096,16384,0.15139467186397976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,4096,16384,0.2421555519104004
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,4096,16384,0.3180026743147108
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,5120,1024,0.04554933309555054
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,4096,12288,0.17835822370317247
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,4096,12288,0.11540533436669244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,4096,12288,0.24584354294670951
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,4096,65536,1.1843858295016818
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,4096,10240,0.1514213350084093
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,4096,10240,0.09651288721296523
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,4096,8192,0.12363910675048828
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,4096,8192,0.1703324450386895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,4096,10240,0.20922311147054037
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,4096,8192,0.07922399706310697
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,4096,7168,0.06896089182959662
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,4096,7168,0.10638666815227932
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,4096,6144,0.09178666936026679
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,4096,7168,0.1512897809346517
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,4096,6144,0.1323688824971517
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,4096,5120,0.0779235561688741
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,4096,5120,0.11403111616770427
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,4096,5120,0.050986667474110924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,4096,4096,0.0648968882030911
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,4096,6144,0.06031822495990329
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,4096,3584,0.037328001525666975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,4096,4096,0.0963111122449239
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,4096,4096,0.04192444350984362
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,4096,3584,0.059207108285692006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,4096,3072,0.0329146683216095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,4096,3584,0.08792355325486924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,4096,3072,0.05239822136031257
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,4096,2560,0.028575999869240656
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,4096,3072,0.0793022248480055
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,4096,2560,0.04626933402485318
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,4096,2048,0.023394667439990576
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,4096,2048,0.03859022259712219
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,4096,1536,0.019123555885420907
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,4096,2560,0.07113333543141682
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,4096,2048,0.060495111677381724
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,4096,1536,0.032800889677471586
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,4096,1024,0.014341332846217685
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,4096,1536,0.05197066730923123
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,4096,1024,0.02626577847533756
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,4096,768,0.011965333587593503
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,4096,1024,0.04332888788647122
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,4096,768,0.023056889573733013
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,4096,512,0.009568000005351173
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,4096,768,0.03931822379430135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,4096,512,0.02011999984582265
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,4096,256,0.006989333364698622
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,4096,256,0.01740800009833442
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,4096,512,0.033915556139416166
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,4096,128,0.0058444444504049085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,4096,256,0.03085600005255805
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,4096,128,0.01643822259373135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,4096,64,0.005116444494989184
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,4096,128,0.02896266513400608
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,4096,32,0.005340444544951121
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,4096,64,0.016506666938463848
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,4096,32,0.016415110892719693
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3584,16384,0.2375048796335856
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3584,16384,0.13514133294423422
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,3584,16384,0.31975822978549534
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,5120,256,0.03319911162058512
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,5120,128,0.029472000069088403
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3584,12288,0.10289511415693496
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3584,12288,0.17757511138916016
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,3584,12288,0.24524799982706705
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3584,10240,0.15054221947987875
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3584,10240,0.08409777614805435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3584,8192,0.12310755252838135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,3584,10240,0.2099857727686564
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,3584,8192,0.17109689447614881
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3584,8192,0.0692488882276747
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3584,65536,0.9067351023356119
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3584,7168,0.10610044664806789
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3584,6144,0.09156888723373413
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,3584,65536,1.1798061794704862
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3584,7168,0.06053777535756429
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,3584,7168,0.15152622593773737
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,3584,6144,0.13292266262902155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3584,6144,0.05245955453978645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3584,5120,0.07772444354163276
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3584,4096,0.036271111832724676
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3584,4096,0.0641440020667182
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,3584,5120,0.11471377478705512
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3584,3584,0.03247733248604669
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,3584,4096,0.09627377986907959
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3584,3584,0.058837334314982094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3584,5120,0.044841776291529335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3584,3072,0.02830222249031067
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,3584,3584,0.08868622117572361
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3584,3072,0.052411556243896484
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3584,2560,0.024057777391539678
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,3584,3072,0.07954755756590101
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3584,2560,0.045929776297675244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3584,2048,0.020279111133681405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3584,2048,0.03849155704180399
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,3584,2560,0.07130489084455702
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3584,1536,0.01626488897535536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,3584,2048,0.0609875586297777
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3584,1536,0.032871110571755305
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3584,65536,0.5860071182250977
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3584,1024,0.012103999654452005
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,3584,1536,0.05235200126965841
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3584,1024,0.025801777839660645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3584,768,0.00996800015370051
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,3584,1024,0.04425866736306084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3584,768,0.02271911170747545
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3584,512,0.007673777639865875
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,3584,768,0.03973066806793213
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3584,512,0.01995644470055898
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3584,256,0.005941333456171884
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,3584,512,0.03467199868626065
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3584,128,0.005064888960785336
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,3584,256,0.03067200051413642
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3584,256,0.017648888958825003
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3584,64,0.004881777697139316
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3584,128,0.016546666622161865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3584,32,0.0052915554907586836
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,5120,512,0.036770668294694685
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3584,64,0.01646933290693495
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,3584,128,0.029231998655531142
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3584,32,0.016507556041081745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3072,16384,0.23658845159742567
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3072,12288,0.08849422136942546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3072,16384,0.11615910794999863
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,3072,16384,0.2941039933098687
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3072,12288,0.17797511153750947
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3072,10240,0.1497431066301134
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3072,8192,0.1232080062230428
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,3072,12288,0.22938312424553764
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3072,10240,0.07435288694169787
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,3072,10240,0.19500888718499076
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3072,7168,0.10567555824915568
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3072,8192,0.06081777811050415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,3072,8192,0.1586488882700602
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,3072,7168,0.1399431096182929
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3072,6144,0.09096978108088176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3072,7168,0.05366222063700358
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3072,5120,0.039557334449556135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,3072,6144,0.1224222183227539
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3072,65536,0.9020275539822049
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3072,6144,0.04664444592263964
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3072,5120,0.07788622379302979
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3072,4096,0.03259822063975864
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3072,4096,0.06428355640835233
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3072,3584,0.02924799919128418
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,3072,5120,0.10619644323984782
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,3072,4096,0.08992444144354926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3072,3584,0.058966219425201416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3072,3072,0.025428444147109985
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,3072,3584,0.08234844605127971
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3072,3072,0.052059554391437106
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3072,2560,0.022026666336589392
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3072,2560,0.04551555713017782
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,3072,3072,0.07448533508512709
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3072,2048,0.01854311095343696
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,3072,2560,0.06676089101367526
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3072,2048,0.03864533371395535
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3072,1536,0.015171556009186639
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,3072,2048,0.0570639967918396
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,3072,65536,1.0831449296739366
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3072,1536,0.032864888509114586
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3072,1024,0.011492444409264458
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,3072,1536,0.04987377921740214
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3072,1024,0.02581600017017788
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3072,768,0.00976533359951443
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,3072,1024,0.041499555110931396
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3072,768,0.022338666849666174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3072,512,0.007567110988828871
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,3072,768,0.03570311268170675
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3072,512,0.020093333390023973
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3072,256,0.005821333163314396
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,3072,512,0.03285244438383315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3072,256,0.017747556169827778
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3072,128,0.004971555537647671
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,3072,256,0.02975822157329983
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3072,128,0.01640977793269687
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3072,64,0.004381333374314838
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,3072,128,0.02849066588613722
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3072,32,0.004655111167165968
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3072,64,0.01642933322323693
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,3072,32,0.01639111174477471
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2560,16384,0.23376001252068412
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,3072,65536,0.4498888651529948
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,2560,16384,0.28594933615790474
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2560,16384,0.10141422351201375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2560,12288,0.17679644955529106
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2560,12288,0.07670222388373481
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2560,10240,0.1499119997024536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,2560,12288,0.22188622421688506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,2560,10240,0.1895973285039266
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2560,8192,0.12238132953643799
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2560,10240,0.06519466638565063
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,2560,8192,0.15398755338456896
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2560,8192,0.05305244525273641
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2560,7168,0.10573599735895793
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2560,7168,0.047228445609410606
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2560,6144,0.04119999872313605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,2560,7168,0.13673333326975504
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2560,6144,0.09107200304667155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2560,5120,0.03534933262401157
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,2560,6144,0.11977156003316243
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2560,65536,0.8961768680148654
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2560,5120,0.07717244492636786
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2560,4096,0.029209779368506536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,2560,5120,0.10362044307920669
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2560,4096,0.06381422281265259
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2560,3584,0.026128000683254663
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2560,3584,0.05848177936342028
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,2560,4096,0.08824177583058675
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2560,3072,0.022919111781650122
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2560,3072,0.051687111457188926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,2560,3584,0.08128177457385592
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2560,2560,0.020021332634819877
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,2560,3072,0.0728915532430013
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2560,2048,0.01704355577627818
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,2560,65536,1.047943115234375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2560,2560,0.04499377806981405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2560,1536,0.013987556099891663
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,2560,2560,0.06597244739532471
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2560,2048,0.03788977861404419
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,2560,2048,0.05663822094599406
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2560,1536,0.03251733382542928
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2560,1024,0.010787555740939246
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2560,768,0.009229333036475712
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,2560,1536,0.04956266615125868
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2560,1024,0.024591111474566992
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,2560,1024,0.040720889965693154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2560,768,0.02200800014866723
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2560,512,0.0075128889746136135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2560,256,0.005716444303592046
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,2560,768,0.03510488735304938
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2560,512,0.019736889335844252
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2560,65536,0.39048088921440977
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,2560,512,0.031911998987197876
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2560,256,0.017215111189418368
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2560,128,0.004869333157936732
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,2560,256,0.02911111050181919
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2560,128,0.016213332613309223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2560,64,0.00432977784011099
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2560,32,0.004602666530344221
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,2560,128,0.028023110495673284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2560,64,0.016138666205936007
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2560,32,0.016168889072206285
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2048,12288,0.14954400062561035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2048,12288,0.059544000360700816
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2048,16384,0.19555111726125082
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2048,16384,0.0783804456392924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2048,10240,0.12712621688842773
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2048,10240,0.05054577853944567
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2048,8192,0.10416177908579509
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2048,8192,0.04099466734462314
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2048,7168,0.03637333379851447
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2048,7168,0.08986844619115193
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2048,65536,0.7383057806226941
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2048,6144,0.03226222263442145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2048,6144,0.0775573319858975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2048,65536,0.3047991063859728
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2048,5120,0.029527111185921565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2048,5120,0.06746666961246066
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,2048,16384,0.275673786799113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2048,4096,0.022806222240130108
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,2048,12288,0.21340355608198378
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2048,4096,0.05537066525883145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2048,3584,0.020503999458418954
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2048,3584,0.05074488785531786
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,2048,10240,0.18178489473130968
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2048,3072,0.018294221825069852
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,2048,8192,0.1482142210006714
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2048,3072,0.04537777768241035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2048,2560,0.015821332732836407
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2048,2560,0.040151112609439425
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,2048,7168,0.13057243824005127
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2048,2048,0.013374221821626028
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,2048,6144,0.1144711176554362
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,2048,65536,1.009454197353787
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2048,2048,0.033406221204333834
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2048,1536,0.011094222466150919
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2048,1536,0.029190222422281902
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2048,1024,0.008481777376598781
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2048,1024,0.0226248883538776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,2048,5120,0.09905599885516697
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2048,768,0.007106666763623555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2048,768,0.020384889509942796
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,2048,3584,0.07776444488101535
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,2048,4096,0.08454489045672947
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2048,512,0.006027555714050929
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2048,512,0.018171555466122098
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2048,256,0.004875555634498596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2048,256,0.016102222932709586
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,2048,3072,0.07070222165849474
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2048,128,0.004275555411974589
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2048,128,0.015015999476114908
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,2048,2560,0.06263733572430082
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2048,64,0.0038515557017591265
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,2048,1536,0.04732799861166212
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2048,64,0.014992000328169929
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,2048,2048,0.054233776198493115
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,2048,32,0.004046222401989831
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,2048,32,0.014908444550302295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,2048,1024,0.03806844353675842
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,2048,768,0.03425511055522495
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1536,16384,0.059009777175055615
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1536,16384,0.19537333647410074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1536,65536,0.7309208975897895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1536,65536,0.2350675529903836
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1536,12288,0.15004621611701116
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1536,12288,0.04717333449257744
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1536,10240,0.03882133298450046
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1536,10240,0.12726222144232854
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1536,8192,0.03201777736345927
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1536,8192,0.10331555869844224
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,2048,512,0.03102488981352912
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,2048,256,0.028679112593332928
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1536,7168,0.02824977702564663
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1536,7168,0.09003377623028225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,2048,128,0.027079110344250996
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1536,6144,0.024694222542974684
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1536,6144,0.07731111182106866
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1536,5120,0.021228444245126512
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1536,5120,0.0661697785059611
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,1536,10240,0.18004177676306832
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,1536,16384,0.2724826600816515
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1536,4096,0.017880888448821176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,1536,12288,0.2108880016538832
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1536,4096,0.05494577685991923
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1536,3584,0.01605422298113505
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1536,3584,0.05043288734224108
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1536,3072,0.014316444595654806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,1536,65536,0.9990604188707141
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1536,3072,0.044630222850375705
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1536,2560,0.012607999973826937
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1536,2560,0.038839999172422625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,1536,8192,0.14674399958716497
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1536,2048,0.010876444478829702
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1536,2048,0.03335466649797227
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,1536,7168,0.13033955627017552
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1536,1536,0.00886933339966668
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1536,1536,0.028566221396128338
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,1536,6144,0.11363111601935492
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1536,1024,0.006909333169460297
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1536,1024,0.02221244408024682
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,1536,5120,0.09848533074061076
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1536,768,0.006082666830884085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1536,768,0.02018666598531935
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,1536,3584,0.0772435532675849
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1536,512,0.005231110999981563
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1536,512,0.018039110634062026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,1536,2560,0.06238489018546211
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1536,256,0.004354666504595014
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1536,256,0.015771556231710646
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,1536,4096,0.08427466948827107
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1536,128,0.003921777837806278
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1536,128,0.014901333385043673
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,1536,3072,0.07003288798862033
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1536,64,0.003563555578390757
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1536,64,0.014848889576064216
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1536,32,0.003726222034957674
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1536,32,0.01478666729397244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,1536,2048,0.05424799852901035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1024,65536,0.15843733151753744
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1024,65536,0.7309084468417697
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1024,16384,0.04757422208786011
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1024,16384,0.19570933447943795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,1536,768,0.03449066811137729
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1024,12288,0.03812888926929898
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1024,12288,0.14978488286336264
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,1536,1536,0.047472887569003634
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1024,10240,0.02907022171550327
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1024,10240,0.12698488765292698
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,1536,1024,0.03741600116093954
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1024,8192,0.028689778513378564
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1024,8192,0.10221155484517415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,1536,128,0.02724444369475047
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1024,7168,0.02181422213713328
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1024,7168,0.0899511112107171
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,1536,256,0.02819288770357768
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1024,6144,0.019698666201697457
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1024,6144,0.07736977603700426
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,1536,512,0.030755556292004053
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1024,5120,0.017824888229370117
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1024,5120,0.0655848913722568
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,1024,16384,0.2600657675001356
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1024,4096,0.016062221593326993
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1024,4096,0.054647111230426364
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,1024,65536,0.9476213455200195
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1024,3584,0.013674666484196981
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1024,3584,0.05040622088644239
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,1024,12288,0.2000159952375624
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1024,3072,0.011672000090281168
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1024,3072,0.04388977752791511
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,1024,10240,0.1710817813873291
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1024,2560,0.010553778045707278
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1024,2560,0.03926400012440152
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,1024,8192,0.13935289118025038
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1024,2048,0.008450667063395182
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,1024,7168,0.12346133920881484
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1024,2048,0.032776888873842024
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1024,1536,0.007167999943097432
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1024,1536,0.02699111070897844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,1024,6144,0.10850844118330215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1024,1024,0.00573333352804184
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1024,1024,0.022055110997623865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,1024,5120,0.09436977571911281
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1024,768,0.0052888890107472735
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1024,768,0.020075556304719713
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,1024,4096,0.0808968875143263
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1024,512,0.004752000172932942
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,1024,3584,0.07406400309668647
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1024,512,0.017852443787786696
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1024,256,0.004056888735956616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1024,256,0.015726221932305228
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,1024,3072,0.06700088580449422
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1024,128,0.00361333332127995
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1024,128,0.014767999450365702
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,1024,2560,0.05971466832690769
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1024,64,0.003351111171974076
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1024,64,0.014697778556081982
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,1024,32,0.0034888887570963967
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,1024,32,0.014412444498803882
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,768,65536,0.12335022290547688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,768,65536,0.7305742369757758
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,1024,2048,0.05200799968507555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,768,16384,0.037660443120532565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,768,16384,0.19559021790822348
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,1024,1536,0.04541511005825467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,768,12288,0.030236442883809406
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,768,12288,0.1483395497004191
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,1024,1024,0.03647289011213515
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,768,10240,0.02530844509601593
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,768,10240,0.12664000193277994
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,1024,768,0.03313688768280877
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,768,8192,0.022832888695928786
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,1024,256,0.028145776854621038
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,768,8192,0.10255999697579278
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,768,7168,0.020484444167878892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,1024,512,0.02993333339691162
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,768,7168,0.08963466352886623
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,1024,128,0.026224889689021643
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,768,6144,0.016335111525323655
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,768,6144,0.07708622349633111
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,768,5120,0.014463999205165438
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,768,5120,0.06561866733762953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,768,65536,0.9503688812255859
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,768,4096,0.014189332723617554
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,768,4096,0.054655111498302884
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,768,16384,0.2602764500512017
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,768,3584,0.011483555866612328
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,768,3584,0.049829334020614624
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,768,12288,0.1999537812338935
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,768,3072,0.010094222095277574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,768,3072,0.04368444283803304
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,768,10240,0.17098132769266763
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,768,2560,0.008921778036488427
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,768,2560,0.038655999634000994
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,768,8192,0.14022578133477107
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,768,2048,0.007500444849332173
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,768,2048,0.03256266646915012
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,768,7168,0.12358044253455268
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,768,1536,0.006239999913507038
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,768,6144,0.10836711194780137
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,768,1536,0.0265964451763365
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,768,1024,0.005129777722888523
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,768,1024,0.022016000416543748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,768,5120,0.09411822424994574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,768,768,0.0047484445903036325
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,768,768,0.019913777709007263
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,768,4096,0.08064711093902588
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,768,512,0.004239999999602635
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,768,512,0.017793777916166518
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,768,3584,0.07332444190979004
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,768,256,0.0037208890749348533
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,768,256,0.015770667129092746
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,768,3072,0.0664551125632392
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,768,2560,0.05908799833721585
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,768,128,0.0033946666452619764
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,768,128,0.01462933255566491
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,768,64,0.003136000078585413
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,768,768,0.03279022375742594
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,768,64,0.014567999376191033
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,768,32,0.0032755554550223877
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,768,32,0.014384888940387301
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,512,65536,0.08432088957892524
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,512,65536,0.7289822366502551
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,512,16384,0.029301332102881536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,768,1536,0.044065776798460216
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,512,16384,0.19595466719733345
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,512,12288,0.024671110841963027
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,768,2048,0.05152177810668945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,512,12288,0.1503866646024916
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,512,10240,0.022777777579095628
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,512,10240,0.12666489018334284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,768,1024,0.03584355447027419
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,512,8192,0.0199617776605818
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,512,8192,0.1025164458486769
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,768,512,0.029804444975323145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,512,7168,0.015736889508035447
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,512,7168,0.08952622281180488
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,768,256,0.02739199995994568
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,512,6144,0.013316444224781461
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,512,6144,0.07687289184994169
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,768,128,0.026310222016440496
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,512,5120,0.011759999725553723
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,512,5120,0.06479644775390625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,512,16384,0.24848267767164442
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,512,4096,0.010337777435779572
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,512,4096,0.054281777805752225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,512,65536,0.9073315726386176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,512,12288,0.19223288695017496
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,512,3584,0.009135999613338048
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,512,3584,0.04976977904637655
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,512,3072,0.008376888930797577
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,512,10240,0.16350221633911133
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,512,3072,0.04343733191490173
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,512,2560,0.007473777565691207
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,512,2560,0.03855822152561612
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,512,8192,0.13442844814724392
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,512,2048,0.006440889090299606
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,512,2048,0.03194488750563727
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,512,7168,0.11915289031134711
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,512,1536,0.005768888940413793
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,512,1536,0.0266266663869222
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,512,6144,0.10400266779793634
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,512,1024,0.004839111119508743
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,512,1024,0.021928000781271193
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,512,5120,0.09070399734708999
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,512,768,0.004425777743260066
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,512,768,0.01958933307064904
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,512,4096,0.07765511009428236
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,512,3584,0.07130310932795207
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,512,512,0.003991999973853429
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,512,512,0.01753155555990007
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,512,256,0.0035706667436493766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,512,256,0.015468445089128284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,512,3072,0.06454488966200086
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,512,128,0.0032746667663256326
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,512,128,0.014244443840450711
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,512,2560,0.05760888920889961
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,512,64,0.0031235555393828284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,512,64,0.014185777968830533
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,512,32,0.0033155555526415506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,512,32,0.013968888256284924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,256,65536,0.052856889035966664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,256,65536,0.7304862340291342
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,512,2048,0.050428443484836154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,256,16384,0.021700445148679946
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,256,16384,0.19461333751678467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,512,1536,0.04350577791531881
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,256,12288,0.017091555727852713
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,256,12288,0.14956088860829672
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,512,1024,0.0351235568523407
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,256,10240,0.015430221954981485
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,512,768,0.032065778970718384
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,256,10240,0.12570488452911377
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,256,8192,0.013350221845838757
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,256,8192,0.10314844714270698
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,512,512,0.02923999892340766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,512,256,0.02747111188040839
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,256,7168,0.012131555212868584
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,256,7168,0.08958133061726888
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,256,6144,0.01110400011142095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,256,6144,0.07643555270300971
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,512,128,0.026389333936903212
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,256,5120,0.010282666318946415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,256,5120,0.06454488966200086
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,256,65536,0.8920550876193576
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,256,4096,0.009505777723259395
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,256,4096,0.05399466554323832
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,256,16384,0.24371732605828178
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,256,3584,0.00812266684240765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,256,3584,0.049045334259668984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,256,12288,0.1893999973932902
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,256,3072,0.007320000065697565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,256,3072,0.04296355446179708
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,256,10240,0.1613182226816813
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,256,2560,0.007302222152551015
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,256,8192,0.13254933887057835
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,256,2560,0.03793777690993415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,256,2048,0.006626666833957036
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,256,2048,0.030607110924190942
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,256,7168,0.11728266874949138
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,256,1536,0.005850666513045629
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,256,6144,0.10232444604237874
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,256,1536,0.02625688910484314
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,256,1024,0.004860444615284602
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,256,1024,0.02165066699186961
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,256,5120,0.08928000264697605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,256,768,0.00442755553457472
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,256,768,0.019523556033770244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,256,4096,0.07672622468736437
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,256,512,0.003882666842805015
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,256,512,0.017300445172521803
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,256,3584,0.06980622477001615
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,256,256,0.003395555540919304
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,256,256,0.015302222636010913
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,256,3072,0.06338400310940213
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,256,128,0.003136000078585413
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,256,128,0.014188443620999655
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,256,2560,0.056484447585211865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,256,64,0.002921777880854077
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,256,2048,0.04892622099982368
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,256,64,0.01418133411142561
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,256,32,0.003003555453485913
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,256,32,0.013918222652541267
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,128,65536,0.04421866602367825
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,128,65536,0.7307955423990885
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,128,16384,0.012772444221708508
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,256,1536,0.040807111395729914
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,128,16384,0.1956391069624159
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,128,12288,0.010670222342014313
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,128,12288,0.1499306625790066
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,256,1024,0.0347360008292728
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,128,10240,0.00983911090426975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,128,10240,0.1267306672202216
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,256,768,0.031792001591788396
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,128,8192,0.009860444400045607
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,128,8192,0.10253866513570149
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,256,512,0.028974221812354192
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,128,7168,0.00945955514907837
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,128,7168,0.08951733509699504
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,256,256,0.02715911136733161
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,128,6144,0.008755555583371056
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,128,6144,0.076255997021993
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,256,128,0.026408889227443274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,128,5120,0.008294222255547842
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,128,5120,0.06475733386145698
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,128,65536,0.886897775861952
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,128,4096,0.007673777639865875
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,128,4096,0.05395199855168661
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,128,16384,0.24367732471889922
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,128,3584,0.008367111285527548
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,128,3584,0.04896355668703715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,128,12288,0.1882693370183309
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,128,3072,0.007678222325113084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,128,3072,0.04321066538492838
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,128,10240,0.16078221797943115
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,128,2560,0.0064702220261096954
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,128,2560,0.03645955522855123
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,128,8192,0.13155821959177652
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,128,2048,0.005743999862008625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,128,2048,0.030384000804689195
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,128,7168,0.11729689439137776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,128,1536,0.005072888814740711
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,128,1536,0.02622311148378584
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,128,6144,0.10224711232715183
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,128,1024,0.004334222111437055
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,128,1024,0.021538666552967493
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,128,5120,0.08963733249240452
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,128,768,0.004005333201752769
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,128,768,0.019457777341206867
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,128,4096,0.07598577605353461
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,128,512,0.003581333491537306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,128,512,0.01721955504682329
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,128,3584,0.06937422355016072
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,128,256,0.0031759999692440033
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,128,256,0.01517777807182736
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,128,3072,0.06278222137027316
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,128,128,0.002959111084540685
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,128,128,0.014122666584120857
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,128,2560,0.05618311299218071
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,128,64,0.0028000000036425064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,128,64,0.013980444934633044
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,128,32,0.0028640000770489373
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,128,32,0.01386044422785441
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,64,65536,0.0448035564687517
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,128,2048,0.046911110480626426
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,64,16384,0.011728000309732227
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,64,16384,0.19507200188106963
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,64,12288,0.010196444061067369
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,64,65536,0.7302897771199545
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,64,10240,0.00944622192117903
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,64,12288,0.14955022599962023
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,64,8192,0.008795555267069075
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,64,10240,0.1265653371810913
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,64,7168,0.00833511104186376
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,64,8192,0.10223111179139878
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,64,6144,0.007837333612971837
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,64,7168,0.08934755457772149
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,64,5120,0.007373332977294922
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,64,6144,0.07633244329028659
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,64,4096,0.006963555183675554
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,64,5120,0.0645351078775194
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,64,3584,0.007266666326257918
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,64,4096,0.053844445281558566
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,128,1536,0.04057777921358744
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,64,3072,0.0069413334131240845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,64,3584,0.048659556441836886
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,64,2560,0.006232000059551663
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,64,3072,0.04265422291225857
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,64,2048,0.005531555662552516
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,64,2560,0.03663377629386054
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,64,2048,0.030478222502602473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,64,1536,0.005565333283609814
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,64,1536,0.026392888691690233
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,64,1024,0.004592888885074192
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,64,768,0.004176889028814104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,64,1024,0.021593777669800654
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,64,512,0.0037360000941488477
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,64,768,0.019476445184813607
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,64,256,0.0033608888172441055
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,64,512,0.01732711162832048
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,64,128,0.003063111048605707
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,64,256,0.01516266663869222
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,64,128,0.014098667436175875
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,64,64,0.002979555477698644
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,64,32,0.003068444422549672
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,64,64,0.013903111219406128
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,64,32,0.013855111267831592
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,32,65536,0.043064001533720225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,32,16384,0.01089511149459415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,32,16384,0.1948853333791097
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,32,12288,0.009499555660618676
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,32,65536,0.7299911181131998
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,32,10240,0.008581333690219456
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,32,12288,0.14942488405439588
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,32,8192,0.007780444290902879
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,32,10240,0.12670844131045872
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,32,7168,0.007245333658324347
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,32,8192,0.10214400291442871
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,32,6144,0.006647111227114995
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,32,7168,0.08956533008151585
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,32,5120,0.007045333584149678
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,32,6144,0.07607111003663805
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,32,4096,0.006663110935025745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,32,5120,0.0648604432741801
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,32,4096,0.053808887799580894
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,32,3584,0.007236444287829929
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,32,3584,0.049053334527545504
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,32,3072,0.00686488880051507
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,32,2560,0.006200888918505774
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,32,3072,0.04295466674698723
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,32,2560,0.037230223417282104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,32,2048,0.005488889084921942
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,32,1536,0.004807999978462855
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,32,2048,0.030504888958401147
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,32,1024,0.004087111188305749
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,32,1536,0.0262737770875295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,32,768,0.0037537775933742523
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,32,1024,0.02156977852185567
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,32,512,0.0033946666452619764
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,32,768,0.019549333386951022
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,32,256,0.003043555551105075
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,32,512,0.01718488832314809
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,128,1024,0.03460089034504361
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,32,128,0.0028479999552170434
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,32,256,0.015141333142916361
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,32,128,0.014128888646761576
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,32,64,0.0026986667265494666
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,32,64,0.013854222165213691
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1024,32,32,0.0027155555370781156
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1024,32,32,0.013831111292044321
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,128,768,0.03179377648565505
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,128,512,0.02894222074084812
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,128,256,0.027038221557935078
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1024,128,128,0.026179555389616225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,65536,12288,0.8537315792507596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,65536,16384,1.180574205186632
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,65536,16384,1.3792311350504558
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,65536,10240,0.721225791507297
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,65536,12288,1.0185164345635307
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,65536,12288,1.5624230702718098
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,65536,10240,1.2776710722181532
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,65536,10240,0.8404942088656955
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,65536,8192,0.5879715813530816
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,65536,7168,0.518580436706543
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,65536,8192,0.8926817576090494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,65536,8192,0.6863110860188802
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,65536,7168,0.8037938012017144
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,65536,6144,0.42771199014451766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,65536,7168,0.5910319752163357
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,65536,6144,0.5030346446567112
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,65536,5120,0.36200533972846133
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,65536,6144,0.7097795804341634
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,65536,16384,2.0451422797309027
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,65536,4096,0.2896897792816162
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,65536,5120,0.4327369001176622
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,65536,5120,0.5893146726820204
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,65536,3584,0.2615279886457655
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,65536,4096,0.3472782240973578
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,65536,4096,0.46205419964260525
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,65536,3072,0.222304900487264
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,65536,3584,0.3167688846588135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,65536,3072,0.3528177738189697
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,65536,3072,0.2790542178683811
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,65536,2560,0.19067111280229357
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,65536,2560,0.23372801144917807
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,65536,1536,0.12289422088199192
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,65536,2048,0.153056886461046
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,65536,2560,0.2965759966108534
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,65536,2048,0.19864000214470756
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,65536,3584,0.4075004524654812
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,65536,1536,0.18743022282918295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,65536,2048,0.2424586613972982
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,65536,1536,0.16026043891906738
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,65536,1024,0.09200711382759942
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,65536,768,0.07843289110395643
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,65536,1024,0.13362844785054526
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,65536,1024,0.12050044536590576
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,65536,768,0.1058960027164883
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,65536,512,0.06590755780537923
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,65536,256,0.0562302205297682
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,65536,768,0.10292444626490276
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,65536,512,0.07909866836335924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,65536,128,0.041768001185523145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,65536,256,0.06711022059122722
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,65536,64,0.03783555494414435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,65536,256,0.05119466781616211
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,65536,128,0.05197244551446703
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,65536,128,0.06493688954247369
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,65536,32,0.037941333320405744
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,65536,64,0.052206221554014415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,65536,32,0.053187555736965604
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,65536,512,0.08216000265545316
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,16384,16384,0.3712915579477946
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,16384,16384,0.4527280065748427
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,16384,12288,0.2681235472361247
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,16384,12288,0.33917866812811953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,16384,16384,0.44849157333374023
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,16384,12288,0.3445759879218207
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,16384,10240,0.23009954558478463
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,16384,10240,0.28459644317626953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,16384,8192,0.18895467122395834
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,16384,10240,0.2922542095184326
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,16384,8192,0.23014667299058703
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,16384,7168,0.16278933154212102
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,16384,8192,0.23921688397725424
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,16384,7168,0.20286044809553358
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,16384,6144,0.1390613317489624
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,16384,7168,0.21247555149926078
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,16384,6144,0.18677600224812826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,16384,6144,0.17514755990770128
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,16384,5120,0.11960266696082221
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,16384,65536,1.5417102177937825
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,16384,5120,0.14800355169508192
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,16384,5120,0.16186932722727457
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,16384,4096,0.09853955772187974
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,16384,4096,0.1345813274383545
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,16384,3584,0.0897128913137648
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,16384,4096,0.119741333855523
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,16384,3584,0.10644355747434829
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,16384,3072,0.07965777979956733
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,16384,3584,0.12288088268703884
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,16384,3072,0.09303022093243069
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,16384,3072,0.10830844773186578
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,16384,2560,0.06934844122992621
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,16384,2560,0.0956382221645779
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,16384,2048,0.058816889921824135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,16384,2560,0.07945422331492107
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,16384,1536,0.049341334236992725
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,16384,2048,0.08364444308810765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,16384,1536,0.07104000118043688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,16384,2048,0.06600266695022583
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,16384,1024,0.03863733344607883
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,16384,1536,0.05211288730303446
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,16384,768,0.031735999716652766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,16384,1024,0.0558640029695299
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,16384,768,0.03359555535846286
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,16384,512,0.02438666588730282
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,16384,512,0.02960444490114848
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,16384,768,0.051078223519855075
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,16384,1024,0.03827733463711209
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,16384,65536,1.8422133127848308
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,16384,256,0.017216889394654166
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,16384,512,0.044491555955674916
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,16384,256,0.025272000167104933
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,16384,128,0.013325333595275879
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,16384,256,0.03946400019857619
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,16384,128,0.02349511120054457
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,16384,64,0.01164533363448249
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,16384,64,0.022663111488024395
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,16384,32,0.012050666742854647
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,16384,32,0.022952889402707417
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,16384,128,0.03672000103526645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,12288,16384,0.323853333791097
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,12288,12288,0.2526506582895915
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,12288,12288,0.23837155765957305
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,12288,16384,0.33613154623243546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,12288,10240,0.20183200306362578
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,12288,10240,0.211371554268731
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,16384,65536,2.0141128963894315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,12288,8192,0.17055200205908883
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,12288,8192,0.17133422692616782
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,12288,65536,1.294374254014757
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,12288,7168,0.15092178185780844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,12288,7168,0.1427991125318739
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,12288,6144,0.12991378042432997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,12288,6144,0.12246133221520318
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,12288,65536,1.4478168487548828
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,12288,5120,0.11093066798316108
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,12288,5120,0.10632177856233384
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,12288,16384,0.38084888458251953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,12288,4096,0.08974311086866592
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,12288,4096,0.08771200312508477
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,12288,12288,0.29424889882405597
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,12288,3584,0.07944266663657294
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,12288,10240,0.2463315592871772
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,12288,3584,0.07982933521270752
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,12288,8192,0.20231733057234022
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,12288,3072,0.06900533040364583
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,12288,3072,0.07044355736838447
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,12288,2560,0.05927733580271403
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,12288,2560,0.061090668042500816
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,12288,7168,0.18134489324357775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,12288,65536,1.501309288872613
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,12288,2048,0.048806223604414195
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,12288,2048,0.05433422327041626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,12288,1536,0.03895555602179633
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,12288,1536,0.04478044311205546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,12288,6144,0.15830043951670328
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,12288,1024,0.027956446011861164
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,12288,1024,0.034317334493001304
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,12288,5120,0.13742844263712564
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,12288,768,0.0229404436217414
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,12288,768,0.030256887276967365
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,12288,4096,0.11505244837866889
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,12288,512,0.017656889226701524
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,12288,512,0.026672000686327618
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,12288,3584,0.10459111134211223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,12288,256,0.012587555580668978
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,12288,256,0.02331377731429206
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,12288,2560,0.08306844366921319
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,12288,128,0.009628444082207149
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,12288,128,0.020682666036817763
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,12288,3072,0.09429688586129083
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,12288,64,0.008055999875068665
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,12288,64,0.020804444948832195
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,12288,32,0.008458666503429413
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,12288,32,0.02085955606566535
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,12288,1536,0.06392000118891399
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,12288,2048,0.07310044765472412
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,12288,768,0.04732799861166212
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,12288,1024,0.05053955647680494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,10240,16384,0.2804933389027913
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,10240,16384,0.2990310986836751
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,12288,256,0.038161777787738375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,10240,65536,1.114747577243381
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,10240,12288,0.226111994849311
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,12288,128,0.03445422318246629
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,10240,12288,0.20793155829111734
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,10240,10240,0.17535644107394746
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,10240,10240,0.18992355134752062
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,12288,512,0.04134577843878005
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,10240,8192,0.14318133725060359
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,10240,8192,0.15329866939120823
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,10240,65536,1.3062239752875435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,10240,7168,0.126108447710673
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,10240,7168,0.1355039940940009
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,10240,6144,0.11664622359805638
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,10240,6144,0.10857066843244766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,10240,5120,0.09854044516881307
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,10240,5120,0.09304355250464545
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,10240,16384,0.3560755517747667
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,10240,65536,1.3564302656385634
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,10240,4096,0.07991377512613933
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,10240,4096,0.07888622416390313
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,10240,12288,0.27384622891743976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,10240,3584,0.07216799921459623
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,10240,3584,0.07092355357276069
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,10240,10240,0.2312240070766873
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,10240,3072,0.06392977635065715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,10240,3072,0.062302225165896945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,10240,8192,0.18903288576338026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,10240,2560,0.05354400144682991
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,10240,2560,0.05414577656322055
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,10240,2048,0.0440924432542589
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,10240,7168,0.1732382244533963
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,10240,2048,0.04775999983151754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,10240,1536,0.03447644578086005
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,10240,1536,0.04040888945261637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,10240,6144,0.14923911624484593
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,10240,1024,0.02516711089346144
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,10240,1024,0.031344887283113264
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,10240,5120,0.12921778361002603
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,10240,768,0.02040177749262916
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,10240,768,0.027928001350826685
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,10240,3584,0.10018577840593125
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,10240,512,0.016190222567982145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,10240,512,0.02497688929239909
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,10240,2560,0.07831999990675184
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,10240,256,0.012967111335860359
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,10240,256,0.020980444219377305
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,10240,4096,0.10919733179940118
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,10240,128,0.011149333582984077
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,10240,128,0.01942133406798045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,10240,1536,0.060468441910213895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,10240,64,0.010602666272057427
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,10240,64,0.01953155630164676
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,10240,32,0.011040888726711273
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,10240,32,0.019344889455371432
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,10240,3072,0.0886533326572842
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,10240,2048,0.06925511360168457
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,10240,768,0.044544888867272266
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,10240,1024,0.04779466655519274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,8192,65536,0.9347306357489692
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,8192,16384,0.22871910201178658
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,8192,16384,0.2448364363776313
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,8192,12288,0.17355822192298043
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,8192,12288,0.17842755052778456
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,10240,128,0.03204977843496535
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,10240,512,0.03913866811328464
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,8192,65536,1.0649200015597873
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,10240,256,0.03451111250453525
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,8192,10240,0.14533066749572754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,8192,10240,0.15022488435109457
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,8192,8192,0.1234968900680542
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,8192,8192,0.11875555250379775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,8192,7168,0.10834578010771011
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,8192,7168,0.10294310914145575
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,8192,6144,0.09041688839594524
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,8192,6144,0.09408533573150635
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,8192,5120,0.07536533143785265
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,8192,5120,0.08025511105855306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,8192,16384,0.31302044126722545
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,8192,4096,0.06220799684524536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,8192,65536,1.1893315845065646
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,8192,4096,0.06720088587866889
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,8192,3584,0.055211553970972695
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,8192,3584,0.06145866711934408
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,8192,12288,0.24068000581529406
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,8192,3072,0.04827555682924059
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,8192,3072,0.05637511279847887
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,8192,10240,0.20524356100294325
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,8192,2560,0.041230221589406334
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,8192,2560,0.04770311050944858
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,8192,7168,0.14926489194234213
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,8192,8192,0.166975114080641
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,8192,2048,0.040906667709350586
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,8192,2048,0.034056001239352755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,8192,1536,0.027012444204754297
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,8192,1536,0.036002665758132935
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,8192,6144,0.1323431067996555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,8192,1024,0.01999288962946998
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,8192,1024,0.028778665595584448
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,8192,5120,0.11461599667867024
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,8192,768,0.016672000288963318
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,8192,768,0.025791999366548326
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,8192,3584,0.08846933311886257
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,8192,512,0.012879110872745514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,8192,512,0.02273333403799269
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,8192,4096,0.0972977810435825
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,8192,256,0.009315555294354757
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,8192,256,0.01889244384235806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,8192,2560,0.0713919997215271
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,8192,128,0.007213333414660559
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,8192,128,0.017864000466134813
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,8192,3072,0.08006488614612155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,8192,64,0.006263999889294307
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,8192,64,0.018138666947682697
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,8192,32,0.006604444649484422
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,8192,768,0.04228622383541531
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,8192,32,0.018037334084510803
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,8192,1536,0.05592088566886055
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,8192,2048,0.061675555176205106
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,8192,1024,0.044848889112472534
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,7168,16384,0.19815999931759307
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,7168,65536,0.9512169096204969
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,7168,16384,0.23782576455010307
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,7168,12288,0.17892532878451875
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,7168,12288,0.14978755844963923
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,7168,10240,0.12604799535539415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,7168,10240,0.15135910775926378
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,8192,128,0.03150755498144362
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,8192,256,0.032933334509531655
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,7168,65536,0.785553773244222
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,7168,8192,0.12138222323523627
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,7168,8192,0.10131644540362889
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,7168,7168,0.10611644718382095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,7168,7168,0.08946310811572605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,7168,6144,0.07765511009428236
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,8192,512,0.0373271107673645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,7168,6144,0.09340088897281223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,7168,5120,0.06573422087563409
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,7168,5120,0.07979822158813477
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,7168,16384,0.307549344168769
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,7168,4096,0.054005334774653115
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,7168,4096,0.06689777639177111
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,7168,65536,1.1389164394802518
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,7168,12288,0.2366480032602946
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,7168,3584,0.04877066612243652
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,7168,3584,0.060896886719597705
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,7168,3072,0.04196711050139534
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,7168,3072,0.05576355589760674
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,7168,10240,0.20045067204369438
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,7168,2560,0.035946667194366455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,7168,2560,0.04741600155830383
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,7168,8192,0.16480888260735407
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,7168,2048,0.029534220695495605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,7168,2048,0.04048266675737169
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,7168,7168,0.1475048859914144
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,7168,1536,0.02349511120054457
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,7168,1536,0.03548089000913832
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,7168,6144,0.128346668349372
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,7168,1024,0.017465778523021273
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,7168,1024,0.028446220689349707
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,7168,5120,0.11228799819946289
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,7168,768,0.014389332797792224
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,7168,768,0.02550044490231408
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,7168,3584,0.08713866604699029
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,7168,512,0.011109333071443768
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,7168,512,0.021520000365045335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,7168,4096,0.09550755553775364
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,7168,256,0.008273777862389883
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,7168,256,0.01871200071440803
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,7168,3072,0.0793271130985684
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,7168,128,0.007301333049933116
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,7168,128,0.017794667018784415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,7168,2560,0.0699253347184923
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,7168,64,0.006256888724035687
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,7168,64,0.01817244456874
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,7168,32,0.00665244460105896
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,7168,32,0.01798577772246467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,7168,1536,0.055604444609748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,7168,2048,0.06062755319807264
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,7168,768,0.04184533490075005
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,6144,16384,0.1685617764790853
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,7168,1024,0.0440977778699663
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,6144,16384,0.23485867182413736
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,6144,65536,0.9262729220920138
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,6144,65536,0.698670228322347
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,6144,12288,0.17811555332607695
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,6144,12288,0.12791555457644993
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,6144,10240,0.14854221873813206
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,6144,10240,0.1082551081975301
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,6144,8192,0.08723466926150852
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,7168,256,0.033610665135913424
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,6144,8192,0.12116977903578018
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,7168,128,0.031062222189373438
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,6144,7168,0.10684711403316921
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,6144,7168,0.07717421982023451
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,7168,512,0.037103109889560275
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,6144,6144,0.06717599762810601
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,6144,6144,0.09294311205546062
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,6144,5120,0.056905779573652476
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,6144,5120,0.07916533284717135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,6144,16384,0.3139244450463189
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,6144,12288,0.2404257721371121
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,6144,4096,0.066520889600118
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,6144,4096,0.04673511121008131
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,6144,3584,0.041583110888799034
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,6144,65536,1.1672257317437065
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,6144,3584,0.06046666701634725
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,6144,3072,0.036377777655919395
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,6144,3072,0.055567999680837
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,6144,10240,0.20400533411237928
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,6144,2560,0.031018667750888403
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,6144,2560,0.046145776907602944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,6144,8192,0.1656079954571194
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,6144,2048,0.025980444418059453
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,6144,2048,0.03999377621544732
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,6144,7168,0.14835111300150552
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,6144,1536,0.02093066606256697
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,6144,1536,0.034563554657830134
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,6144,6144,0.1306497785780165
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,6144,1024,0.015747555428081088
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,6144,1024,0.028212444649802312
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,6144,5120,0.11289955510033502
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,6144,768,0.013111111190583972
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,6144,768,0.025076443950335186
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,6144,4096,0.09641244676378037
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,6144,512,0.010430222584141625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,6144,3584,0.08727822038862441
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,6144,512,0.021178666088316176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,6144,256,0.007416889071464539
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,6144,256,0.01849244369400872
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,6144,3072,0.07904089159435697
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,6144,128,0.006125333408514659
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,6144,128,0.017514665921529133
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,6144,2560,0.06998755534489949
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,6144,64,0.005344888816277186
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,6144,64,0.01780533293883006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,6144,32,0.005696889013051987
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,6144,32,0.017664889494578045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,6144,2048,0.06104533539877998
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,6144,1536,0.05412711037529839
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,6144,768,0.042118221521377563
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,6144,1024,0.04456888967090183
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,5120,16384,0.19605599509345162
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,5120,65536,0.7546097967359754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,5120,16384,0.1521377828386095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,5120,12288,0.14957333935631648
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,5120,12288,0.11547466119130452
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,5120,65536,0.6195733282301161
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,5120,10240,0.09693155686060588
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,5120,10240,0.12568088372548422
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,5120,8192,0.07870578103595309
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,5120,8192,0.10142755508422852
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,6144,256,0.03247288862864176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,6144,512,0.037235554721620344
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,6144,128,0.030613332986831665
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,5120,7168,0.06966311401791044
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,5120,7168,0.09035644266340469
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,5120,6144,0.07905777957704332
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,5120,6144,0.06091555621888903
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,5120,5120,0.05152444375885857
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,5120,5120,0.06805777549743652
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,5120,16384,0.267410675684611
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,5120,4096,0.042066666815016
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,5120,4096,0.05706577830844455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,5120,12288,0.20554488235049775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,5120,3584,0.03763555486996969
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,5120,3584,0.0521982212861379
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,5120,65536,0.9754301706949869
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,5120,3072,0.033209777540630765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,5120,10240,0.17459377977583143
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,5120,3072,0.04820977648099264
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,5120,2560,0.028463111983405218
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,5120,2560,0.040602667464150324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,5120,8192,0.14352088504367405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,5120,2048,0.023616888456874426
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,5120,2048,0.035135110219319664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,5120,7168,0.12812799877590603
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,5120,1536,0.01901244454913669
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,5120,6144,0.11306577258639866
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,5120,1536,0.03095910946528117
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,5120,1024,0.014392000105645923
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,5120,1024,0.02508977717823452
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,5120,5120,0.09956889020072089
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,5120,768,0.012054222325483957
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,5120,768,0.02197333342499203
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,5120,3584,0.0779822203848097
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,5120,512,0.009720889230569204
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,5120,512,0.019183110859658983
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,5120,2560,0.06264266702863905
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,5120,256,0.007090667055712805
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,5120,4096,0.085116446018219
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,5120,256,0.017215111189418368
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,5120,128,0.0058293334311909145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,5120,128,0.01629599928855896
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,5120,3072,0.07111022207472059
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,5120,64,0.005417777846256892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,5120,64,0.016554666890038382
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,5120,32,0.005611555443869696
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,5120,32,0.016178665889634024
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,5120,1536,0.04965688784917196
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,5120,2048,0.05486933390299479
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,5120,768,0.03772266705830892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,4096,65536,0.740971565246582
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,4096,65536,0.4741911358303494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,4096,16384,0.11655467086368138
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,4096,16384,0.19584443834092882
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,5120,1024,0.04054755634731717
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,4096,12288,0.14927821689181858
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,4096,12288,0.08847644593980576
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,4096,10240,0.12607644663916692
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,4096,10240,0.07419733206431071
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,4096,8192,0.06042489078309801
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,4096,8192,0.10066044330596924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,5120,128,0.030008001459969416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,5120,256,0.03128355410363939
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,4096,7168,0.08879022465811835
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,4096,7168,0.053605334626303784
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,5120,512,0.034761776526769005
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,4096,6144,0.07824622260199653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,4096,6144,0.04669777883423699
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,4096,5120,0.03973689013057285
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,4096,5120,0.06737777921888563
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,4096,65536,0.8963137732611762
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,4096,4096,0.032638221979141235
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,4096,4096,0.05724000268512302
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,4096,12288,0.1890017721388075
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,4096,3584,0.029046220911873713
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,4096,16384,0.2451324462890625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,4096,3584,0.052137778864966504
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,4096,3072,0.02555022140343984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,4096,3072,0.04775555597411262
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,4096,10240,0.1606302261352539
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,4096,2560,0.022384888596004907
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,4096,2560,0.0402782228257921
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,4096,8192,0.13205777274237737
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,4096,2048,0.01867911054028405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,4096,2048,0.03555644551912943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,4096,7168,0.11831288867526585
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,4096,1536,0.015135111080275642
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,4096,1536,0.03070577647950914
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,4096,6144,0.10445066955354478
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,4096,1024,0.011561777856614856
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,4096,1024,0.02365066607793172
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,4096,5120,0.09195466836293538
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,4096,768,0.009696000152164036
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,4096,768,0.021348444951905146
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,4096,3584,0.07303822040557861
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,4096,512,0.007619555625650618
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,4096,512,0.019184889064894784
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,4096,4096,0.07874666982226901
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,4096,256,0.00589155571328269
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,4096,2560,0.05868622329499987
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,4096,256,0.017268444101015728
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,4096,128,0.004957333207130432
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,4096,128,0.016193777322769165
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,4096,768,0.03586577706866794
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,4096,64,0.004456888884305954
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,4096,64,0.016373333003785875
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,4096,32,0.0047324444684717385
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,4096,32,0.016132444143295288
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,4096,3072,0.06648088826073541
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,4096,1536,0.047268443637424044
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,4096,2048,0.052654223309622876
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3584,65536,0.7416177855597602
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,4096,1024,0.03891911109288534
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3584,16384,0.101583997408549
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3584,16384,0.19563021924760607
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3584,12288,0.07737421989440918
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3584,12288,0.1495235496097141
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3584,65536,0.389066669676039
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3584,10240,0.06532266404893664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3584,10240,0.1259102159076267
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3584,8192,0.05386844608518812
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3584,8192,0.10069777568181355
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,4096,128,0.02923999892340766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3584,7168,0.04710400104522705
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3584,7168,0.08871644735336304
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,4096,256,0.030204445123672485
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3584,6144,0.04156800111134847
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3584,6144,0.07761777771843804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,4096,512,0.03148444493611654
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3584,5120,0.03559644354714288
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3584,5120,0.06763466861512926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,3584,16384,0.2420275476243761
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3584,4096,0.02940800123744541
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3584,4096,0.056609776284959584
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,3584,12288,0.18646221690707734
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3584,3584,0.026353778110610113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3584,3584,0.05164088805516561
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,3584,65536,0.8778159883287219
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3584,3072,0.02325155503220028
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3584,3072,0.04721688893106249
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,3584,10240,0.15874844127231175
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3584,2560,0.020216888851589628
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3584,2560,0.03977155685424805
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,3584,8192,0.13016711341010198
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3584,2048,0.017311111092567444
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3584,2048,0.035107556316587664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,3584,7168,0.11668445004357232
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3584,1536,0.014069333672523499
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3584,1536,0.030220445659425523
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,3584,6144,0.10332710875405206
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3584,1024,0.01127199994193183
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3584,1024,0.023235556152131822
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,3584,5120,0.0908622211880154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3584,768,0.009441778063774109
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3584,768,0.021032000581423443
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,3584,3584,0.07212533553441365
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3584,512,0.007703999678293864
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,3584,2560,0.058034665054745145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3584,512,0.018927999668651156
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3584,256,0.006056889063782162
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3584,256,0.016926222377353244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,3584,3072,0.06574310859044392
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,3584,4096,0.0778577791319953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3584,128,0.005321777943107817
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3584,128,0.016062221593326993
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3584,64,0.004888888862397936
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3584,64,0.01625155574745602
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3584,32,0.005139555368158552
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3584,32,0.015847999188635085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,3584,2048,0.05190666516621908
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,3584,1536,0.046985778543684215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,3584,768,0.0352497763103909
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3072,65536,0.744865788353814
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3072,65536,0.34700621498955625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3072,16384,0.08699555529488458
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3072,16384,0.19604267014397514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3072,12288,0.1489831076727973
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3072,12288,0.06685066885418363
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3072,10240,0.056314667065938316
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3072,10240,0.1250942283206516
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,3584,1024,0.03867910967932807
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3072,8192,0.04582133226924472
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3072,8192,0.10005066792170207
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,3584,256,0.03018222252527873
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3072,7168,0.04064088728692796
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3072,7168,0.08772177828682794
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,3584,128,0.02918133470747206
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3072,6144,0.03522577881813049
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3072,6144,0.07767999834484525
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,3584,512,0.03228266702757941
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3072,5120,0.030221333106358845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3072,5120,0.06590221987830268
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,3072,16384,0.24351112047831217
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3072,4096,0.025447110335032146
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3072,4096,0.056511110729641385
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,3072,65536,0.8927368587917752
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,3072,12288,0.18694844510820177
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3072,3584,0.02241777711444431
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3072,3584,0.051573332813051015
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3072,3072,0.019870221614837646
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3072,3072,0.046358221107059054
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,3072,10240,0.1590862274169922
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3072,2560,0.01737422247727712
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3072,2560,0.03967466619279649
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,3072,8192,0.13028532928890652
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3072,2048,0.014825777875052558
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3072,2048,0.03484800126817491
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,3072,7168,0.11642311678992377
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3072,1536,0.012111110819710625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,3072,6144,0.10327644480599298
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3072,1536,0.030072000291612413
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3072,1024,0.009535999761687385
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3072,1024,0.023384888966878254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,3072,5120,0.09071733554204305
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3072,768,0.007883555359310573
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3072,768,0.02093688812520769
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,3072,4096,0.0785084433025784
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3072,512,0.006439111298984951
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3072,512,0.01887822151184082
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,3072,3584,0.07235822412702772
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3072,256,0.0051893335249688895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3072,256,0.016932444439993966
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,3072,3072,0.06539911031723022
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3072,128,0.004516444272465176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3072,128,0.015998222761683993
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,3072,2560,0.0580915543768141
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3072,64,0.004031110968854693
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3072,64,0.01609688831700219
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,3072,32,0.0042364444169733255
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,3072,32,0.01591377788119846
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,3072,2048,0.051810665263069995
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2560,65536,0.5907671186659071
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2560,65536,0.3098062144385444
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2560,16384,0.07880177762773302
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2560,16384,0.15915466679467097
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,3072,1536,0.04687022169431051
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,3072,768,0.03628177775277032
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2560,12288,0.05968888600667318
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2560,12288,0.12218222353193496
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2560,10240,0.05043377810054355
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,3072,1024,0.03740266627735562
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2560,10240,0.102966222498152
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2560,8192,0.04184444414244758
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2560,8192,0.08223911126454671
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,3072,256,0.030335999197430078
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,3072,512,0.03153511219554477
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2560,7168,0.03684799869855245
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2560,7168,0.07175466749403211
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,3072,128,0.028945777151319716
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2560,6144,0.032157331705093384
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2560,6144,0.06485866838031344
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2560,5120,0.027666666441493567
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2560,5120,0.05633778042263455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,2560,65536,0.8088169097900391
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2560,4096,0.02276800076166789
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,2560,16384,0.22516443994310167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2560,4096,0.047798222965664335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2560,3584,0.02053777707947625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2560,3584,0.043741332160102
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,2560,12288,0.1730684439341227
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2560,3072,0.01825066738658481
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2560,3072,0.039565334717432656
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,2560,10240,0.14610755443572998
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2560,2560,0.015825778245925903
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2560,2560,0.03374044431580438
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,2560,8192,0.12060977352990045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2560,2048,0.013541333377361298
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2560,2048,0.03033777740266588
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,2560,7168,0.10848000314500596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2560,1536,0.011198221809334226
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2560,1536,0.025812443759706285
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,2560,6144,0.09636088874604966
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,2560,5120,0.08466400040520562
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2560,1024,0.008945778012275696
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2560,1024,0.021776000658671062
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2560,768,0.007575111256705389
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2560,768,0.01941244469748603
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,2560,4096,0.07350489166047838
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,2560,3584,0.06798488563961454
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2560,512,0.006046222315894232
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2560,512,0.017648888958825003
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2560,256,0.004884444591071871
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2560,256,0.015966221690177917
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,2560,3072,0.06124711036682129
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2560,128,0.004306666553020477
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2560,128,0.01498133275243971
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,2560,2560,0.054696887731552124
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2560,64,0.003984000119898054
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2560,64,0.014934221903483072
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2560,32,0.004041777716742622
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2560,32,0.014700444208251106
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,2560,2048,0.048730668094423085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2048,65536,0.590502209133572
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,2560,768,0.03404444456100464
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,2560,1536,0.04485599862204658
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2048,65536,0.23335732354058158
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2048,16384,0.1585599978764852
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,2560,1024,0.03506133291456435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2048,12288,0.12091378370920818
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2048,16384,0.08292444547017415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2048,12288,0.06255022022459242
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2048,10240,0.10168088807000054
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2048,10240,0.05173866616355049
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2048,8192,0.04186577929390801
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2048,8192,0.08164088593588935
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,2560,256,0.029136889510684546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2048,7168,0.03643733263015747
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,2560,512,0.030291554000642564
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2048,7168,0.07347110907236735
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,2560,128,0.02844088938501146
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2048,6144,0.030213332838482324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2048,6144,0.06438933478461371
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2048,5120,0.025942222939597234
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2048,5120,0.05614488654666477
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,2048,16384,0.22245332929823133
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2048,4096,0.021849777963426378
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2048,4096,0.0471111105548011
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,2048,65536,0.8048382335238986
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2048,3584,0.019455111689037748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,2048,12288,0.17239556047651503
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2048,3584,0.04306311077541775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2048,3072,0.01662400033738878
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,2048,10240,0.14601600170135498
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2048,3072,0.039220445685916476
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2048,2560,0.012888000243239932
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,2048,8192,0.12019999821980794
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2048,2560,0.034021334515677557
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2048,2048,0.011067555182509951
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2048,2048,0.030230220821168687
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,2048,7168,0.10817066828409831
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2048,1536,0.009342222577995723
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2048,1536,0.02548355526394314
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,2048,6144,0.09577155775494045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2048,1024,0.007426666716734569
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2048,1024,0.020891555481486853
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,2048,5120,0.08473510874642266
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2048,768,0.006229333165619109
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2048,768,0.019173334042231243
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,2048,2560,0.05520177880922953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2048,512,0.005244444641802046
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,2048,3584,0.06799199846055773
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2048,512,0.017443555924627516
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2048,256,0.00443733317984475
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2048,256,0.015623999966515435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,2048,4096,0.07341155740949842
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2048,128,0.003923555629120933
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2048,128,0.014719110396173267
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,2048,3072,0.061298668384552
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2048,64,0.0035546666218174826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2048,64,0.01475288967291514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,2048,2048,0.048680000834994845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,2048,32,0.0037582222786214617
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,2048,32,0.01459822224246131
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,2048,768,0.03372088736957974
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1536,65536,0.5902586513095432
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1536,65536,0.17701866891649035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1536,16384,0.05361955695682102
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1536,16384,0.15837066703372532
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,2048,1536,0.044234666559431285
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1536,12288,0.043381333351135254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1536,12288,0.12195555369059245
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1536,10240,0.03840355409516229
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1536,10240,0.10223644309573704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,2048,1024,0.035339554150899254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1536,8192,0.027240888939963445
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1536,8192,0.08133866389592488
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,2048,128,0.028428445259730022
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1536,7168,0.024335111180941265
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,2048,256,0.029184887806574505
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1536,7168,0.07339644432067871
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1536,6144,0.021448888712459143
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1536,6144,0.06418222188949585
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,2048,512,0.03017866611480713
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1536,5120,0.018485332528750103
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1536,5120,0.05470577875773112
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,1536,16384,0.22204000420040557
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1536,4096,0.015593777100245157
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,1536,65536,0.7960249053107368
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1536,4096,0.04695822132958306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1536,3584,0.014168888330459595
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1536,3584,0.042881776889165245
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,1536,12288,0.17060711648729113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1536,3072,0.012693333129088083
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1536,3072,0.03892177674505446
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,1536,10240,0.14574044280582005
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1536,2560,0.012677333421177335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1536,2560,0.03402488761478
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,1536,8192,0.11978222264183892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1536,2048,0.011401777466138205
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1536,2048,0.03002133303218418
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,1536,7168,0.10791199737124973
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1536,1536,0.0092613332801395
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,1536,6144,0.0951742198732164
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1536,1536,0.025352888637118872
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1536,1024,0.007171555525726742
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1536,1024,0.020800888538360596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,1536,5120,0.08434310886594985
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1536,768,0.006232000059551663
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1536,768,0.018975999620225694
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,1536,4096,0.07297066847483318
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1536,512,0.005252444495757421
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,1536,3584,0.06711200210783216
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1536,512,0.017154667112562392
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1536,256,0.0042239998777707415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1536,256,0.01551644504070282
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,1536,3072,0.06067733632193672
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1536,128,0.003738666574160258
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1536,128,0.014688889185587565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,1536,2560,0.05479733480347527
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,1536,2048,0.04845688740412394
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1536,64,0.003456888927353753
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1536,64,0.014703111516104804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1536,32,0.00351555562681622
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1536,32,0.014466666513019137
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1024,65536,0.5876586702134874
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1024,65536,0.12312532795800103
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1024,16384,0.037851555479897395
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1024,16384,0.15904177559746638
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,1536,1536,0.04271999994913737
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1024,12288,0.030408889055252075
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1024,12288,0.12166310681237115
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,1536,1024,0.035531553957197405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1024,10240,0.0377511117193434
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1024,10240,0.10117689106199478
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,1536,256,0.028823110792371962
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,1536,768,0.03386844529045953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1024,8192,0.03231822119818793
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1024,8192,0.08184266752666898
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1024,7168,0.02681422233581543
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,1536,512,0.030072000291612413
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1024,7168,0.07317155599594116
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,1536,128,0.027845333019892376
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1024,6144,0.023938667443063524
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1024,6144,0.06370577547285292
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1024,5120,0.02016799979739719
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1024,5120,0.055239998632007174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,1024,16384,0.2101751168568929
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1024,4096,0.01796444422668881
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1024,4096,0.0464915566974216
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,1024,65536,0.7528239885965983
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1024,3584,0.015323556131786771
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1024,3584,0.04204622242185804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,1024,12288,0.16147732734680176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1024,3072,0.013145777914259167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1024,3072,0.03819555706448025
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,1024,10240,0.13710578282674155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1024,2560,0.008942222429646386
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1024,2560,0.033379554748535156
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,1024,7168,0.10167911317613389
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,1024,8192,0.11365421613057454
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1024,2048,0.00742133292886946
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1024,2048,0.029282665914959375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1024,1536,0.006573333508438534
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,1024,5120,0.07977510823143853
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1024,1536,0.024638222323523626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1024,1024,0.005504888792832692
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,1024,6144,0.09045777718226115
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1024,1024,0.020357333951526217
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1024,768,0.004950222041871813
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1024,768,0.018719111879666645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,1024,3584,0.06313421991136339
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1024,512,0.004295999805132548
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1024,512,0.01700266698996226
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,1024,4096,0.06905066967010498
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1024,256,0.0037760001917680106
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1024,256,0.015372445185979208
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,1024,2560,0.05190044310357836
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1024,128,0.0034773333205117118
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1024,128,0.01462399959564209
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,1024,3072,0.05807022253672282
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1024,64,0.003277333246337043
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1024,64,0.014534221755133735
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,1024,32,0.0033742222521040174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,1024,32,0.014261333478821648
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,768,65536,0.09603733486599392
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,1024,768,0.03233777814441257
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,768,65536,0.5886977513631185
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,768,16384,0.031918221049838595
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,768,16384,0.1582453383339776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,1024,1536,0.041724443435668945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,1024,2048,0.046168890264299184
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,768,12288,0.02608355548646715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,768,12288,0.12054044670528835
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,768,10240,0.023464888334274292
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,1024,1024,0.033005333609051175
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,768,10240,0.10044711165957981
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,768,8192,0.018547554810841877
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,768,8192,0.0804799993832906
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,1024,256,0.028631998433007136
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,768,7168,0.016787555482652452
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,768,7168,0.07312355438868205
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,1024,512,0.028936889436509874
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,768,6144,0.014751111467679342
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,768,6144,0.06387022468778822
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,1024,128,0.027388445205158655
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,768,5120,0.01294133315483729
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,768,5120,0.055312888489829175
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,768,65536,0.7310542000664605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,768,4096,0.011160000330872007
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,768,16384,0.20356355773078072
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,768,4096,0.046258668104807533
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,768,3584,0.010145777629481422
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,768,12288,0.1566151115629408
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,768,3584,0.041802667909198336
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,768,3072,0.009257777697510189
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,768,3072,0.037696888049443565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,768,10240,0.13400622208913168
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,768,2560,0.010145777629481422
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,768,2560,0.03331466515858968
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,768,8192,0.11028711001078288
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,768,2048,0.008909333083364699
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,768,2048,0.029320889049106177
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,768,7168,0.09962311055925156
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,768,1536,0.007359999749395583
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,768,1536,0.024375110864639282
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,768,6144,0.08816888597276475
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,768,1024,0.005965333431959152
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,768,1024,0.020285333196322124
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,768,5120,0.07804977893829346
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,768,768,0.005212444398138259
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,768,768,0.018777777751286823
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,768,4096,0.06725688775380452
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,768,512,0.004513777792453766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,768,512,0.016737777325842116
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,768,3584,0.06215555800331963
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,768,256,0.00351200004418691
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,768,256,0.01514311134815216
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,768,3072,0.056417776478661426
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,768,128,0.0032417778339650896
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,768,2560,0.05116000109248691
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,768,128,0.014370666609870063
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,768,64,0.003044444446762403
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,768,64,0.014286221729384528
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,768,32,0.0030826667530669104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,768,32,0.014058666096793281
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,512,65536,0.06703911225001018
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,512,65536,0.5888569090101454
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,768,2048,0.04526755544874403
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,512,16384,0.024847111768192712
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,512,16384,0.15676977899339464
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,768,1536,0.04030133287111918
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,512,12288,0.019186667270130582
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,512,12288,0.11962489287058513
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,768,1024,0.033231112692091204
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,512,10240,0.019242667489581637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,512,10240,0.1017342209815979
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,768,768,0.03221777743763394
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,512,8192,0.016321778297424316
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,512,8192,0.08144266737831964
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,768,256,0.028096887800428603
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,512,7168,0.014607111612955729
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,768,512,0.028911110427644517
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,512,7168,0.07238222493065728
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,512,6144,0.013043555948469373
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,768,128,0.027394667267799377
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,512,6144,0.06299377812279595
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,512,5120,0.011571555501884885
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,512,5120,0.05450488792525398
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,512,65536,0.722450680202908
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,512,4096,0.009935999910036722
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,512,4096,0.045765333705478244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,512,16384,0.20029244158003065
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,512,3584,0.009279111193286048
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,512,3584,0.04148088892300924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,512,12288,0.15467466248406306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,512,3072,0.008278222547637092
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,512,10240,0.13151289357079401
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,512,3072,0.037490665912628174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,512,2560,0.007366221812036302
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,512,2560,0.03296622302797105
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,512,8192,0.10918400022718643
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,512,2048,0.006529777828190062
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,512,2048,0.02772177755832672
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,512,6144,0.08696889215045506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,512,7168,0.09789599974950154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,512,1536,0.0057795556883017225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,512,1536,0.024057777391539678
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,512,1024,0.004722666823201709
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,512,5120,0.07687199778027005
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,512,1024,0.020026667250527274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,512,768,0.00434311106801033
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,512,768,0.018379555808173288
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,512,4096,0.06680978006786771
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,512,512,0.003938666648334927
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,512,512,0.016736888223224215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,512,3584,0.0616133345497979
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,512,256,0.0034497777620951333
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,512,256,0.015014222926563688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,512,3072,0.055697778860727944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,512,128,0.0032088889016045462
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,512,128,0.014316444595654806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,512,2560,0.05040622088644239
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,512,64,0.0030177777840031516
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,512,2048,0.04447199900945028
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,512,64,0.014155555102560254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,512,32,0.003112888791494899
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,512,32,0.013950222068362765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,256,65536,0.044975999328825206
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,256,65536,0.5887706544664171
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,256,16384,0.01500266624821557
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,256,16384,0.15891467200385198
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,512,1536,0.039664000272750854
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,256,12288,0.015824889143308003
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,256,12288,0.12064266204833984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,512,768,0.03197600113021003
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,256,10240,0.014175110393100314
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,512,1024,0.03291822142071194
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,256,10240,0.10211111439598931
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,256,8192,0.012627555264366997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,256,8192,0.08120710982216729
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,512,512,0.028736889362335205
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,256,7168,0.01183377785815133
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,256,7168,0.07249422205819024
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,512,256,0.028315554062525432
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,256,6144,0.010837333069907295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,256,6144,0.06291111310323079
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,512,128,0.027497778336207073
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,256,5120,0.009869332942697737
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,256,5120,0.05418489045566983
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,256,16384,0.19372177124023438
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,256,4096,0.008796444369686974
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,256,65536,0.6923991309271919
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,256,4096,0.04562933246294657
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,256,3584,0.008070222205585903
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,256,3584,0.04120799899101257
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,256,12288,0.14951022466023764
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,256,3072,0.007296000089910295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,256,10240,0.12717066870795354
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,256,3072,0.03666222095489502
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,256,2560,0.006579555571079254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,256,2560,0.031022220849990845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,256,8192,0.1046595573425293
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,256,2048,0.0057502223385704895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,256,2048,0.027327999472618103
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,256,7168,0.09470311138365005
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,256,1536,0.005122666557629903
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,256,1536,0.024080889092551336
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,256,6144,0.08437866634792751
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,256,1024,0.004332444320122401
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,256,1024,0.02010311186313629
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,256,5120,0.07460888889100817
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,256,768,0.004007110993067424
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,256,768,0.018424888451894123
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,256,4096,0.06414310799704658
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,256,512,0.00369777778784434
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,256,512,0.016737777325842116
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,256,3584,0.059657779004838735
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,256,256,0.0032666667054096856
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,256,256,0.015029332704014249
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,256,3072,0.05447733402252197
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,256,128,0.002945777856641345
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,256,128,0.014169777433077494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,256,2560,0.04874933428234524
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,256,64,0.0028568889117903183
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,256,64,0.014150222142537435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,256,32,0.00291377781993813
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,256,32,0.013888888888888888
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,128,65536,0.046118223004870944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,128,65536,0.5878426763746474
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,256,2048,0.04118311074044969
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,128,16384,0.012288000020715924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,128,16384,0.1584577825334337
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,256,1536,0.03766399953100417
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,128,12288,0.010551999840471478
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,128,12288,0.12199644247690837
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,256,1024,0.03191466795073615
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,128,10240,0.009483555952707926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,128,10240,0.10198311010996501
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,256,768,0.031108445591396753
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,128,8192,0.00869777798652649
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,128,8192,0.08057066467073229
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,256,512,0.02827111217710707
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,128,7168,0.008213332957691616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,128,7168,0.07249333461125691
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,256,256,0.02781955401102702
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,128,6144,0.007768889268239339
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,128,6144,0.06268088685141669
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,256,128,0.02693155573474036
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,128,5120,0.008602666358153025
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,128,5120,0.05377777748637729
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,128,65536,0.6897004445393881
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,128,4096,0.007923555870850882
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,128,4096,0.044329779015647046
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,128,16384,0.1929759979248047
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,128,3584,0.0075777777367168
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,128,3584,0.040672888358434044
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,128,12288,0.14913154972924128
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,128,3072,0.00720355576939053
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,128,3072,0.0363031095928616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,128,10240,0.12657066186269125
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,128,2560,0.0063528890411059065
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,128,2560,0.03087555699878269
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,128,8192,0.10424533155229355
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,128,2048,0.00564533347884814
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,128,2048,0.02716711163520813
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,128,7168,0.09458755784564549
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,128,1536,0.004965333475006951
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,128,1536,0.02403733299838172
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,128,6144,0.08415022161271836
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,128,1024,0.004231111043029361
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,128,1024,0.01999200052685208
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,128,5120,0.07409599754545423
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,128,768,0.0038764443662431505
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,128,768,0.01828799976242913
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,128,4096,0.06389333142174615
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,128,512,0.0035751110149754416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,128,512,0.016582222448454965
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,128,3584,0.05878844526078966
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,128,256,0.003137777869900068
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,128,3072,0.05279733406172859
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,128,256,0.014797333214018079
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,128,2560,0.047080000241597496
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,128,128,0.0028942223224374983
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,128,128,0.014019555515713163
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,128,64,0.0027431110954946945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,128,64,0.013921777407328287
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,128,32,0.002805333377586471
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,128,32,0.013657777673668332
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,64,65536,0.036622222926881574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,64,65536,0.5885457992553711
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,64,16384,0.009936889012654623
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,64,16384,0.15856977303822836
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,64,12288,0.008944888909657797
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,64,12288,0.12144622537824844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,64,10240,0.008245333201355403
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,64,10240,0.10218844148847793
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,64,8192,0.0075635554061995606
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,64,8192,0.0809804466035631
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,64,7168,0.00721599989467197
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,64,7168,0.07223377625147502
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,64,6144,0.006859555426571105
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,64,6144,0.06316088967853122
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,64,5120,0.00721866637468338
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,64,5120,0.05464977688259549
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,128,2048,0.040885332557890154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,64,4096,0.006809777683681912
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,64,3584,0.007239999870459239
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,64,4096,0.04464177621735466
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,64,3072,0.006783999916579988
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,64,3584,0.040983110666275024
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,64,2560,0.006235555642180973
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,64,3072,0.03554133243030972
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,64,2048,0.005554666535721884
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,64,2560,0.031122667921913996
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,64,1536,0.00488800017370118
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,64,2048,0.027199111051029626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,64,1024,0.004192000048028098
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,64,1536,0.023929778072569106
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,64,768,0.003826666623353958
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,64,1024,0.020037333170572918
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,64,512,0.0034977777136696708
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,64,768,0.018231999542978074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,128,1536,0.03736266824934218
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,64,256,0.0031439999325407874
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,64,512,0.016516443755891588
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,64,128,0.00286311118139161
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,64,256,0.014728888869285583
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,64,128,0.013971555564138623
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,64,64,0.0027013334135214486
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,64,32,0.002806222273243798
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,64,64,0.013890667094124688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,64,32,0.013677332964208392
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,32,16384,0.009299555586444007
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,32,65536,0.03569688730769687
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,32,12288,0.007880888879299164
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,32,16384,0.15853777196672228
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,32,10240,0.007515555454625024
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,32,65536,0.5890106625027126
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,32,12288,0.12123021814558242
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,32,8192,0.007256888680987888
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,32,10240,0.10203733046849568
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,32,7168,0.0069191112286514705
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,32,8192,0.08075466420915392
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,32,6144,0.006667555620272954
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,32,7168,0.07212711042828031
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,32,5120,0.007311999797821045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,32,6144,0.06268266836802165
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,32,4096,0.0069226668112807805
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,32,5120,0.05432888865470886
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,32,3584,0.007194666398896112
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,32,4096,0.04497866829236349
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,32,3072,0.006839111033413146
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,32,3584,0.04054755634731717
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,32,2560,0.006175999840100606
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,32,3072,0.03585422370168898
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,32,2048,0.005506666584147348
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,32,2560,0.030784888399971858
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,128,1024,0.03168088859981961
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,32,2048,0.02719822194841173
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,32,1536,0.004798222333192825
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,32,1024,0.00407911092042923
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,32,1536,0.023936000135209825
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,128,768,0.030976888206270006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,32,768,0.003752888904677497
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,32,1024,0.02001066671477424
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,32,768,0.01830577850341797
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,32,512,0.0033760000434186724
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,32,256,0.0030577776746617425
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,32,512,0.016467556357383728
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,32,256,0.014680888917711047
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,32,128,0.002834666727317704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,32,128,0.013956444131003486
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,32,64,0.0026702222724755607
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,32,64,0.013823111024167804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,768,32,32,0.0026737778551048706
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,768,32,32,0.013780444032616086
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,128,512,0.028078221612506445
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,128,256,0.02754666739039951
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,65536,12288,0.5441439946492513
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,65536,12288,0.6827511257595487
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,65536,16384,0.9309173160129123
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,65536,16384,0.7702373398674859
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,65536,10240,0.45696269141303164
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,65536,10240,0.8280026647779676
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,65536,10240,0.5726515452067057
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,65536,12288,0.9328230751885308
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,65536,16384,1.3534800211588542
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,65536,8192,0.38398488362630206
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,65536,8192,0.4595315721299913
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,65536,8192,0.6022293302747939
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,65536,7168,0.31946754455566406
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,65536,6144,0.2729244497087267
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,65536,7168,0.5301031006707085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,65536,7168,0.4059973292880588
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,65536,6144,0.45695998933580184
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,768,128,128,0.02693866689999898
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,65536,6144,0.3521840042538113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,65536,5120,0.2331599924299452
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,65536,5120,0.3821813265482585
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,65536,3584,0.16844711038801405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,65536,4096,0.192267550362481
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,65536,5120,0.2983022265964084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,65536,3584,0.22131733099619547
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,65536,4096,0.31413155131869847
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,65536,4096,0.2469902303483751
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,65536,3072,0.14576000637478298
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,65536,3584,0.27481600973341197
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,65536,2560,0.12609422206878662
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,65536,2560,0.17058755291832817
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,65536,3072,0.19572799735599092
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,65536,2048,0.16727911101447213
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,65536,2048,0.10476622316572402
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,65536,3072,0.2370133399963379
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,65536,2048,0.14426399601830378
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,65536,1536,0.0857342216703627
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,65536,1024,0.06661333640416463
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,65536,2560,0.20214311281840006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,65536,1536,0.12810666031307644
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,65536,768,0.057632002565595836
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,65536,1024,0.09191733598709106
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,65536,1536,0.11713243855370416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,65536,768,0.07374488645129733
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,65536,1024,0.09187733464770848
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,65536,512,0.049534221490224205
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,65536,768,0.07759555843141344
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,65536,256,0.03711288836267259
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,65536,128,0.030117332935333252
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,65536,256,0.04193955659866333
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,65536,512,0.06573333342870076
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,65536,128,0.039451556073294744
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,65536,256,0.05580621957778931
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,65536,64,0.02645066711637709
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,65536,128,0.052982221047083534
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,65536,32,0.027102222045262653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,65536,64,0.04081155525313483
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,65536,512,0.05598400036493937
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,65536,32,0.042360888587103955
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,16384,12288,0.17874222331576875
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,16384,16384,0.3104337851206462
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,16384,16384,0.24168711238437227
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,16384,12288,0.23934311336941191
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,16384,16384,0.29952534039815265
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,16384,12288,0.22707200050354004
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,16384,10240,0.15013866954379612
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,16384,10240,0.18998755349053276
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,16384,10240,0.20385332902272543
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,16384,8192,0.1264479955037435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,16384,8192,0.15358044041527644
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,16384,8192,0.17026044262780082
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,16384,7168,0.11203377776675755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,16384,7168,0.15222222275204128
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,16384,7168,0.1356524493959215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,16384,6144,0.09710933102501763
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,16384,6144,0.1357182264328003
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,16384,6144,0.11703466044531928
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,16384,5120,0.08490222030215794
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,16384,65536,0.9880755742390951
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,16384,4096,0.06936266687181261
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,16384,5120,0.11760621600680882
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,16384,4096,0.08102577924728394
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,16384,4096,0.09867378075917561
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,16384,5120,0.09955288966496785
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,16384,3584,0.06374755832884046
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,16384,3072,0.057371556758880615
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,16384,3584,0.09051377905739678
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,16384,3584,0.07242400116390653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,16384,3072,0.06233955754174126
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,16384,3072,0.0820195542441474
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,16384,2560,0.05117333266470167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,16384,2560,0.05331644415855408
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,16384,2048,0.04236355423927307
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,16384,2560,0.07434578074349298
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,16384,1536,0.03587822119394938
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,16384,2048,0.04448711209826999
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,16384,1536,0.0548986660109626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,16384,1024,0.02958933181232876
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,16384,1536,0.03528533379236857
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,16384,2048,0.06405778063668145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,16384,1024,0.04674488968319363
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,16384,768,0.021020443903075323
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,16384,768,0.026735110415352717
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,16384,512,0.016075554821226332
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,16384,1024,0.025557332568698462
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,16384,768,0.04234666625658671
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,16384,512,0.02461333407296075
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,16384,256,0.011538666983445486
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,16384,256,0.021372444099850122
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,16384,512,0.03911644551489089
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,16384,65536,1.2312471601698134
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,16384,128,0.008799999952316284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,16384,256,0.03486844566133287
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,16384,128,0.018935110833909776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,16384,64,0.007375111182530721
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,16384,128,0.032263110081354775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,16384,32,0.00775733341773351
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,16384,32,0.020260444945759244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,16384,64,0.02146311104297638
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,12288,16384,0.20085155963897705
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,12288,16384,0.2658177746666802
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,12288,12288,0.15005689197116429
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,12288,16384,0.2297297848595513
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,12288,12288,0.1743919981850518
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,12288,12288,0.20468533039093018
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,12288,10240,0.12854044967227513
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,12288,10240,0.14559288819630942
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,12288,10240,0.17314578427208793
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,12288,8192,0.10804710785547893
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,12288,8192,0.11742933591206868
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,12288,7168,0.09493333101272583
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,12288,8192,0.14417333073086208
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,16384,65536,1.354498651292589
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,12288,65536,0.794473754035102
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,12288,7168,0.10442133082283868
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,12288,7168,0.13147111733754477
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,12288,6144,0.08293866448932223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,12288,6144,0.11875643995073105
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,12288,5120,0.07297066847483318
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,12288,6144,0.09077244334750706
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,12288,5120,0.10174400276607937
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,12288,5120,0.07634755637910631
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,12288,4096,0.06039288971159193
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,12288,4096,0.062055110931396484
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,12288,4096,0.08566488822301228
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,12288,3584,0.05525066786342197
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,12288,3584,0.05561333232455783
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,12288,3072,0.04900177650981479
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,12288,3072,0.0712311135398017
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,12288,3584,0.0782595541742113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,12288,3072,0.04889333248138428
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,12288,2048,0.03664799862437778
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,12288,2560,0.06615022155973646
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,12288,2560,0.04188088907135857
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,12288,2560,0.04511377877659268
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,12288,2048,0.05688799752129448
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,12288,1536,0.03173866536882188
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,12288,1024,0.02000977761215634
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,12288,2048,0.03447377681732178
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,12288,1024,0.027080888549486797
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,12288,1536,0.04970666766166687
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,12288,768,0.0166657782263226
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,12288,1024,0.042656001117494374
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,12288,1536,0.02721422248416477
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,12288,768,0.023922666907310486
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,12288,512,0.013059555656380124
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,12288,768,0.0399617784553104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,12288,512,0.021713778376579285
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,12288,256,0.00925688859489229
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,12288,512,0.03666755557060242
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,12288,256,0.019720888800091214
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,12288,128,0.007190222541491191
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,12288,256,0.03341066506173875
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,12288,128,0.017657778329319425
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,12288,64,0.006063111126422882
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,12288,128,0.030495110485288832
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,12288,32,0.006463999963468975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,12288,64,0.01796000036928389
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,12288,32,0.018096889058748882
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,12288,65536,0.993168936835395
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,10240,16384,0.19770577218797472
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,10240,12288,0.14946756098005506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,10240,16384,0.19151111443837485
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,10240,12288,0.1461040046479967
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,10240,10240,0.12166578239864773
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,10240,10240,0.1268142196867201
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,10240,65536,0.7887475755479602
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,10240,8192,0.10531110896004571
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,10240,8192,0.09838666518529256
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,12288,65536,1.049438264634874
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,10240,7168,0.08729866478178237
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,10240,7168,0.09702399704191421
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,10240,6144,0.07664089070426093
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,10240,6144,0.08233510785632663
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,10240,65536,0.8321190940009223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,10240,5120,0.0650177796681722
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,10240,5120,0.07360888852013482
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,10240,16384,0.24475733439127603
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,10240,4096,0.05265511075655619
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,10240,4096,0.059579557842678495
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,10240,12288,0.1888400051328871
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,10240,10240,0.16353066762288412
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,10240,3584,0.04718044400215149
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,10240,3584,0.05411110983954536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,10240,65536,0.9154258304172092
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,10240,3072,0.04138222336769104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,10240,3072,0.052058666944503784
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,10240,2560,0.03630400035116408
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,10240,8192,0.13570488823784724
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,10240,2560,0.042872889174355396
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,10240,7168,0.12315377924177383
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,10240,2048,0.02992977698644002
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,10240,2048,0.03626222080654568
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,10240,1536,0.023615999354256526
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,10240,1536,0.031003554662068684
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,10240,6144,0.10978044403923883
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,10240,1024,0.017602667212486267
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,10240,1024,0.02646488944689433
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,10240,5120,0.0963360004954868
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,10240,768,0.0144533332851198
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,10240,768,0.02416533397303687
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,10240,4096,0.08138933446672228
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,10240,512,0.011611555185582904
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,10240,512,0.021336000826623704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,10240,3584,0.07512978050443861
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,10240,256,0.008550222549173567
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,10240,256,0.019359111785888672
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,10240,2560,0.06255288918813069
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,10240,128,0.006551110910044775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,10240,128,0.017621333400408428
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,10240,1536,0.04760533240106371
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,10240,64,0.0058622223635514574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,10240,64,0.018180444836616516
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,10240,32,0.006120000034570694
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,10240,32,0.018248000078731112
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,10240,3072,0.06783911254670885
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,10240,768,0.038650665018293597
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,10240,2048,0.05475022064314949
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,8192,16384,0.15198311540815565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,10240,1024,0.04131288992034064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,8192,65536,0.6417537795172797
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,8192,16384,0.1595404413011339
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,8192,12288,0.1207360029220581
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,8192,16384,0.2186444467968411
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,8192,12288,0.11549954944186741
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,8192,65536,0.80276001824273
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,8192,12288,0.16947110493977866
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,8192,10240,0.10253689024183486
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,8192,10240,0.1461884445614285
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,8192,8192,0.08575110965304905
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,8192,10240,0.09709688689973618
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,8192,8192,0.07872977521684435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,8192,8192,0.12245244450039333
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,8192,7168,0.06958488623301189
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,8192,7168,0.07760000228881836
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,8192,7168,0.11071555482016669
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,8192,6144,0.06055555740992228
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,8192,6144,0.0682720012134976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,8192,6144,0.09847733047273423
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,8192,5120,0.059975114133622914
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,8192,5120,0.05134222242567274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,8192,5120,0.08672977818383111
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,8192,4096,0.04956533180342781
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,8192,3584,0.04534577661090427
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,10240,128,0.030057777961095173
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,8192,4096,0.07396355602476332
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,8192,4096,0.04224977890650431
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,8192,3584,0.03771377934349908
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,8192,3072,0.040878223048316106
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,8192,3584,0.06800355513890584
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,10240,256,0.0325164430671268
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,8192,3072,0.062337776025136314
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,8192,2560,0.03889155719015334
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,8192,2560,0.05841244591606987
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,8192,2048,0.031311111317740545
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,8192,3072,0.032997333341174655
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,8192,2560,0.02853155467245314
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,8192,1536,0.0192284451590644
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,8192,2048,0.023800889650980633
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,8192,2048,0.049752000305387706
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,8192,1536,0.027458666099442378
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,8192,1024,0.014468444718254937
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,8192,1536,0.043848888741599194
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,8192,1024,0.023215999205907185
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,8192,768,0.012227555943859948
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,8192,1024,0.038486222426096596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,8192,768,0.021271111236678228
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,8192,512,0.009729777773221334
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,8192,768,0.03646755549642775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,8192,512,0.019724445210562814
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,8192,65536,0.5930968920389811
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,8192,256,0.007198221981525421
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,8192,512,0.03382222188843621
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,8192,128,0.01608711149957445
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,8192,256,0.016777777009540133
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,8192,128,0.0058133333093590206
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,8192,256,0.029930667744742498
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,8192,64,0.00508355556262864
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,8192,32,0.005624889085690181
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,8192,64,0.01640266676743825
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,8192,128,0.029867556360032823
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,8192,32,0.016372443901167977
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,7168,16384,0.15960710578494602
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,7168,16384,0.21812444263034395
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,7168,16384,0.14834045039282903
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,10240,512,0.03559289044804043
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,7168,12288,0.1078888906372918
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,7168,12288,0.12080444229973687
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,7168,12288,0.16892355018191865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,7168,10240,0.102874669763777
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,7168,10240,0.14501599470774332
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,7168,8192,0.08541866805818345
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,7168,10240,0.08446755674150254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,7168,8192,0.12135999732547337
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,7168,65536,0.6087564362419976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,7168,8192,0.07648089196946886
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,7168,7168,0.07688977983262804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,7168,7168,0.1097502244843377
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,7168,7168,0.06068978044721815
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,7168,6144,0.0680053300327725
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,7168,6144,0.05290133423275418
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,7168,4096,0.049249778191248574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,7168,5120,0.04494222336345249
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,7168,6144,0.09771022531721328
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,7168,5120,0.05915021896362305
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,7168,5120,0.08604355653127034
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,7168,4096,0.07366844680574205
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,7168,3584,0.04507111178504097
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,7168,4096,0.03706577751371596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,7168,65536,0.8005519972907172
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,7168,3072,0.04062844316164652
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,7168,3584,0.06820355521308051
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,7168,3072,0.06207288636101616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,7168,3584,0.032615999380747475
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,7168,2560,0.037773334317737155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,7168,2048,0.02049777739577823
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,7168,3072,0.028535998529858057
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,7168,2048,0.03143289022975498
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,7168,2560,0.058295110861460366
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,7168,2560,0.024664888779322307
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,7168,2048,0.05014222198062473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,7168,1536,0.01680444512102339
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,7168,1536,0.02714666724205017
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,7168,1024,0.01257777793539895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,7168,1536,0.04412533177269829
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,7168,1024,0.023349333140585158
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,7168,768,0.01033688916100396
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,7168,1024,0.039022223816977605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,7168,768,0.021363554729355708
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,7168,512,0.008248888783984715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,7168,768,0.037103109889560275
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,7168,512,0.018551111221313477
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,7168,256,0.00602311102880372
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,7168,512,0.032594667540656194
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,7168,128,0.01626311077011956
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,7168,128,0.004981333182917701
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,7168,256,0.016972444123691983
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,7168,256,0.030433777305814955
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,7168,64,0.004856889032655292
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,7168,32,0.005197333378924264
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,7168,64,0.01626755628320906
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,7168,128,0.029654221402274236
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,7168,32,0.01613333324591319
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,6144,16384,0.15963821940951878
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,6144,16384,0.19359289275275338
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,7168,65536,0.5923173162672255
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,6144,12288,0.12075200345781113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,6144,12288,0.0891466670566135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,6144,12288,0.14981777138180202
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,6144,16384,0.116047117445204
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,6144,10240,0.07516622543334961
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,6144,10240,0.10311110814412434
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,6144,10240,0.12950311766730413
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,6144,8192,0.08489333258734809
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,6144,8192,0.10907289054658677
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,6144,8192,0.060805333985222705
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,6144,7168,0.05381866627269321
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,6144,7168,0.07663110891977946
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,6144,6144,0.0672524438963996
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,6144,65536,0.6067306730482314
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,6144,7168,0.09891288810306126
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,6144,5120,0.058712886439429395
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,6144,6144,0.0882071124182807
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,6144,6144,0.04736444354057312
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,6144,4096,0.0488391121228536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,6144,5120,0.07830044296052721
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,6144,4096,0.06686933173073663
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,6144,3584,0.045033779409196645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,6144,5120,0.04011199871699015
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,6144,3584,0.061943113803863525
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,6144,3072,0.041308446062935725
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,6144,3584,0.02941333254178365
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,6144,4096,0.03299466768900553
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,6144,2560,0.022375999225510493
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,6144,3072,0.05763644642300076
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,6144,2560,0.03669155637423197
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,6144,3072,0.025826666090223525
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,6144,2048,0.0188737776544359
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,6144,2560,0.05339822173118591
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,6144,2048,0.031145777967241075
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,6144,1536,0.015409777561823526
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,6144,65536,0.7048426734076606
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,6144,2048,0.046514666742748685
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,6144,1536,0.027107555005285475
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,6144,1024,0.011690666278203329
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,6144,768,0.00998222248421775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,6144,1024,0.0232933329211341
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,6144,1536,0.04121333360671997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,6144,512,0.007897777689827813
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,6144,768,0.02013511127895779
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,6144,1024,0.036730666955312095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,6144,768,0.03484355409940084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,6144,512,0.019377777973810833
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,6144,256,0.0058559998869895935
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,6144,512,0.030906667311986286
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,6144,256,0.0170817772547404
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,6144,128,0.00500177757607566
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,6144,256,0.028959999481836956
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,6144,128,0.016114667057991028
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,6144,64,0.004443555656406614
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,6144,128,0.028370665179358587
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,6144,64,0.016229333149062265
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,6144,32,0.004734222259786394
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,6144,32,0.01612622208065457
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,5120,16384,0.1587031152513292
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,5120,16384,0.186008890469869
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,6144,65536,0.452498647901747
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,5120,16384,0.10186311271455552
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,5120,12288,0.11988622612423366
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,5120,12288,0.14536444346110025
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,5120,12288,0.07748355468114217
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,5120,10240,0.10250044531292385
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,5120,10240,0.0655004448360867
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,5120,10240,0.1250400013393826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,5120,8192,0.08721599976221721
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,5120,8192,0.1054240001572503
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,5120,8192,0.05330222182803684
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,5120,7168,0.07625333468119304
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,5120,7168,0.047431998782687716
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,5120,6144,0.06709955798255073
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,5120,6144,0.08602933088938396
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,5120,5120,0.058627552456325956
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,5120,7168,0.09556533230675592
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,5120,5120,0.07591378026538424
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,5120,6144,0.041505777173572116
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,5120,5120,0.03545511100027297
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,5120,4096,0.04843466811709934
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,5120,4096,0.06514755884806316
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,5120,4096,0.029221332735485498
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,5120,3584,0.044850667317708336
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,5120,3584,0.06010755565431383
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,5120,3584,0.02647999922434489
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,5120,3072,0.02334311107794444
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,5120,3072,0.04067822297414144
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,5120,2560,0.020212444994184706
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,5120,3072,0.05535911189185249
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,5120,2560,0.03620355659061008
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,5120,2048,0.01713511182202233
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,5120,65536,0.609191099802653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,5120,2560,0.05237244566281637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,5120,2048,0.030607110924190942
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,5120,1536,0.014160000615649752
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,5120,1536,0.026742221580611333
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,5120,2048,0.04449244340260824
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,5120,1024,0.011113777756690979
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,5120,1024,0.022231111923853557
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,5120,1536,0.04063377777735392
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,5120,768,0.009657777845859528
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,5120,1024,0.034565332863065935
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,5120,768,0.020631111330456205
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,5120,512,0.007704888780911763
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,5120,768,0.03323555654949612
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,5120,512,0.018160889546076458
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,5120,256,0.006021333237489064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,5120,512,0.030192000998391047
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,5120,256,0.0165928883685006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,5120,128,0.00490133340160052
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,5120,256,0.02833599845568339
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,5120,128,0.015928000211715698
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,5120,65536,0.6577742364671495
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,5120,64,0.004309333446953031
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,5120,128,0.02805866797765096
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,5120,32,0.004566222015354368
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,5120,64,0.016051555673281353
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,5120,32,0.015740444262822468
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,4096,16384,0.1219831175274319
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,5120,65536,0.39175645510355633
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,4096,12288,0.059902224275800914
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,4096,16384,0.07813333140479194
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,4096,12288,0.09249333540598552
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,4096,10240,0.08013777600394355
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,4096,10240,0.05063910947905647
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,4096,8192,0.06697511010699801
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,4096,8192,0.041708442899915904
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,4096,7168,0.0372559991147783
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,4096,7168,0.06135288874308268
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,4096,65536,0.45461421542697483
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,4096,6144,0.03247200118170844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,4096,6144,0.05390755666626824
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,4096,65536,0.3067733446756999
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,4096,5120,0.0277128881878323
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,4096,5120,0.04770311050944858
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,4096,16384,0.17259466648101807
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,4096,4096,0.022837332553333704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,4096,4096,0.03964799973699782
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,4096,12288,0.13531555069817439
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,4096,3584,0.020785777105225455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,4096,3584,0.03707289033465915
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,4096,10240,0.11678400304582383
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,4096,3072,0.01833777791923947
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,4096,3072,0.033519112401538424
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,4096,8192,0.09816355837715997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,4096,2560,0.015928000211715698
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,4096,2560,0.030404445197847154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,4096,65536,0.6248817973666722
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,4096,2048,0.013703111145231458
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,4096,7168,0.08905866410997178
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,4096,2048,0.026569777064853247
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,4096,1536,0.011339555184046427
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,4096,1536,0.023743111226293776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,4096,6144,0.08035288916693793
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,4096,1024,0.00906666699382994
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,4096,1024,0.020420443680551317
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,4096,5120,0.07143821981218126
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,4096,768,0.007453333172533248
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,4096,768,0.018090666996108163
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,4096,4096,0.060995552274915904
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,4096,512,0.005953777581453323
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,4096,3584,0.05675466855367025
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,4096,512,0.016688888271649677
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,4096,256,0.004865777575307422
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,4096,256,0.015573332707087198
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,4096,3072,0.05250755614704556
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,4096,128,0.004312889029582341
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,4096,128,0.01479555500878228
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,4096,2560,0.04894577794604831
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,4096,64,0.0038702223036024305
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,4096,64,0.014836443795098199
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,4096,32,0.004109333372778363
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,4096,32,0.014512888259357877
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,4096,2048,0.04327822062704298
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,4096,1536,0.038867556386523776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,4096,1024,0.03477599885728624
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3584,65536,0.4490150875515408
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3584,16384,0.07121866941452026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3584,16384,0.1218328873316447
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3584,12288,0.05256355471081204
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3584,65536,0.2782595687442356
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3584,12288,0.09192888604270087
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,4096,768,0.03221244282192654
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3584,10240,0.044665777020984225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3584,10240,0.0786951118045383
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3584,8192,0.036186665296554565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3584,8192,0.06589600112703112
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,4096,512,0.029529776838090684
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,4096,256,0.028032888968785603
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,4096,128,0.02767555581198798
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3584,7168,0.060221334298451744
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3584,7168,0.032384001546435885
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3584,6144,0.028396444188223943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3584,6144,0.0531217787000868
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3584,5120,0.02440711193614536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3584,5120,0.04694844285647074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,3584,16384,0.1778808832168579
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3584,4096,0.020473778247833252
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3584,4096,0.03925866550869412
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,3584,65536,0.6417333285013834
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3584,3584,0.018319111731317308
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3584,3584,0.03619822197490268
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,3584,12288,0.13636710908677843
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3584,3072,0.016383111476898193
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3584,3072,0.03372622198528714
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,3584,10240,0.11780888504452175
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3584,2560,0.014461333552996317
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3584,2560,0.030559112628300984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,3584,8192,0.09847288661532932
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3584,2048,0.012296000288592445
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3584,2048,0.026540444956885442
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,3584,7168,0.0898231135474311
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3584,1536,0.010296888649463654
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3584,1536,0.02346666653951009
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,3584,6144,0.08055200179417928
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3584,1024,0.008185777399275038
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3584,1024,0.019392000304328073
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,3584,5120,0.07204266389211018
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3584,768,0.006714666469229593
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3584,768,0.017958222164048087
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,3584,4096,0.061703999837239586
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3584,512,0.005594666633341048
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3584,512,0.016602666841612924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,3584,3584,0.05720800161361694
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3584,256,0.004597333156400257
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3584,256,0.015365334020720588
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,3584,3072,0.052617778380711876
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3584,128,0.004055110944641961
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3584,128,0.014674666855070325
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,3584,2560,0.049426668220096164
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3584,64,0.003751111113362842
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3584,64,0.014536889062987434
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3584,32,0.003944000022278892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3584,32,0.014504000544548035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,3584,2048,0.04361688759591845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,3584,1536,0.039056890540652804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,3584,768,0.0314026673634847
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,3584,1024,0.03323288758595785
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3072,65536,0.44826846652560765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3072,16384,0.060312887032826744
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3072,16384,0.1202195617887709
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3072,65536,0.23625599013434517
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3072,12288,0.090902222527398
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3072,12288,0.04573066698180305
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3072,10240,0.07815200090408325
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3072,10240,0.03898933198716905
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3072,8192,0.03199199835459391
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3072,8192,0.06553422080145942
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,3584,512,0.029486222399605647
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,3584,256,0.028010666370391846
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3072,7168,0.057800889015197754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3072,7168,0.028745777077145044
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,3584,128,0.02784622377819485
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3072,6144,0.025176000263955858
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3072,6144,0.052592890130149
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3072,5120,0.021632888250880774
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3072,5120,0.04686399963166979
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,3072,16384,0.1712986628214518
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3072,4096,0.018210666047202218
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3072,4096,0.03906222184499105
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,3072,12288,0.13376000192430285
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3072,3584,0.0163102216190762
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,3072,10240,0.11555200152926975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3072,3584,0.03588355580965678
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3072,3072,0.014608888162506951
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3072,3072,0.03283199999067519
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,3072,65536,0.6087306870354546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3072,2560,0.012998222476906247
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3072,2560,0.03010222315788269
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,3072,8192,0.09736800193786621
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3072,2048,0.011184888581434885
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3072,2048,0.02632355524433984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,3072,7168,0.08872710996203953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3072,1536,0.009356444080670675
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3072,1536,0.023187556200557288
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,3072,6144,0.07979466517766316
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3072,1024,0.0074026667409473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3072,1024,0.019285332825448778
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,3072,5120,0.07053511010275947
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3072,768,0.006265777680608962
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3072,768,0.018054222067197163
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,3072,3584,0.056346668137444385
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3072,512,0.005311111195219888
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3072,512,0.016570667425791424
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,3072,2560,0.04886755678388807
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3072,256,0.004415110995372136
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3072,256,0.015392000476519266
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,3072,4096,0.06074844466315376
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3072,128,0.003967111309369405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3072,128,0.014626666903495789
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,3072,768,0.03174933460023668
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3072,64,0.0035706667436493766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3072,64,0.014647111296653748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,3072,32,0.0037617778612507712
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,3072,32,0.014243554737832812
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,3072,3072,0.051924443907207914
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,3072,2048,0.042893333567513355
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,3072,1536,0.03849511014090644
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2560,16384,0.06671910815768771
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2560,65536,0.4465057584974501
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2560,65536,0.2039395570755005
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2560,16384,0.12032711505889893
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2560,12288,0.09056444300545587
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,3072,1024,0.03438399897681342
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2560,12288,0.04381422201792399
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2560,10240,0.07705866628222995
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2560,10240,0.0341635545094808
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2560,8192,0.028385778268178303
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2560,8192,0.06465333037906222
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,3072,128,0.026947556270493403
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2560,7168,0.025928889711697895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2560,7168,0.05810044209162394
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,3072,256,0.027484445108307734
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2560,6144,0.021597334080272253
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2560,6144,0.05231377813551161
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,3072,512,0.02920711040496826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2560,5120,0.018887110882335238
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2560,5120,0.04624888963169522
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,2560,16384,0.16702489058176676
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2560,4096,0.015802666544914246
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2560,4096,0.03850488861401876
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,2560,12288,0.13098043865627712
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2560,3584,0.014410666293568082
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2560,3584,0.035459554857677884
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,2560,10240,0.11320443948109944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2560,3072,0.012794666820102267
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2560,3072,0.03276622295379639
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,2560,65536,0.5847057766384548
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2560,2560,0.011374221907721626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2560,2560,0.02998133169280158
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,2560,8192,0.09566044145160252
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2560,2048,0.00998933364947637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2560,2048,0.026205332742797002
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,2560,6144,0.07817066378063627
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,2560,7168,0.08708266417185466
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2560,1536,0.008641777767075432
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2560,1536,0.022296888960732356
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2560,1024,0.0063688887490166565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2560,1024,0.019090667366981506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,2560,5120,0.06996177964740329
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2560,768,0.0058853332367208265
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2560,768,0.0177502218219969
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,2560,3584,0.05594578054216173
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2560,512,0.004907555464241239
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2560,512,0.01643111142847273
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,2560,2560,0.04778133498297798
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2560,256,0.004222222086456087
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2560,256,0.015319110618697273
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,2560,4096,0.06004533502790663
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2560,128,0.0038524443904558816
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2560,128,0.014589332871966891
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,2560,3072,0.05086133215162489
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2560,64,0.003495111233658261
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2560,64,0.014567999376191033
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2560,32,0.0036915557252036203
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2560,32,0.014329777823554145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,2560,2048,0.04205866654713949
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2048,65536,0.1605786614947849
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2048,65536,0.44762664371066624
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2048,16384,0.04787733488612705
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,2560,768,0.031039112144046362
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2048,16384,0.12054577138688828
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2048,12288,0.03860355416933695
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2048,12288,0.09080177545547485
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,2560,1536,0.03800355394681295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2048,10240,0.03637955586115519
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2048,10240,0.0767128864924113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,2560,1024,0.03308533297644721
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2048,8192,0.028343998723559912
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2048,8192,0.06446399953630236
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,2560,128,0.027154665854242113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2048,7168,0.024422221713595923
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2048,7168,0.05882755253050062
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,2560,256,0.027414222558339436
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2048,6144,0.02102488941616482
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2048,6144,0.051948444710837476
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,2560,512,0.028616888655556574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2048,5120,0.018231999542978074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2048,5120,0.044812444183561534
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,2048,65536,0.5541982120937771
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,2048,16384,0.1565635601679484
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2048,4096,0.01681422193845113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2048,4096,0.03888355692227682
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,2048,12288,0.12071200211842854
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2048,3584,0.014891554911931356
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2048,3584,0.03566133313708835
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2048,3072,0.012372444073359171
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2048,3072,0.03276977936426798
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,2048,10240,0.1050257749027676
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2048,2560,0.011424000064531961
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2048,2560,0.02944355540805393
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,2048,8192,0.08889600303437974
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2048,2048,0.009758222434255812
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2048,2048,0.02556266718440586
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,2048,7168,0.08122133546405368
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2048,1536,0.008178667061858708
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2048,1536,0.023033777872721355
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,2048,6144,0.07355378071467082
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2048,1024,0.006351999938488007
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2048,1024,0.018969777557584975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,2048,5120,0.06557955344518025
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2048,768,0.005614222337802251
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2048,768,0.017656000124083627
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,2048,4096,0.056596444712744824
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,2048,3584,0.0527617765797509
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2048,512,0.0047315553658538395
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2048,512,0.01628266606065962
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2048,256,0.004117333226733738
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2048,256,0.015287111202875773
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,2048,3072,0.04892355534765455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2048,128,0.003654222107595868
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2048,128,0.014574221438831754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,2048,2560,0.04639022217856514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2048,64,0.003367111086845398
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2048,64,0.014236445228258768
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,2048,32,0.0035413333939181436
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,2048,32,0.014097778333557976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1536,65536,0.12309243943956162
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,2048,2048,0.04067822297414144
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1536,65536,0.4474017884996202
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1536,16384,0.03802488909827338
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1536,16384,0.12065333790249294
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,2048,1536,0.036835554573271014
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,2048,1024,0.03111999895837572
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1536,12288,0.030540443129009668
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1536,12288,0.09010933505164252
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1536,10240,0.02941688895225525
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1536,10240,0.07702844672732882
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,2048,768,0.029932445949978296
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1536,8192,0.025224889318148296
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1536,8192,0.06293155749638875
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,2048,256,0.02720533311367035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1536,7168,0.02154133386082119
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1536,7168,0.058746669027540416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,2048,128,0.026664000418451097
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,2048,512,0.028338667419221666
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1536,6144,0.017467555072572496
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1536,6144,0.05183822247717115
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1536,5120,0.016200888488027785
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1536,5120,0.04514133267932468
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,1536,16384,0.15843733151753744
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1536,4096,0.015215999550289579
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,1536,65536,0.558459546830919
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1536,4096,0.03782666722933451
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1536,3584,0.01275022245115704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1536,3584,0.03502666619088914
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,1536,12288,0.12253244717915852
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1536,3072,0.010485333700974783
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1536,3072,0.032320890161726214
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,1536,10240,0.1059982246822781
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1536,2560,0.009485333330101436
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1536,2560,0.0295448899269104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,1536,8192,0.08998222483528985
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1536,2048,0.008028444316652086
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1536,2048,0.02462399999300639
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,1536,7168,0.08221244149737887
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1536,1536,0.006863111009200414
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1536,1536,0.02181244393189748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,1536,6144,0.07404800256093343
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1536,1024,0.0053173332578606075
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1536,1024,0.01884977850649092
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,1536,5120,0.06569244464238484
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1536,768,0.004829333474238713
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1536,768,0.01744888888465034
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,1536,3584,0.0522035559018453
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1536,512,0.004228444563017951
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1536,512,0.01610133383009169
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,1536,4096,0.056377775139278836
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1536,256,0.003699555579158995
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1536,256,0.015055111712879606
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,1536,2560,0.045399109522501625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1536,128,0.003383111208677292
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,1536,3072,0.04822399881150988
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1536,128,0.014470222923490735
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1536,64,0.0031475555151700974
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1536,64,0.014337778091430664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1536,32,0.003273777870668305
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1536,32,0.014170666535695394
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1024,65536,0.08418755398856269
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,1536,1536,0.0351893338892195
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1024,65536,0.44750308990478516
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,1536,2048,0.04033244318432278
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1024,16384,0.12018933561113145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1024,16384,0.028428445259730022
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1024,12288,0.023401778605249193
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,1536,768,0.02951288885540432
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,1536,1024,0.03042311138576931
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1024,12288,0.09035911162694295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1024,10240,0.02127911150455475
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1024,10240,0.07677777608235677
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1024,8192,0.018950222267044913
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1024,8192,0.06415822108586629
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,1536,256,0.026711111267407734
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1024,7168,0.014866666661368476
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1024,7168,0.0579342246055603
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,1536,128,0.02666666607062022
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1024,6144,0.013380444712109037
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1024,6144,0.0507279998726315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,1536,512,0.028323554330401953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1024,5120,0.012920888761679331
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1024,5120,0.04435555471314324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,1024,16384,0.14481866359710693
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1024,4096,0.011046222514576383
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1024,4096,0.03792000148031447
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,1024,12288,0.11210489273071289
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1024,3584,0.009301332963837517
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,1024,65536,0.5119848781161839
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,1024,10240,0.09713155693478054
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1024,3584,0.0346942212846544
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1024,3072,0.008453333543406593
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1024,3072,0.03155288762516446
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1024,2560,0.007343110938866933
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,1024,8192,0.08345777458614773
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1024,2560,0.028354667954974707
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1024,2048,0.006653333289755716
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1024,2048,0.024498666326204937
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,1024,7168,0.076328886879815
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1024,1536,0.005744888964626525
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1024,1536,0.021496888664033677
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,1024,6144,0.06879378027386136
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1024,1024,0.004831999954250124
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1024,1024,0.018623110320832994
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,1024,5120,0.06088266770044962
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1024,768,0.004487110922733943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1024,768,0.017343110508388944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,1024,4096,0.05324177940686544
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1024,512,0.003974222060706881
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,1024,3584,0.04942489001486036
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1024,512,0.015863999724388123
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1024,256,0.003563555578390757
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1024,256,0.014731556177139282
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,1024,3072,0.045943111181259155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1024,128,0.0033155555526415506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,1024,2560,0.04353866643375821
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1024,128,0.014210666219393412
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,1024,2048,0.0389342208703359
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1024,64,0.003118222165438864
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1024,64,0.01404355631934272
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,1024,32,0.0032702222880389956
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,1024,32,0.013998222019937305
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,768,65536,0.06742666827307807
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,768,65536,0.4464249081081814
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,768,16384,0.023762666516833838
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,768,16384,0.12069066365559895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,1024,1536,0.033448888195885554
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,768,12288,0.019023999571800232
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,768,12288,0.08997955587175156
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,1024,1024,0.030047112041049536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,768,10240,0.019010666343900893
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,768,10240,0.07646310991711087
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,1024,768,0.02914311157332526
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,768,8192,0.01793155570824941
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,768,8192,0.0634746683968438
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,1024,512,0.027635556128289964
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,1024,256,0.026611554953787062
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,768,7168,0.014751111467679342
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,768,7168,0.05780711438920763
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,1024,128,0.026496888862715826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,768,6144,0.013072888884279462
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,768,6144,0.05083466569582621
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,768,5120,0.011583111352390714
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,768,5120,0.04429777794414096
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,768,16384,0.14341155687967935
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,768,4096,0.010048000348938836
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,768,65536,0.5061679946051704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,768,4096,0.03735466798146566
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,768,3584,0.009323555562231276
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,768,3584,0.034372445609834455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,768,12288,0.11089244153764512
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,768,3072,0.008358221915033128
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,768,3072,0.03086044390996297
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,768,10240,0.09623110956615871
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,768,2560,0.007600888609886169
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,768,2560,0.028008888165156048
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,768,8192,0.08258755339516534
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,768,2048,0.00646666685740153
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,768,2048,0.02419822249147627
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,768,7168,0.07541066408157349
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,768,1536,0.005688000056478713
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,768,6144,0.0680053300327725
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,768,1536,0.0214355554845598
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,768,1024,0.004765333400832282
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,768,1024,0.01868977811601427
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,768,5120,0.060736887984805636
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,768,768,0.004328000048796336
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,768,768,0.01734133395883772
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,768,4096,0.05260089039802551
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,768,512,0.0038817777401871155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,768,512,0.01588444411754608
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,768,3584,0.04900266726811727
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,768,256,0.003419555516706573
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,768,256,0.01478488908873664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,768,3072,0.04553066690762838
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,768,128,0.003181333343187968
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,768,128,0.014096000128322177
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,768,2560,0.042825778325398765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,768,64,0.003028444531891081
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,768,64,0.014176888598336114
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,768,32,0.0030666666312350165
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,768,32,0.0138568886452251
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,768,2048,0.03657244311438666
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,512,65536,0.048523555199305214
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,512,65536,0.44721778233846027
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,512,16384,0.01962222158908844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,512,16384,0.12078400452931722
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,768,1536,0.032962666617499456
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,768,1024,0.02981333269013299
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,512,12288,0.015392000476519266
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,512,12288,0.09068622191747029
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,512,10240,0.01404622197151184
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,512,10240,0.07691999938752916
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,768,768,0.029098666376537744
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,512,8192,0.013502221968438892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,512,8192,0.06324888600243463
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,768,512,0.027481777800454035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,512,7168,0.012347555822796293
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,512,7168,0.056327111191219754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,768,256,0.026564444104830425
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,512,6144,0.011393778026103973
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,512,6144,0.05060799916585287
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,768,128,0.02641955514748891
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,512,5120,0.010463111102581024
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,512,5120,0.044255110952589244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,512,16384,0.14122045040130615
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,512,4096,0.009333333207501305
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,512,4096,0.03710755705833435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,512,65536,0.4938986566331651
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,512,3584,0.008077333370844523
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,512,3584,0.03401777810520596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,512,12288,0.10907555288738674
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,512,3072,0.007138667007287343
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,512,3072,0.031249778138266668
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,512,10240,0.09540888998243545
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,512,2560,0.007271111011505127
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,512,8192,0.08153777652316622
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,512,2560,0.027355555031034682
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,512,2048,0.006630222416586346
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,512,7168,0.07452444235483806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,512,2048,0.02411288850837284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,512,1536,0.005798222290145026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,512,1536,0.021369778447681006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,512,6144,0.0668791135152181
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,512,1024,0.004671111288997862
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,512,1024,0.018438221679793466
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,512,5120,0.059593776861826576
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,512,768,0.004378666480382283
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,512,768,0.017136888371573556
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,512,4096,0.051743110020955406
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,512,512,0.003834666477309333
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,512,512,0.01570133368174235
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,512,3584,0.04806844393412272
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,512,256,0.003412444351447953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,512,256,0.014769777655601501
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,512,3072,0.0451128880182902
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,512,128,0.003100444459252887
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,512,128,0.014078223043017916
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,512,2560,0.041831112570232816
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,512,64,0.0029440000653266907
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,512,64,0.014017777310477363
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,512,32,0.00299822228650252
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,512,32,0.013731555806265937
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,256,65536,0.04539377821816338
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,256,65536,0.44719733132256406
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,512,2048,0.035691556003358625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,256,16384,0.012785777449607849
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,256,16384,0.12072000238630508
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,512,1536,0.03260711166593764
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,256,12288,0.01090222183201048
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,512,1024,0.029468443658616807
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,256,12288,0.08948977788289388
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,256,10240,0.010021333065297868
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,256,10240,0.07615111271540324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,512,768,0.028452446063359577
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,256,8192,0.010041777458455827
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,256,8192,0.06308799982070923
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,512,256,0.02643288837538825
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,512,512,0.027423999375767175
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,256,7168,0.010051555931568146
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,256,7168,0.05694933070076836
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,256,6144,0.008919999831252629
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,256,6144,0.05043911271625095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,512,128,0.02605244517326355
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,256,5120,0.008566222257084316
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,256,5120,0.04400622182422214
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,256,65536,0.49236265818277997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,256,4096,0.007830222447713217
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,256,4096,0.03664799862437778
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,256,16384,0.13966933886210123
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,256,3584,0.00775022225247489
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,256,3584,0.03253066539764404
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,256,12288,0.10796533028284709
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,256,3072,0.007271111011505127
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,256,3072,0.029469334416919287
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,256,10240,0.09425689114464654
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,256,2560,0.006516444600290722
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,256,2560,0.027012444204754297
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,256,8192,0.08032266961203681
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,256,2048,0.005753777921199799
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,256,2048,0.023894222246276006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,256,7168,0.07302133242289226
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,256,1536,0.005023111071851519
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,256,1536,0.021175111333529156
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,256,6144,0.06629866361618042
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,256,1024,0.004311999926964442
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,256,1024,0.018417777286635507
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,256,5120,0.05950755543178982
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,256,768,0.003927999900446998
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,256,768,0.017148445049921673
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,256,4096,0.05141777793566386
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,256,512,0.003579555700222651
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,256,512,0.015779554843902588
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,256,3584,0.04753333330154419
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,256,256,0.0032382222513357797
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,256,256,0.014519999424616495
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,256,3072,0.04203022188610501
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,256,128,0.002941333378354708
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,256,128,0.013869333598348828
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,256,2560,0.04012088974316915
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,256,64,0.0028213332924577924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,256,64,0.013878222141000958
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,256,32,0.002871111035346985
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,256,32,0.013583111266295115
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,128,65536,0.035387555758158364
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,128,65536,0.4473484357198079
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,256,2048,0.03527377711402045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,128,16384,0.010406222608354356
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,128,16384,0.1211164395014445
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,256,1536,0.032243556446499295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,128,12288,0.009394666386975182
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,128,12288,0.08918933073679607
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,256,1024,0.029249777396519978
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,128,10240,0.00850755555762185
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,256,768,0.028148445818159316
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,128,10240,0.07636800077226427
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,256,512,0.027094221777386133
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,128,8192,0.008845333423879411
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,128,8192,0.06335644589530097
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,128,7168,0.00832711077398724
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,128,7168,0.057189332114325635
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,256,256,0.026368000441127356
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,128,6144,0.007850666840871176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,128,6144,0.049357334772745766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,256,128,0.02642755541536543
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,128,5120,0.0074142225914531285
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,128,5120,0.043140444490644664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,128,65536,0.4927235709296332
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,128,4096,0.0069919998447100324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,128,4096,0.0351893338892195
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,128,16384,0.1396791140238444
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,128,3584,0.007261333366235097
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,128,3584,0.032279110617107816
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,128,12288,0.10880800088246663
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,128,3072,0.007002666592597961
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,128,3072,0.02936089038848877
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,128,10240,0.09406222237481011
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,128,2560,0.00629066675901413
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,128,2560,0.02676444417900509
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,128,8192,0.08025244209501478
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,128,2048,0.005616888817813661
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,128,2048,0.023775999744733173
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,128,7168,0.07303466399510701
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,128,1536,0.004872888740566042
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,128,1536,0.02110933264096578
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,128,6144,0.06639733579423693
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,128,1024,0.004345777961942884
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,128,1024,0.018417777286635507
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,128,5120,0.058916442924075656
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,128,768,0.003983111017280155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,128,768,0.016928889685206942
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,128,4096,0.049851556619008384
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,128,512,0.003698666476541095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,128,512,0.015623999966515435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,128,3584,0.045223110251956515
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,128,256,0.0032382222513357797
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,128,256,0.014440889159838358
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,128,3072,0.04197511076927185
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,128,128,0.0030053332448005676
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,128,128,0.01386844449573093
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,128,2560,0.03990488913324144
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,128,64,0.002847111059559716
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,128,64,0.013631111217869652
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,128,32,0.003022222262289789
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,128,32,0.013560888667901358
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,64,65536,0.027512888113657635
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,64,65536,0.4462204509311252
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,64,16384,0.009082666701740688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,64,16384,0.1203120019700792
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,64,12288,0.007993777592976889
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,64,12288,0.08987200260162354
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,64,10240,0.007610666255156199
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,64,10240,0.07599021991093953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,64,8192,0.007127111156781514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,64,8192,0.06282844146092732
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,64,7168,0.006955555743641323
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,64,7168,0.05684533384111193
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,64,6144,0.006697777658700943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,64,6144,0.049810667832692467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,64,5120,0.007069333559936947
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,64,5120,0.0425546665986379
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,64,4096,0.006664000037643645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,128,2048,0.03554755449295044
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,64,4096,0.03498133354716831
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,64,3584,0.007137777904669444
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,64,3072,0.006795555353164673
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,64,3584,0.0320720010333591
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,64,2560,0.006143110907740063
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,64,3072,0.029570665624406602
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,64,2048,0.005515555540720622
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,64,2560,0.026732444763183594
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,64,1536,0.0047786666287316214
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,64,2048,0.02379111117786831
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,64,1536,0.02102577851878272
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,64,1024,0.004108444270160464
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,128,1536,0.03228888909022013
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,128,1024,0.02924088968171014
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,64,768,0.0037671112351947357
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,64,1024,0.01831022236082289
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,64,768,0.01699911057949066
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,64,512,0.0034142221427626083
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,64,256,0.003014222201373842
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,64,512,0.015531554818153381
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,64,256,0.014504889647165934
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,64,128,0.0028853333658642242
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,64,128,0.013870221873124441
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,64,64,0.0027146666414207886
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,64,64,0.01371911085314221
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,64,32,0.0027253333893087176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,64,32,0.013591110706329346
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,32,16384,0.007809778054555257
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,32,65536,0.029630220598644678
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,32,12288,0.007182222273614671
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,32,16384,0.12039377954271104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,32,10240,0.0070426662762959795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,32,12288,0.0900497767660353
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,32,65536,0.44725778367784286
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,32,8192,0.006979555719428592
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,32,10240,0.07591911156972249
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,128,768,0.028279999891916912
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,32,7168,0.007184888753626082
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,32,8192,0.06304711103439331
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,32,6144,0.00664444433318244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,32,7168,0.05673155519697401
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,32,5120,0.006974221931563483
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,32,6144,0.04968177941110399
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,32,4096,0.006610666712125142
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,32,5120,0.04292177822854784
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,32,3584,0.007104888558387756
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,32,4096,0.03501155641343858
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,32,3072,0.006735111276308696
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,32,3584,0.032124443186653986
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,32,2560,0.006155555446942647
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,32,3072,0.029296000798543293
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,32,2048,0.005456888841258154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,32,2560,0.026787555880016748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,32,1536,0.00480266660451889
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,32,2048,0.023794665932655334
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,32,1024,0.004042666819360522
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,32,1536,0.021012443635198805
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,32,768,0.0037395556767781577
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,32,1024,0.018383999665578205
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,32,512,0.0033928888539473214
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,32,768,0.016931555337376065
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,32,256,0.0030497778207063675
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,32,512,0.01568800045384301
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,32,128,0.0028071111689011254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,32,256,0.014301333162519666
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,32,128,0.01384888920519087
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,32,64,0.0026666666898462507
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,128,512,0.027267555395762127
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,512,32,32,0.0026622222115596137
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,32,64,0.013696889082590738
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,512,32,32,0.01351288871632682
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,128,256,0.025979555315441553
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,65536,12288,0.8185902171664767
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,65536,16384,1.0937048594156902
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,65536,12288,0.718506654103597
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,65536,10240,0.6798924340142144
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,65536,10240,0.5986915694342719
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,65536,16384,0.9834480285644531
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,65536,8192,0.5455830891927084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,65536,8192,0.45906755659315324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,65536,7168,0.480754640367296
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,512,128,128,0.026027555267016094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,65536,7168,0.40340532196892637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,65536,6144,0.4107937812805176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,65536,6144,0.3481030993991428
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,65536,12288,0.5182204246520996
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,65536,5120,0.34292088614569766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,65536,16384,0.723490662044949
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,65536,5120,0.2921706570519341
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,65536,4096,0.2380586730109321
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,65536,4096,0.276546663708157
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,65536,3584,0.2428222232394748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,65536,3584,0.210371560520596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,65536,8192,0.35756177372402614
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,65536,10240,0.4391288757324219
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,65536,3072,0.2121244404051039
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,65536,3072,0.1840640041563246
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,65536,2560,0.15657333532969156
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,65536,2560,0.17441956202189127
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,65536,2048,0.1284337838490804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,65536,7168,0.31760621070861816
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,65536,2048,0.1440275510152181
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,65536,1536,0.1005013320181105
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,65536,6144,0.2739742332034641
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,65536,1536,0.11050666703118218
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,65536,1024,0.07251199748780993
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,65536,1024,0.07916266388363309
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,65536,5120,0.23410044776068795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,65536,768,0.059303998947143555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,65536,768,0.06331111325158013
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,65536,4096,0.19684888256920707
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,65536,512,0.045109334919187755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,65536,3584,0.17993422349294028
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,65536,512,0.04825422167778015
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,65536,256,0.031031999323103163
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,65536,256,0.0410426656405131
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,65536,3072,0.1566622257232666
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,65536,128,0.0247022228108512
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,65536,128,0.03831200136078729
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,65536,2560,0.13833422131008571
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,65536,64,0.021893334057595994
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,65536,64,0.040722668170928955
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,65536,32,0.02183466653029124
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,65536,32,0.043902221653196544
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,65536,1536,0.0978542235162523
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,65536,2048,0.11468444930182563
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,65536,256,0.05296355485916138
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,65536,512,0.06009422408209907
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,65536,128,0.04682933290799459
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,65536,1024,0.07763910955852933
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,65536,768,0.06939555539025201
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,16384,16384,0.31023910310533315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,16384,16384,0.2284017668830024
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,16384,12288,0.23413332303365073
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,16384,12288,0.17369777626461455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,16384,65536,1.2519475089179146
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,16384,10240,0.14640355110168457
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,16384,10240,0.1973031096988254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,16384,8192,0.15951022836897108
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,16384,8192,0.11917866600884332
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,16384,7168,0.14170666535695395
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,16384,7168,0.10420799917644924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,16384,65536,1.0419493781195748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,16384,6144,0.12443021933237712
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,16384,6144,0.0902586645550198
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,16384,5120,0.07646310991711087
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,16384,5120,0.10378399822447036
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,16384,12288,0.1890711122088962
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,16384,16384,0.2399982346428765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,16384,4096,0.08602933088938396
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,16384,4096,0.061979558732774526
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,16384,3584,0.05517333414819506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,16384,8192,0.13566933737860784
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,16384,3584,0.07730399900012545
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,16384,10240,0.16146666473812527
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,16384,65536,0.9256791008843316
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,16384,3072,0.06857066684299044
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,16384,3072,0.04849777619043986
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,16384,2560,0.057962669266594775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,16384,2560,0.041546665959888034
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,16384,7168,0.1208675569958157
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,16384,2048,0.03450489044189453
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,16384,2048,0.04986044433381823
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,16384,1536,0.027435556054115295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,16384,1536,0.04110488957828946
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,16384,6144,0.1088542209731208
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,16384,1024,0.020358221398459542
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,16384,1024,0.03228088882234361
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,16384,5120,0.09443822171952988
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,16384,768,0.01686844395266639
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,16384,768,0.027507555153634813
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,16384,3584,0.07703822188907199
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,16384,512,0.01309333327743742
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,16384,512,0.02418666746881273
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,16384,4096,0.08173955811394586
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,16384,256,0.0096133334769143
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,16384,256,0.02071911096572876
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,16384,2560,0.06295822064081828
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,16384,128,0.007375111182530721
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,16384,128,0.018971555762820773
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,16384,3072,0.06938488615883721
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,16384,64,0.006340444501903322
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,16384,64,0.019137778215938143
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,16384,32,0.006592888798978593
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,16384,32,0.01997866729895274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,16384,1536,0.04941244588957893
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,16384,768,0.03958933220969306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,16384,2048,0.05413155423270332
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,16384,1024,0.0420595539940728
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,12288,16384,0.2476444509294298
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,12288,16384,0.1694400045606825
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,12288,65536,0.9967173470391167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,12288,12288,0.1287813319100274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,12288,12288,0.1868000030517578
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,16384,128,0.029401779174804688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,16384,256,0.03380444314744737
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,12288,10240,0.15739999877081975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,12288,10240,0.10902577638626099
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,16384,512,0.0374017788304223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,12288,8192,0.12915555636088052
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,12288,8192,0.0878595577345954
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,12288,65536,0.681854248046875
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,12288,7168,0.11461510923173691
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,12288,7168,0.07814133167266846
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,12288,6144,0.10059377882215713
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,12288,6144,0.06749777661429511
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,12288,5120,0.057393776045905225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,12288,5120,0.08505688773261176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,12288,16384,0.22853689723544648
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,12288,4096,0.04719733198483785
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,12288,4096,0.07082221905390422
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,12288,65536,0.8607030974494086
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,12288,3584,0.04269688990381029
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,12288,3584,0.06375466452704535
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,12288,12288,0.17696355448828804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,12288,10240,0.154567109213935
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,12288,3072,0.036878221564822726
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,12288,3072,0.05739911397298177
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,12288,2560,0.031806220610936485
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,12288,2560,0.04844533403714498
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,12288,8192,0.1285324494043986
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,12288,2048,0.026686223016844854
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,12288,2048,0.04178222351604038
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,12288,7168,0.11528000566694473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,12288,1536,0.02130222154988183
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,12288,1536,0.03495733274353875
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,12288,6144,0.1021066639158461
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,12288,1024,0.015957333975368075
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,12288,1024,0.027902222341961328
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,12288,5120,0.0899733304977417
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,12288,768,0.013409777647919126
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,12288,768,0.024800888366169397
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,12288,4096,0.07800622118843926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,12288,3584,0.07233777973386976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,12288,512,0.010928000013033548
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,12288,512,0.022077333596017625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,12288,256,0.007804444266690149
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,12288,256,0.018756444255510967
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,12288,3072,0.06621244218614367
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,12288,128,0.006308444258239534
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,12288,128,0.017818666166729398
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,12288,2560,0.0602417786916097
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,12288,64,0.005522666705979242
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,12288,64,0.018072000808186002
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,12288,32,0.005864000154866113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,12288,32,0.018090666996108163
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,12288,2048,0.05254400107595655
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,12288,1536,0.0477786660194397
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,12288,1024,0.04168088899718391
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,12288,768,0.03882222374280294
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,10240,16384,0.22926667001512316
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,10240,16384,0.15310843785603842
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,10240,65536,0.9234142303466797
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,10240,12288,0.1733715534210205
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,10240,12288,0.11624356110890706
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,12288,128,0.02897955642806159
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,12288,256,0.031481779283947416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,12288,512,0.037074665228525795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,10240,10240,0.09792088800006443
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,10240,10240,0.14396622445848253
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,10240,8192,0.11990043852064346
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,10240,8192,0.07923555374145508
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,10240,7168,0.10504355695512559
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,10240,7168,0.07046488920847575
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,10240,6144,0.061292443010542125
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,10240,6144,0.09367289145787557
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,10240,65536,0.6058008935716417
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,10240,5120,0.052480889691246875
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,10240,5120,0.0797191129790412
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,10240,16384,0.29853688346015084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,10240,4096,0.0425235562854343
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,10240,4096,0.06594755252202351
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,10240,65536,1.0825555589463975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,10240,12288,0.2274728880988227
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,10240,3584,0.038338667816585965
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,10240,3584,0.06014221906661987
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,10240,10240,0.19255733489990234
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,10240,3072,0.053772442870669894
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,10240,3072,0.03345955411593119
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,10240,2560,0.02910311023394267
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,10240,2560,0.04540266593297323
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,10240,8192,0.1632888846927219
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,10240,2048,0.024282667371961806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,10240,2048,0.03927110963397556
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,10240,7168,0.1431164476606581
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,10240,1536,0.01958755486541324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,10240,1536,0.03284800052642822
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,10240,6144,0.12201510535346137
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,10240,1024,0.01461777753300137
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,10240,1024,0.02671555512481266
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,10240,5120,0.10596088568369548
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,10240,768,0.0124195557501581
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,10240,768,0.02441155579355028
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,10240,4096,0.09182933304044937
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,10240,512,0.009914666414260864
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,10240,512,0.02164266672399309
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,10240,3584,0.0843048890431722
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,10240,256,0.007205333146784041
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,10240,256,0.018775110443433125
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,10240,3072,0.07359555694792005
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,10240,2560,0.061214221848381885
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,10240,128,0.00582311095462905
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,10240,128,0.017645332548353408
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,10240,64,0.005572444448868434
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,10240,64,0.017816000514560275
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,10240,32,0.005787555542257097
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,10240,32,0.017795556121402316
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,10240,1536,0.04857066604826185
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,10240,2048,0.05453600154982673
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,10240,1024,0.04090933336151971
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,10240,768,0.038809776306152344
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,8192,65536,0.736780431535509
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,8192,16384,0.1835155619515313
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,8192,16384,0.11665422386593288
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,8192,12288,0.13786221875084773
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,8192,12288,0.08879822492599487
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,10240,256,0.03168888886769613
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,10240,512,0.03626577721701728
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,8192,65536,0.4478595521714952
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,8192,10240,0.11689421865675186
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,10240,128,0.028705779049131606
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,8192,8192,0.09652622540791829
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,8192,10240,0.07509244150585599
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,8192,8192,0.062041779359181724
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,8192,7168,0.08709600236680771
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,8192,7168,0.05406755540106031
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,8192,6144,0.076837334367964
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,8192,6144,0.04698488778538174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,8192,5120,0.04017333189646403
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,8192,5120,0.0654604434967041
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,8192,16384,0.16915555795033774
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,8192,4096,0.033315555916892156
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,8192,4096,0.055030223396089345
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,8192,65536,0.6120239893595377
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,8192,12288,0.13252800040774876
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,8192,10240,0.11626933680640326
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,8192,3584,0.05074666606055366
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,8192,3584,0.029482665989134047
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,8192,3072,0.026142223013771906
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,8192,3072,0.04552711049715678
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,8192,8192,0.09810577498541938
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,8192,2560,0.038563556141323514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,8192,2560,0.022574222750133936
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,8192,7168,0.08968000279532538
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,8192,2048,0.01899377836121453
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,8192,2048,0.03376355436113145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,8192,1536,0.015392889579137167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,8192,1536,0.028957333829667833
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,8192,6144,0.08063555426067777
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,8192,1024,0.011828444070286222
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,8192,1024,0.02397955622937944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,8192,5120,0.07117511166466607
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,8192,768,0.010054222411579555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,8192,768,0.02148799929353926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,8192,3584,0.059600002235836454
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,8192,512,0.008231999973456064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,8192,512,0.019871999820073444
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,8192,2560,0.04885955651601156
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,8192,4096,0.06498399708006117
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,8192,256,0.005977777971161737
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,8192,256,0.01722755531469981
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,8192,128,0.0052044445441828836
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,8192,128,0.016352888610627916
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,8192,1536,0.04094666573736403
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,8192,64,0.004566222015354368
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,8192,64,0.016303999556435477
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,8192,32,0.004813333352406819
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,8192,32,0.016466667254765827
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,8192,768,0.03457599878311157
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,8192,3072,0.05396355523003472
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,8192,2048,0.04356888930002848
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,7168,16384,0.1026479999224345
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,7168,16384,0.16907555527157256
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,8192,1024,0.03614844547377692
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,7168,65536,0.6953191227383084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,7168,12288,0.0783182250128852
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,7168,12288,0.12921510802374944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,7168,65536,0.40665777524312335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,7168,10240,0.06621155473921034
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,7168,10240,0.10901421970791286
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,7168,8192,0.05436355537838406
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,8192,128,0.02776177724202474
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,7168,8192,0.09126933415730794
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,8192,256,0.03000977635383606
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,7168,7168,0.04795911245875888
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,7168,7168,0.08015111419889662
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,7168,6144,0.04224266608556112
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,8192,512,0.032313777340783015
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,7168,6144,0.07157955567042033
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,7168,5120,0.03621422251065572
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,7168,5120,0.061546663443247475
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,7168,16384,0.20231289333767363
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,7168,65536,0.7748097843594022
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,7168,4096,0.05321955680847168
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,7168,4096,0.030012445317374334
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,7168,3584,0.02693333393997616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,7168,3584,0.04702488912476433
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,7168,12288,0.15240800380706787
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,7168,3072,0.024010666542583044
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,7168,3072,0.04318933354483711
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,7168,10240,0.13380711608462864
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,7168,2560,0.020780444145202637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,7168,2560,0.0370071099864112
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,7168,8192,0.11168711715274388
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,7168,2048,0.01754577789041731
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,7168,2048,0.03239555491341485
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,7168,7168,0.10152710808648004
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,7168,1536,0.014460444450378418
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,7168,1536,0.028019554085201685
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,7168,6144,0.08992089165581597
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,7168,1024,0.011423110961914062
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,7168,1024,0.023544000254737005
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,7168,5120,0.07796621984905668
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,7168,768,0.009941333697901832
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,7168,768,0.021571555071406897
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,7168,3584,0.06416000260247125
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,7168,512,0.008295111358165741
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,7168,512,0.018710222509172227
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,7168,4096,0.06773244672351413
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,7168,256,0.006353777729802662
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,7168,256,0.01697422232892778
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,7168,2560,0.05269599954287211
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,7168,128,0.005488889084921942
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,7168,128,0.016153777639071148
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,7168,3072,0.05749689208136665
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,7168,64,0.004951999833186467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,7168,64,0.01640533407529195
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,7168,32,0.005245333330498801
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,7168,32,0.016304888659053378
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,7168,1536,0.043218665652804904
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,7168,2048,0.046219554212358266
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,7168,768,0.03569333420859443
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,6144,65536,0.6520577536688911
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,6144,16384,0.08819466829299927
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,7168,1024,0.037894222471449114
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,6144,16384,0.163372450404697
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,6144,65536,0.34899200333489316
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,6144,12288,0.12295377254486084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,6144,12288,0.06880622439914279
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,7168,512,0.03297066688537598
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,7168,256,0.03029511041111416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,6144,10240,0.10351022084554036
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,6144,10240,0.05712088611390856
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,6144,8192,0.08672977818383111
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,6144,8192,0.04629955689112345
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,6144,7168,0.07710844278335571
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,7168,128,0.02814488940768772
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,6144,7168,0.041490667396121554
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,6144,6144,0.06865688827302721
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,6144,6144,0.03590133455064561
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,6144,5120,0.03080000148879157
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,6144,5120,0.05874400006400215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,6144,16384,0.1625608868069119
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,6144,4096,0.02560088866286808
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,6144,4096,0.050620443291134305
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,6144,65536,0.5834373368157281
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,6144,12288,0.1263777812321981
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,6144,3584,0.02353511088424259
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,6144,3584,0.04619822237226698
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,6144,10240,0.11042222049501206
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,6144,3072,0.040848887628979154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,6144,3072,0.02029333346419864
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,6144,2560,0.017626666360431247
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,6144,2560,0.03552533189455668
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,6144,8192,0.09352533022562663
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,6144,2048,0.015103111664454142
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,6144,2048,0.031423999203575984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,6144,7168,0.08527911371654934
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,6144,1536,0.012549333274364471
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,6144,1536,0.027220444546805486
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,6144,6144,0.07733955648210314
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,6144,1024,0.00980799976322386
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,6144,1024,0.022913777165942725
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,6144,5120,0.06852089034186469
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,6144,768,0.008557333714432186
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,6144,768,0.020303999384244282
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,6144,4096,0.06059910853703817
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,6144,512,0.006834666762087081
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,6144,512,0.01848088867134518
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,6144,3584,0.05780533287260267
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,6144,256,0.0053280000057485365
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,6144,256,0.01703555550840166
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,6144,2560,0.04786844385994805
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,6144,3072,0.05260533425543043
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,6144,128,0.004608889006906086
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,6144,128,0.016173332929611206
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,6144,64,0.004034666551484002
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,6144,64,0.016121778223249648
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,6144,32,0.004341333276695675
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,6144,32,0.016120889120631747
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,6144,2048,0.04281422164705065
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,6144,1536,0.04015377826160855
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,6144,768,0.03370489014519586
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,6144,1024,0.03580888774659898
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,5120,16384,0.1592257817586263
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,5120,16384,0.07857777674992879
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,5120,65536,0.628419558207194
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,5120,12288,0.12011910809410943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,5120,12288,0.060473779837290444
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,5120,10240,0.051022221644719444
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,5120,65536,0.3134400049845378
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,5120,10240,0.10155022144317627
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,5120,8192,0.041987554894553296
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,5120,8192,0.08481422397825454
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,6144,512,0.031003554662068684
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,6144,256,0.029510223203235205
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,5120,7168,0.037090665764278836
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,6144,128,0.02760444415940179
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,5120,7168,0.07543822129567464
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,5120,6144,0.06779111093944974
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,5120,6144,0.03257066673702664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,5120,5120,0.02801688843303257
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,5120,5120,0.05764355262120565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,5120,16384,0.20015733771853975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,5120,4096,0.023520888553725347
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,5120,4096,0.049413333336512245
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,5120,65536,0.7159982257419162
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,5120,3584,0.021021333005693223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,5120,12288,0.14969333012898764
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,5120,3584,0.04478933413823446
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,5120,3072,0.018602665927675035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,5120,3072,0.04032266802257962
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,5120,10240,0.1284168826209174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,5120,2560,0.016360888878504436
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,5120,2560,0.03487911158137851
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,5120,8192,0.10980800125333999
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,5120,2048,0.013771555489963956
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,5120,2048,0.03084533413251241
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,5120,7168,0.09691022502051459
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,5120,1536,0.011410666836632622
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,5120,1536,0.026523555318514507
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,5120,6144,0.0856871141327752
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,5120,1024,0.00907555553648207
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,5120,5120,0.07453778054979113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,5120,1024,0.02188088827662998
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,5120,768,0.007924444145626491
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,5120,768,0.019575110740131803
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,5120,3584,0.06122577852673001
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,5120,512,0.0063075555695427795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,5120,4096,0.06448444392946032
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,5120,512,0.01808888879087236
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,5120,256,0.004875555634498596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,5120,256,0.016732444365819294
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,5120,3072,0.05415199862586128
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,5120,128,0.004269333349333869
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,5120,128,0.015985776980717976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,5120,2560,0.04822399881150988
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,5120,64,0.0041680000722408295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,5120,64,0.01590666671593984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,5120,32,0.004304000073009067
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,5120,32,0.015857777661747403
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,5120,2048,0.042768001556396484
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,5120,1536,0.03997155692842271
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,4096,65536,0.23570754792955187
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,5120,1024,0.034824887911478676
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,4096,65536,0.4801777733696832
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,5120,768,0.03239466746648153
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,4096,16384,0.0947635571161906
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,4096,16384,0.12137866020202637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,4096,12288,0.09083822038438584
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,4096,12288,0.06686577532026503
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,4096,10240,0.07891911268234253
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,4096,10240,0.052037331793043345
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,4096,8192,0.04761866728464762
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,5120,256,0.029678222205903795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,4096,8192,0.06564978096220228
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,5120,512,0.03139288889037238
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,4096,7168,0.060284442371792264
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,5120,128,0.0273644444015291
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,4096,7168,0.03788355655140347
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,4096,6144,0.05382755398750305
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,4096,6144,0.03029777937465244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,4096,5120,0.028254220883051556
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,4096,5120,0.046816001335779824
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,4096,65536,0.514460457695855
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,4096,16384,0.14512266053093806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,4096,4096,0.040222220950656466
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,4096,4096,0.02310933338271247
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,4096,12288,0.11430399947696263
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,4096,3584,0.01957066688272688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,4096,3584,0.03704977697796292
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,4096,3072,0.017024889588356018
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,4096,3072,0.033498668008380465
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,4096,10240,0.1004142231411404
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,4096,2560,0.013051555388503604
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,4096,2560,0.029520889123280842
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,4096,8192,0.0863271090719435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,4096,2048,0.011193777951929303
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,4096,2048,0.026429333620601233
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,4096,7168,0.0789804458618164
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,4096,1536,0.009511999785900116
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,4096,1536,0.02367288867632548
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,4096,6144,0.07130755318535699
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,4096,1024,0.007756444315115611
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,4096,1024,0.01937866707642873
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,4096,5120,0.06359022193484835
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,4096,768,0.006585777633719974
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,4096,768,0.0178195552693473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,4096,3584,0.052925331725014575
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,4096,512,0.005398222141795688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,4096,512,0.01661688917213016
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,4096,768,0.030784000953038532
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,4096,256,0.004462222258249919
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,4096,256,0.015615110596021017
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,4096,2560,0.04494400156868828
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,4096,128,0.003951999876234266
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,4096,128,0.015019555886586508
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,4096,4096,0.05468888746367561
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,4096,64,0.003528000166018804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,4096,64,0.014689778288205465
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,4096,1536,0.03809422254562378
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,4096,32,0.0037706668178240457
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,4096,32,0.014595554934607612
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,4096,3072,0.04898311032189263
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,4096,2048,0.03974489039844937
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3584,65536,0.46118842230902773
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3584,16384,0.0631635586420695
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,4096,1024,0.033025778002209134
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3584,16384,0.11537688308291966
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3584,12288,0.08810577789942424
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3584,12288,0.05165422293874952
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3584,10240,0.03982222080230713
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3584,65536,0.2136497762468126
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3584,10240,0.07263911432690091
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3584,8192,0.03256355391608344
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3584,8192,0.06294844547907512
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,4096,128,0.026557332939571802
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3584,7168,0.028943111499150593
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3584,7168,0.057151112291547984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,4096,256,0.028597331709331934
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3584,6144,0.02567022211021847
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3584,6144,0.05263911022080315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,4096,512,0.029686222473780315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3584,5120,0.02199199961291419
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3584,5120,0.04500444398985969
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,3584,16384,0.14281511306762695
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,3584,65536,0.5046079953511556
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3584,4096,0.018234666850831773
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,3584,12288,0.1135022242863973
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3584,4096,0.03866399990187751
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3584,3584,0.01648533344268799
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3584,3584,0.035804443889194064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3584,3072,0.014698666003015308
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3584,3072,0.03264000018437704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,3584,10240,0.0991555584801568
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3584,2560,0.014542222023010254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3584,2560,0.02866222129927741
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,3584,8192,0.08543999989827473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3584,2048,0.012888000243239932
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3584,2048,0.025619554850790236
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,3584,7168,0.07828355497784086
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3584,1536,0.010479110810491772
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,3584,6144,0.07057955529954699
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3584,1536,0.02235466738541921
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3584,1024,0.008164444731341468
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3584,1024,0.019443555010689628
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,3584,5120,0.06261066595713298
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3584,768,0.006970666348934174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3584,768,0.01771022213829888
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,3584,3584,0.05240177777078417
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3584,512,0.005688000056478713
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3584,512,0.016672000288963318
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,3584,4096,0.05430488785107931
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,3584,2560,0.044658667511410184
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3584,256,0.004493333399295807
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3584,256,0.015434667468070984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3584,128,0.004057777838574516
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3584,128,0.01463644372092353
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,3584,3072,0.04864889052179125
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3584,64,0.003556444413132138
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3584,64,0.014647111296653748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3584,32,0.003605333467324575
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3584,32,0.014615999327765571
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,3584,2048,0.03992266787423028
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3072,65536,0.43015556865268284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3072,65536,0.18014132976531982
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3072,16384,0.056961774826049805
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,3584,1536,0.037392000357309975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3072,16384,0.10831377903620402
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3072,12288,0.04526666800181071
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,3584,768,0.030481778913074072
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3072,12288,0.08292710781097412
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3072,10240,0.033632887734307185
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,3584,1024,0.03173244330618117
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3072,10240,0.07160533136791654
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3072,8192,0.027886221806208294
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3072,8192,0.060621334446801074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,3584,128,0.026455110973782007
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3072,7168,0.02504088812404209
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3072,7168,0.054465777344173856
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,3584,256,0.02823466724819607
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3072,6144,0.021943999661339655
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3072,6144,0.04862933357556661
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,3584,512,0.029800001117918227
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3072,5120,0.01908177799648709
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3072,5120,0.04195288817087809
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,3072,16384,0.13852088981204563
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3072,4096,0.016379555066426594
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3072,4096,0.03641422258483039
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,3072,65536,0.4909937646653917
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3072,3584,0.01463822192615933
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3072,3584,0.03366400135887994
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,3072,12288,0.10968444744745891
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3072,3072,0.013325333595275879
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3072,3072,0.031347556246651545
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,3072,10240,0.0956382221645779
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3072,2560,0.013449778159459433
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3072,2560,0.027521777484152053
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,3072,8192,0.08215555879804823
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3072,2048,0.01148088855875863
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3072,2048,0.025191111697090998
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,3072,7168,0.0758426653014289
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3072,1536,0.009363555245929295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3072,1536,0.021707556313938562
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,3072,6144,0.06877244181103177
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3072,1024,0.0074275558193524676
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3072,1024,0.01870666609870063
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,3072,5120,0.06106488572226631
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3072,768,0.006172444257471297
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3072,768,0.01776444415251414
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,3072,4096,0.053504887554380626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3072,512,0.005202666752868229
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3072,512,0.016591111818949383
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,3072,3584,0.05190133386188083
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3072,256,0.0042133331298828125
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3072,256,0.015344889627562629
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,3072,3072,0.04797510968314277
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3072,128,0.0037377778854635027
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3072,128,0.014672888649834527
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,3072,2560,0.04432533184687296
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3072,64,0.00350222239891688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3072,64,0.014488889111412896
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,3072,32,0.003571555432346132
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,3072,2048,0.039557334449556135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,3072,32,0.01440711153878106
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,3072,1536,0.03680533170700073
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2560,65536,0.16247822178734675
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2560,65536,0.39434488614400226
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2560,16384,0.04940088921123081
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,3072,768,0.032060444355010986
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,3072,1024,0.03198933270242479
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2560,12288,0.07906933625539143
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2560,16384,0.10208266973495483
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2560,12288,0.036066667901145086
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2560,10240,0.06831822130415174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2560,10240,0.030785779158274334
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2560,8192,0.025765332910749648
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2560,8192,0.05783110857009888
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,3072,256,0.02814133299721612
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,3072,512,0.029543111721674602
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2560,7168,0.05214577913284302
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2560,7168,0.022809776994917128
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2560,6144,0.01995733380317688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2560,6144,0.047276443905300565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,3072,128,0.026294221480687458
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2560,5120,0.01716088917520311
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2560,5120,0.04066577884886
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,2560,16384,0.12804355886247423
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2560,4096,0.01420177850458357
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,2560,12288,0.10179199775060017
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2560,4096,0.035114665826161705
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,2560,65536,0.4546311166551378
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2560,3584,0.012651555240154266
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2560,3584,0.032989333073298134
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,2560,10240,0.08962311347325642
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2560,3072,0.011311111350854238
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2560,3072,0.030571556753582422
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2560,2560,0.01013422260681788
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2560,2560,0.026983999543719824
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,2560,8192,0.0771920018725925
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2560,2048,0.00886222223440806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2560,2048,0.024442666106753882
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,2560,7168,0.07148977782991198
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2560,1536,0.007655111451943715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2560,1536,0.02132533325089349
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,2560,6144,0.06476444668240018
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2560,1024,0.0063635553750726916
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2560,1024,0.018594667315483093
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,2560,5120,0.05740977658165825
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2560,768,0.005574222240183089
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2560,768,0.017626666360431247
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,2560,3584,0.04948000113169352
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2560,512,0.004648888690604104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,2560,4096,0.05033422178692288
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2560,512,0.016413332687483895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2560,256,0.004003555410438114
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2560,256,0.015186667442321777
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,2560,3072,0.04597155584229363
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2560,128,0.003744889050722122
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,2560,2560,0.042388445801205106
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2560,128,0.014493332968817817
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2560,64,0.003400000019205941
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2560,64,0.014420444766680399
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2560,32,0.0034888887570963967
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2560,32,0.014191110928853353
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,2560,2048,0.03792533278465271
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2048,65536,0.37970132297939724
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2048,65536,0.12522933218214247
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,2560,1536,0.03544355432192484
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,2560,768,0.029620445436901514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2048,16384,0.046012444628609546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2048,16384,0.09986844327714707
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,2560,1024,0.030233777231640283
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2048,12288,0.0777955585055881
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2048,12288,0.03533333208825853
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2048,10240,0.0675519969728258
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2048,10240,0.028323554330401953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2048,8192,0.02940888868437873
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2048,8192,0.0572942230436537
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,2560,256,0.028071999549865723
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,2560,128,0.026001777913835313
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2048,7168,0.0251857770813836
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2048,7168,0.05068000157674154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2048,6144,0.022081777453422546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,2560,512,0.02884799904293484
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2048,6144,0.0472302238146464
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2048,5120,0.018255111243989732
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2048,5120,0.04015377826160855
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,2048,12288,0.10386488835016887
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2048,4096,0.01627377834584978
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,2048,65536,0.45472086800469297
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,2048,16384,0.13103555308447942
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2048,4096,0.034767998589409724
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2048,3584,0.01386577801571952
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2048,3584,0.032433778047561646
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2048,3072,0.012452444268597497
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2048,3072,0.03029511041111416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,2048,10240,0.09090755383173625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2048,2560,0.008996444443861643
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2048,2560,0.026929777529504564
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,2048,8192,0.0781431131892734
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2048,2048,0.007710222568776872
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2048,2048,0.023306666149033442
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,2048,7168,0.0716080003314548
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,2048,6144,0.06450044446521334
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2048,1536,0.006671999891599019
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2048,1536,0.021049777666727703
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,2048,5120,0.05708711014853584
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2048,1024,0.005333333379692501
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2048,1024,0.01854044364558326
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2048,768,0.00489688871635331
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2048,768,0.017415111263593037
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,2048,4096,0.050082667006386646
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,2048,3584,0.04872266782654656
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2048,512,0.004376000000370873
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2048,512,0.016383111476898193
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2048,256,0.0037946667936113146
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2048,256,0.01512977812025282
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,2048,2560,0.04127111037572225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2048,128,0.0035048888789282907
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2048,128,0.014458666245142618
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,2048,3072,0.044532444741990834
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2048,64,0.0032666667054096856
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2048,64,0.01419911119672987
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,2048,32,0.0034106667670938703
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,2048,32,0.013967111706733704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1536,65536,0.09728088643815781
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,2048,2048,0.036646223730511136
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1536,65536,0.3766462273067898
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,2048,768,0.02902844548225403
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1536,16384,0.09902666674719916
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1536,16384,0.03394666645261977
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1536,12288,0.027961777316199407
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1536,12288,0.07769155502319336
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,2048,1536,0.03314666615592109
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1536,10240,0.024098666177855596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1536,10240,0.06705155637529162
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,2048,1024,0.02940888868437873
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1536,8192,0.01965777741538154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1536,8192,0.055903997686174184
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,2048,128,0.02576711111598545
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1536,7168,0.0173973325226042
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1536,7168,0.0514008899529775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,2048,256,0.027655111418830022
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1536,6144,0.015448888142903646
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1536,6144,0.04594488938649496
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,2048,512,0.028408000866572063
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1536,5120,0.013421333498424955
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1536,5120,0.03934933410750495
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,1536,16384,0.12270666493309869
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1536,4096,0.011546666423479715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1536,4096,0.03457955519358317
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,1536,12288,0.09773866335550944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,1536,65536,0.43020354376898867
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1536,3584,0.010556444525718689
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1536,3584,0.03236622280544705
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1536,3072,0.009535999761687385
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1536,3072,0.029263999727037217
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,1536,10240,0.08574044704437256
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1536,2560,0.009370666411187913
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1536,2560,0.02556444373395708
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,1536,8192,0.07433333661821154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1536,2048,0.0081377774477005
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1536,2048,0.0229413327243593
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,1536,7168,0.06844355662663777
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1536,1536,0.006463999963468975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1536,1536,0.020780444145202637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,1536,6144,0.06181955337524414
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1536,1024,0.0053013331360287136
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1536,1024,0.018277333842383493
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,1536,5120,0.05486844314469231
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1536,768,0.004846222284767362
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1536,768,0.017310221989949543
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,1536,4096,0.04842933350139194
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1536,512,0.0042408886882993905
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1536,512,0.016233777006467182
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,1536,3584,0.04770133230421278
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1536,256,0.003628444340493944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1536,256,0.01484266585773892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,1536,3072,0.04418489005830553
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1536,128,0.0033173333439562055
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1536,128,0.014378666877746582
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,1536,2560,0.0388337771097819
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1536,64,0.003149333306484752
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1536,64,0.014274666706720987
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1536,32,0.003192000091075897
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1536,32,0.013947554760509066
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,1536,2048,0.03442222211096022
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1024,65536,0.06799733638763428
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1024,65536,0.37634934319390195
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1024,16384,0.027085334062576294
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1024,16384,0.09954400195015801
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,1536,1536,0.03325600094265408
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1024,12288,0.024912888805071514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1024,12288,0.07769155502319336
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,1536,1024,0.02975911233160231
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1024,10240,0.02186311119132572
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1024,10240,0.06515022118886311
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,1536,768,0.028967999749713477
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1024,8192,0.016358221570650738
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1024,8192,0.05566933419969347
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,1536,512,0.02864355511135525
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1024,7168,0.014919999572965833
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,1536,256,0.027241778042581346
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1024,7168,0.05077066686418322
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,1536,128,0.025640888346566096
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1024,6144,0.013213333156373767
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1024,6144,0.04546311166551378
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1024,5120,0.011752888560295105
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1024,5120,0.038945777548684016
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,1024,16384,0.11507288614908855
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1024,4096,0.010168000227875179
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,1024,65536,0.40584799978468156
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1024,4096,0.034512887398401894
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1024,3584,0.009398221969604492
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1024,3584,0.03202399942609999
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,1024,12288,0.09208622243669297
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1024,3072,0.008570666942331526
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1024,3072,0.02879200047916836
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,1024,10240,0.08107911215888129
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1024,2560,0.007808888951937358
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1024,2560,0.025116443634033203
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,1024,8192,0.0699217783080207
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1024,2048,0.006487111250559489
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1024,2048,0.022737777895397607
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,1024,7168,0.06455466482374403
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1024,1536,0.005942222144868638
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1024,1536,0.020596444606781006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,1024,6144,0.05871822436650594
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1024,1024,0.0047324444684717385
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1024,1024,0.018033777674039204
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,1024,5120,0.052520000272327
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1024,768,0.00434666665063964
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1024,768,0.017066667477289837
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,1024,4096,0.04658933480580648
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1024,512,0.0038657776183552216
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1024,512,0.015795555379655626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,1024,3584,0.04557244314087761
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1024,256,0.003428444473279847
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1024,256,0.014712888333532544
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,1024,3072,0.041466666592492
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1024,128,0.003185777821474605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1024,128,0.014172444740931192
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,1024,2560,0.03855822152561612
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1024,64,0.003007111036115222
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1024,64,0.014089778065681458
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,1024,32,0.0030497778207063675
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,1024,32,0.013829333086808523
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,1024,2048,0.034529778692457415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,768,65536,0.055756443076663546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,768,65536,0.3748035430908203
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,768,16384,0.020941333638297185
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,768,16384,0.09941333532333374
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,1024,1536,0.03227555420663621
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,768,12288,0.019582221905390423
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,768,12288,0.07751911216311984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,1024,1024,0.02888710962401496
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,768,10240,0.017073776986863878
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,768,10240,0.06648711363474528
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,1024,768,0.028186665640936956
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,768,8192,0.015189333094490899
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,768,8192,0.05568355321884155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,1024,512,0.028216888507207234
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,1024,256,0.027581334114074707
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,768,7168,0.014206222361988492
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,768,7168,0.050147556596332125
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,768,6144,0.013156444662147097
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,768,6144,0.044850667317708336
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,1024,128,0.025489777326583862
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,768,5120,0.012225777738624148
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,768,5120,0.03908355699645148
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,768,16384,0.11415555742051865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,768,65536,0.4014791117774115
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,768,4096,0.00998577806684706
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,768,4096,0.03309866786003113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,768,3584,0.0092622223827574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,768,3584,0.030839999516805012
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,768,12288,0.09186844693289863
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,768,3072,0.008284444610277811
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,768,3072,0.02804533309406704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,768,10240,0.08100444740719266
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,768,2560,0.00701155596309238
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,768,2560,0.024977778395016987
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,768,8192,0.06961244344711304
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,768,2048,0.006093333164850871
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,768,2048,0.022591999835438196
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,768,7168,0.06432266367806329
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,768,1536,0.005470222069157495
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,768,6144,0.058288892110188804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,768,1536,0.020385776956876118
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,768,1024,0.004582222137186262
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,768,1024,0.01796711153454251
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,768,5120,0.05228711167971293
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,768,768,0.004281777888536453
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,768,768,0.01688088807794783
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,768,4096,0.04624799887339274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,768,512,0.0038177776667806837
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,768,512,0.015758223003811307
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,768,3584,0.04567377765973409
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,768,256,0.0033848887930313745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,768,256,0.014697778556081982
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,768,3072,0.04091555542416043
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,768,128,0.003128000017669466
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,768,2560,0.03784088955985175
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,768,128,0.014031999640994601
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,768,2048,0.03329066766632928
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,768,64,0.003018666679660479
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,768,64,0.013948443863126965
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,768,32,0.0030213333666324615
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,768,32,0.013754666679435305
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,512,65536,0.042098667886522084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,512,65536,0.376160012351142
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,512,16384,0.015944888194402058
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,768,1536,0.032030221488740705
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,768,1024,0.02861777941385905
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,512,16384,0.09949422544903225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,512,12288,0.015690666106012132
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,512,12288,0.07699555820888944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,512,10240,0.014382221632533602
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,512,10240,0.06611377663082547
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,768,768,0.02832533253563775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,512,8192,0.01275644451379776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,512,8192,0.05522844526502821
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,768,512,0.027994665834638808
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,512,7168,0.011971555650234222
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,512,7168,0.04956533180342781
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,768,256,0.02718844513098399
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,512,6144,0.011192888849311404
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,512,6144,0.044772442844178945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,768,128,0.02535999980237749
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,512,5120,0.01034577770365609
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,512,5120,0.03896444373660617
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,512,65536,0.39262933201260036
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,512,16384,0.11230044894748265
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,512,4096,0.009353777600659264
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,512,4096,0.033360888560613
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,512,3584,0.008057778080304464
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,512,12288,0.08950310945510864
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,512,3584,0.030035555362701416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,512,3072,0.00721688899728987
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,512,3072,0.027907556957668726
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,512,10240,0.07942933506435819
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,512,2560,0.006562666760550604
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,512,2560,0.02475555572244856
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,512,8192,0.0683751106262207
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,512,2048,0.0058266665372583605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,512,2048,0.022783999641736347
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,512,7168,0.06270222531424628
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,512,1536,0.005094222310516569
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,512,1536,0.02038755516211192
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,512,6144,0.057597332530551486
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,512,1024,0.004376888689067629
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,512,1024,0.01807733376820882
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,512,5120,0.0513991117477417
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,512,768,0.004017777740955353
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,512,768,0.01684888866212633
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,512,3584,0.044079999128977455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,512,512,0.003644444462325838
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,512,4096,0.04581244455443489
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,512,512,0.015556444724400839
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,512,256,0.003272888975010978
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,512,3072,0.041160000695122614
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,512,256,0.014630221658282809
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,512,128,0.002991111121243901
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,512,128,0.0140346669488483
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,512,2560,0.03643644518322415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,512,64,0.0028613333900769553
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,512,64,0.013610666824711693
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,512,32,0.0028657778683635923
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,512,32,0.013554666605260638
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,256,65536,0.038731555143992104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,256,65536,0.37651644812689883
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,512,2048,0.03280622098180983
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,256,16384,0.0123857781291008
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,256,16384,0.09888799985249837
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,512,1024,0.02829688787460327
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,512,1536,0.031769778993394636
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,256,12288,0.010809777511490716
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,256,12288,0.07717244492636786
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,256,10240,0.010101333260536194
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,256,10240,0.06584355566236708
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,512,768,0.02808800008561876
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,256,8192,0.008938666847017076
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,256,8192,0.054930667082468666
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,512,512,0.027768888407283362
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,256,7168,0.008347555167145198
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,256,7168,0.04946133494377136
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,512,256,0.027172444595230952
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,256,6144,0.008015111088752747
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,512,128,0.02556177808178796
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,256,6144,0.04397066765361362
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,256,5120,0.008804444637563493
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,256,5120,0.037568890386157565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,256,65536,0.3877386781904433
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,256,4096,0.00792266676823298
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,256,4096,0.031884445084465876
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,256,16384,0.11075377464294434
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,256,3584,0.007638221813572778
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,256,3584,0.0299271113342709
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,256,12288,0.0886675582991706
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,256,3072,0.007237333390447829
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,256,3072,0.027793778313530818
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,256,10240,0.0783075557814704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,256,8192,0.06748355759514703
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,256,2560,0.0063591111037466265
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,256,2560,0.02494222256872389
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,256,2048,0.005648889061477449
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,256,2048,0.0224924451775021
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,256,7168,0.061999109056260854
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,256,1536,0.004967111266321606
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,256,1536,0.020192000601026747
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,256,6144,0.05666044685575697
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,256,1024,0.004289777742491828
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,256,5120,0.04898933238453335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,256,1024,0.017835555805100333
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,256,768,0.0038737778862317405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,256,768,0.01680444512102339
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,256,4096,0.043040888177024
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,256,512,0.0034862222770849862
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,256,512,0.015292444162898593
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,256,3584,0.041929778125551015
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,256,256,0.0031582222630580268
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,256,256,0.014517333772447375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,256,3072,0.03935022155443827
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,256,128,0.0029013332807355454
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,256,128,0.013926222920417786
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,256,2560,0.03612355391184489
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,256,64,0.0027502222607533136
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,256,64,0.013771555489963956
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,256,32,0.0027840000887711844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,256,32,0.013564444250530668
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,128,65536,0.030456887351142034
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,256,2048,0.032418668270111084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,128,65536,0.37671555413140195
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,128,16384,0.009143999881214565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,128,16384,0.09934133291244507
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,256,1536,0.031707555055618286
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,128,12288,0.009484444227483537
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,128,12288,0.07530044184790717
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,256,768,0.02779910961786906
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,256,1024,0.02815022071202596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,128,10240,0.008488000267081791
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,128,10240,0.06529688835144043
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,128,8192,0.00755466686354743
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,128,8192,0.05442133214738634
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,256,512,0.027859555350409612
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,128,7168,0.007282666862010956
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,128,7168,0.04919111066394382
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,256,256,0.027114666170544092
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,128,6144,0.0069520001610120135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,128,6144,0.04393066631423103
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,256,128,0.025154666768179998
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,128,5120,0.00723555518521203
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,128,5120,0.03674755493799845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,128,65536,0.38688000043233234
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,128,4096,0.006858666737874349
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,128,4096,0.031680001152886286
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,128,16384,0.11033422417110866
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,128,3584,0.007229333122571309
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,128,3584,0.02947288751602173
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,128,12288,0.08854044146007961
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,128,3072,0.0068408888247278
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,128,3072,0.027236445082558527
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,128,10240,0.07766399780909221
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,128,2560,0.006238222122192383
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,128,2560,0.02446311049991184
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,128,8192,0.06698044141133626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,128,2048,0.005570666657553778
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,128,2048,0.02219555609756046
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,128,7168,0.06168444289101494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,128,1536,0.004882666799757215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,128,1536,0.019959110352728102
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,128,6144,0.05582844548755222
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,128,1024,0.0042035554846127825
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,128,1024,0.01756266587310367
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,128,5120,0.04864977796872457
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,128,768,0.003810666501522064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,128,768,0.01628977722591824
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,128,4096,0.04245333207978142
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,128,512,0.0034799998005231223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,128,3584,0.041548444165123835
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,128,512,0.015256888336605497
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,128,256,0.003117333269781537
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,128,256,0.014318222800890604
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,128,3072,0.03865866528617011
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,128,128,0.002862222285734283
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,128,128,0.013720889058378009
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,128,2560,0.03580800029966567
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,128,64,0.0027182222240500977
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,128,2048,0.03207733233769735
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,128,32,0.0028239999794297745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,128,64,0.01368799971209632
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,128,32,0.013415111435784234
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,64,65536,0.024043555061022442
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,64,16384,0.007691555553012424
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,64,65536,0.37668800354003906
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,64,16384,0.09938755300309922
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,64,12288,0.007027555671003129
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,64,10240,0.007194666398896112
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,64,12288,0.0762702226638794
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,64,8192,0.00701333334048589
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,64,10240,0.06572888957129584
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,64,7168,0.006881777611043718
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,64,8192,0.054105778535207115
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,64,7168,0.04752444558673435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,64,6144,0.006671111202902264
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,64,6144,0.042393777105543345
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,64,5120,0.007040888898902469
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,64,5120,0.036518222755855985
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,128,1536,0.03142133355140686
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,64,4096,0.00664533343580034
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,64,3584,0.007062222394678328
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,64,4096,0.03187200095918443
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,64,3584,0.029398222764333088
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,64,3072,0.006882666713661618
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,64,2560,0.006196444647179709
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,64,3072,0.027264000640975103
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,64,2048,0.005440000030729506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,64,2560,0.024655110306209985
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,64,1536,0.004824000100294749
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,64,2048,0.022094221578703985
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,64,1024,0.0040737779604064096
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,64,1536,0.020033778415785897
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,64,768,0.003798222376240624
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,64,1024,0.017482666505707633
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,64,512,0.0033733333564466904
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,64,768,0.01643288963370853
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,64,512,0.01515111161602868
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,64,256,0.003108444520168834
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,64,256,0.01423822177780999
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,64,128,0.0029031110720502008
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,64,64,0.0027377777215507296
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,64,128,0.013707555830478668
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,64,32,0.0026897777699761917
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,64,64,0.01347555551264021
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,64,32,0.013352889153692456
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,32,65536,0.022788445154825848
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,32,16384,0.00850755555762185
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,32,65536,0.37592622968885636
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,32,12288,0.007568000091446771
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,128,1024,0.02811644474665324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,32,16384,0.09916444619496663
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,32,10240,0.007318221860461765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,32,12288,0.076273779074351
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,32,8192,0.007056000332037608
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,32,10240,0.06576089064280193
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,32,7168,0.006815111057625876
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,32,8192,0.05399466554323832
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,32,6144,0.006628444625271692
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,32,7168,0.04903822143872579
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,32,5120,0.0070053330726093715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,32,6144,0.043474667602115206
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,32,4096,0.006720888945791457
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,32,5120,0.036551998721228704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,32,3584,0.0069395556218094295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,32,4096,0.031769778993394636
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,32,3072,0.006692444284756978
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,32,3584,0.029496000872717962
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,32,2560,0.006147555592987273
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,32,3072,0.027228444814682007
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,32,2048,0.005428444594144821
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,32,2560,0.02482755482196808
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,32,1536,0.0047377778424157035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,32,2048,0.022226666410764057
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,32,1024,0.0041351111398802865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,32,1536,0.01995466649532318
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,32,768,0.0037617778612507712
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,32,1024,0.017503110898865592
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,32,768,0.016230222251680162
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,32,512,0.003370666669474708
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,32,256,0.003036444385846456
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,32,512,0.015242666006088257
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,32,128,0.0028275555620590844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,32,256,0.014402666025691561
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,32,128,0.013802666631009845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,32,64,0.0026808888134029177
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,384,32,32,0.0026764445420768526
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,32,64,0.013433777623706393
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,384,32,32,0.013383111192120446
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,128,768,0.027456000447273254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,128,512,0.027801778581407335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,128,256,0.02682488825586107
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,65536,12288,0.29480089081658256
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,65536,16384,0.3905804422166612
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,65536,10240,0.25067112180921763
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,65536,12288,0.47239287694295246
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,65536,10240,0.3954728974236383
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,65536,8192,0.20966755019293892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,65536,8192,0.3172230985429552
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,65536,16384,0.6411102082994249
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,384,128,128,0.02512888941499922
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,65536,7168,0.18627377351125082
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,65536,7168,0.2828231122758654
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,65536,16384,0.48839378356933594
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,65536,12288,0.3746471140119765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,65536,6144,0.15973867310418025
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,65536,5120,0.13723466131422254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,65536,6144,0.24470665719774035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,65536,10240,0.31969955232408315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,65536,5120,0.20805866188473174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,65536,4096,0.11187733544243707
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,65536,4096,0.169542219903734
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,65536,3584,0.1007617778248257
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,65536,3584,0.1506791114807129
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,65536,3072,0.08862844440672134
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,65536,8192,0.2644515567355686
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,65536,3072,0.1323582198884752
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,65536,2560,0.08181866672303942
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,65536,2560,0.11240622732374404
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,65536,7168,0.2341759999593099
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,65536,2048,0.06647644440333049
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,65536,6144,0.2073111136754354
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,65536,2048,0.09240888886981541
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,65536,1536,0.07713155614005195
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,65536,1536,0.05873777468999227
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,65536,5120,0.17772444089253744
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,65536,1024,0.04504444532924228
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,65536,1024,0.05231022172504001
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,65536,768,0.04311555624008179
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,65536,768,0.03943377733230591
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,65536,4096,0.1508257786432902
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,65536,3584,0.13491822613610163
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,65536,512,0.03312089045842489
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,65536,512,0.03456266721089681
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,65536,256,0.023381332556406658
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,65536,256,0.02883022361331516
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,65536,2560,0.10517688592274983
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,65536,128,0.018048889107174344
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,65536,128,0.027355555031034682
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,65536,3072,0.1192622184753418
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,65536,64,0.019166222876972623
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,65536,64,0.028572443458769057
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,65536,32,0.01947111056910621
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,65536,32,0.029863999949561224
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,65536,1536,0.0764462219344245
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,65536,256,0.04450755649142795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,65536,2048,0.09054666757583618
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,65536,512,0.049851556619008384
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,65536,768,0.05750044186909994
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,16384,16384,0.13158933321634927
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,65536,1024,0.06273421976301405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,65536,128,0.037871999873055354
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,16384,16384,0.1581688854429457
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,16384,12288,0.1041991114616394
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,16384,12288,0.12129333284166123
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,16384,10240,0.09090310997433132
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,16384,65536,0.5115511152479384
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,16384,10240,0.10178755389319526
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,16384,8192,0.08281155427296956
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,16384,8192,0.07865155405468412
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,16384,7168,0.0681662228372362
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,16384,7168,0.07345777750015259
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,16384,6144,0.06395288970735338
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,16384,65536,0.6256915728251139
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,16384,6144,0.06090488698747423
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,16384,5120,0.05497511227925619
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,16384,5120,0.054457777076297335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,16384,16384,0.1781840059492323
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,16384,4096,0.045639998382992215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,16384,12288,0.1405351161956787
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,16384,4096,0.04450311263402303
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,16384,10240,0.12310843997531468
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,16384,3584,0.040475553936428495
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,16384,3584,0.04105688797103034
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,16384,3072,0.03593600127432082
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,16384,3072,0.037500444385740496
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,16384,65536,0.6323039796617296
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,16384,8192,0.1019937793413798
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,16384,2560,0.03477688961558872
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,16384,2560,0.03151466780238681
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,16384,7168,0.09209333525763617
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,16384,2048,0.030091553926467896
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,16384,2048,0.02585244509908888
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,16384,1536,0.020963556236690942
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,16384,1536,0.02608000073168013
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,16384,6144,0.08647288878758748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,16384,1024,0.015247111519177755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,16384,1024,0.022525333695941504
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,16384,5120,0.07847733630074395
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,16384,768,0.012627555264366997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,16384,768,0.020682666036817763
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,16384,3584,0.06066489219665527
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,16384,4096,0.06474222077263726
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,16384,512,0.010245333115259806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,16384,512,0.019706666469573975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,16384,256,0.007975999679830339
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,16384,256,0.016732444365819294
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,16384,2560,0.05080533358785841
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,16384,128,0.005850666513045629
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,16384,128,0.015843555331230164
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,16384,3072,0.055324445168177284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,16384,64,0.00510044437315729
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,16384,64,0.0162444445821974
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,16384,1536,0.04246222310596042
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,16384,32,0.005558222118351195
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,16384,32,0.01663733356528812
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,16384,2048,0.04676444331804911
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,16384,768,0.037230223417282104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,16384,1024,0.03804355528619554
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,16384,256,0.031249778138266668
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,12288,65536,0.4766844643486871
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,12288,16384,0.12776977486080593
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,12288,16384,0.11841066678365071
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,16384,512,0.03399111164940728
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,12288,12288,0.10125511222415501
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,16384,128,0.02808800008561876
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,12288,12288,0.09144800239139134
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,12288,10240,0.08760088682174683
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,12288,10240,0.07757511403825548
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,12288,8192,0.0724053316646152
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,12288,8192,0.062472886509365506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,12288,7168,0.06516355276107788
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,12288,7168,0.05596444341871473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,12288,65536,0.46361690097384983
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,12288,6144,0.061315556367238365
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,12288,6144,0.04884799983766344
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,12288,5120,0.04240444302558899
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,12288,5120,0.05227555500136482
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,12288,16384,0.1510631110933092
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,12288,12288,0.120706664191352
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,12288,4096,0.04273955689536201
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,12288,4096,0.03465244505140517
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,12288,10240,0.10719911257425944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,12288,3584,0.031038221385743883
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,12288,3584,0.0388080014122857
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,12288,65536,0.5351422097947862
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,12288,8192,0.08904000123341878
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,12288,3072,0.03654222355948554
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,12288,3072,0.027501333090994094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,12288,2560,0.0320720010333591
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,12288,2560,0.024085332949956257
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,12288,2048,0.02017422186003791
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,12288,2048,0.02880000074704488
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,12288,7168,0.08089066876305474
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,12288,1536,0.016034666034910414
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,12288,1536,0.025104888611369666
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,12288,6144,0.07527466615041097
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,12288,1024,0.012305777933862476
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,12288,1024,0.0221733334991667
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,12288,5120,0.06802488697899713
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,12288,768,0.01055466632048289
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,12288,768,0.020968000094095867
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,12288,3584,0.05429066552056206
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,12288,512,0.008642666869693333
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,12288,512,0.01908266709910499
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,12288,4096,0.05784355269538032
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,12288,256,0.00628355559375551
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,12288,256,0.016789333687888253
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,12288,2560,0.04605688982539707
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,12288,128,0.004931555440028508
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,12288,128,0.01585155559910668
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,12288,3072,0.049997333023283214
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,12288,64,0.004430222428507275
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,12288,64,0.016015110744370353
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,12288,32,0.004790222065316306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,12288,32,0.016306666864289176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,12288,768,0.03525244527392917
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,12288,1536,0.03884977764553494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,12288,2048,0.042528887589772545
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,10240,16384,0.10434666607115005
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,12288,1024,0.03578222129080031
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,10240,16384,0.12735377417670354
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,10240,65536,0.4891306559244792
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,10240,12288,0.08012355698479547
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,10240,12288,0.10034933355119492
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,12288,128,0.02637333340115017
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,12288,256,0.02905422117975023
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,10240,10240,0.08635199732250637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,10240,10240,0.06754577822155423
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,10240,65536,0.4257688787248399
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,10240,8192,0.07166933351092868
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,10240,8192,0.05535911189185249
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,10240,7168,0.06474222077263726
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,12288,512,0.032241778241263494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,10240,7168,0.049442665444480054
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,10240,6144,0.043375998735427856
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,10240,6144,0.05870044231414795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,10240,5120,0.03733155462476943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,10240,5120,0.05181866553094652
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,10240,16384,0.14124977588653564
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,10240,4096,0.031042665243148804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,10240,4096,0.04198933309978909
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,10240,65536,0.4933350880940755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,10240,3584,0.02796622117360433
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,10240,3584,0.038457777765062116
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,10240,12288,0.11228799819946289
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,10240,3072,0.024889777104059856
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,10240,3072,0.03491200009981791
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,10240,10240,0.09933333264456855
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,10240,8192,0.08699555529488458
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,10240,2560,0.031656887796190046
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,10240,2560,0.021673777037196692
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,10240,2048,0.01829333272245195
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,10240,2048,0.027969777584075928
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,10240,7168,0.07688533597522311
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,10240,1536,0.014689778288205465
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,10240,1536,0.02493155499299367
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,10240,6144,0.07161955700980292
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,10240,1024,0.011527111132939657
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,10240,1024,0.021611554755104914
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,10240,5120,0.06485866838031344
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,10240,768,0.010030222435792288
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,10240,768,0.02047911120785607
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,10240,3584,0.05200799968507555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,10240,512,0.008398222426573435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,10240,512,0.018000000052981906
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,10240,4096,0.05543733305401272
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,10240,256,0.005955555372767978
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,10240,256,0.01641244524055057
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,10240,3072,0.04744088980886671
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,10240,128,0.004855111241340637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,10240,128,0.015676443775494892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,10240,2560,0.044066667556762695
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,10240,64,0.004417777889304691
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,10240,64,0.015805333852767944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,10240,32,0.0046968890560997855
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,10240,32,0.015671110815472074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,10240,1536,0.0373617774910397
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,10240,768,0.03479555580351088
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,10240,2048,0.039988444911109075
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,10240,1024,0.03478044602606032
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,8192,16384,0.0910373330116272
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,8192,65536,0.3238168822394477
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,8192,16384,0.08283378018273248
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,8192,12288,0.07320888837178548
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,8192,12288,0.06319910950130887
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,8192,10240,0.05348089006212023
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,8192,65536,0.3202284442053901
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,8192,10240,0.06459911002053155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,8192,8192,0.04409866531689962
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,10240,128,0.026862222287389968
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,10240,256,0.028607110182444256
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,8192,8192,0.05379466546906365
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,8192,7168,0.03898933198716905
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,8192,7168,0.04993599984380934
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,8192,6144,0.034479111433029175
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,8192,6144,0.0452133317788442
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,10240,512,0.030485332012176514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,8192,5120,0.029735111527972754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,8192,5120,0.04061333338419596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,8192,16384,0.12772711118062338
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,8192,4096,0.025320000118679468
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,8192,4096,0.03333866596221924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,8192,12288,0.10235644711388482
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,8192,65536,0.4389386706882053
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,8192,3584,0.02231377859910329
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,8192,3584,0.03175910976197984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,8192,3072,0.02000799940692054
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,8192,3072,0.028697778781255085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,8192,10240,0.0925182236565484
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,8192,2560,0.017278222574128043
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,8192,2560,0.02615822189384037
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,8192,8192,0.07666933536529541
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,8192,2048,0.01424533294306861
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,8192,2048,0.02366222275627984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,8192,7168,0.07066577672958374
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,8192,1536,0.011643555429246692
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,8192,1536,0.021776000658671062
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,8192,6144,0.0652302238676283
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,8192,1024,0.009272000028027428
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,8192,1024,0.019725332657496136
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,8192,5120,0.05959288941489326
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,8192,768,0.008098666866620382
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,8192,768,0.01847555571132236
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,8192,4096,0.05070133341683281
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,8192,512,0.006856888946559694
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,8192,512,0.01643111142847273
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,8192,3584,0.04782488942146301
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,8192,256,0.005025777965784073
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,8192,256,0.015410666664441427
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,8192,3072,0.044060445494121976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,8192,128,0.004472889006137848
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,8192,128,0.014688889185587565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,8192,2560,0.04173955652448866
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,8192,64,0.003998222036494149
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,8192,64,0.014740443891949125
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,8192,32,0.004348444441954295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,8192,32,0.014441778262456259
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,8192,2048,0.037984000311957464
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,8192,1536,0.03572088811132643
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,8192,1024,0.03345777922206455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,8192,768,0.033182223637898765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,7168,16384,0.08913333548439874
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,7168,16384,0.08552533388137817
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,7168,65536,0.3153715663486057
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,7168,12288,0.07227644655439588
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,7168,12288,0.058259553379482694
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,8192,512,0.0288080010149214
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,7168,65536,0.29629866282145184
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,8192,256,0.028460443019866943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,7168,10240,0.04751555456055535
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,7168,10240,0.06368622514936659
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,7168,8192,0.05303200085957845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,8192,128,0.025598223010698955
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,7168,8192,0.04443200098143684
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,7168,7168,0.04805066519313388
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,7168,7168,0.034444444709353976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,7168,6144,0.044403556320402354
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,7168,6144,0.02999733222855462
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,7168,5120,0.025997334056430395
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,7168,5120,0.03901244534386529
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,7168,16384,0.1274062262641059
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,7168,4096,0.023984889189402264
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,7168,65536,0.43860801060994464
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,7168,4096,0.032327999671300255
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,7168,12288,0.1004035539097256
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,7168,3584,0.02117777698569828
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,7168,3584,0.03013955553372701
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,7168,10240,0.08866222037209405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,7168,3072,0.028182221783532038
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,7168,3072,0.017510222064124215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,7168,8192,0.0752426650789049
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,7168,2560,0.025906667113304138
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,7168,2560,0.015119110544522604
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,7168,2048,0.012574222352769641
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,7168,2048,0.023414222730530634
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,7168,7168,0.06892444690068562
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,7168,1536,0.010608889162540436
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,7168,1536,0.021613332960340712
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,7168,6144,0.06373866399129231
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,7168,1024,0.008540444076061249
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,7168,1024,0.019263111882739596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,7168,5120,0.058209776878356934
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,7168,768,0.007517333659860823
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,7168,768,0.017129777206314933
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,7168,4096,0.05046933227115207
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,7168,512,0.006146666490369373
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,7168,512,0.01608622239695655
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,7168,3584,0.04768622252676222
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,7168,256,0.0047333331571684945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,7168,256,0.015211555692884656
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,7168,3072,0.04406933320893181
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,7168,128,0.004208000169859992
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,7168,128,0.01458755632241567
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,7168,2560,0.04135377870665656
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,7168,64,0.0037804444630940757
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,7168,64,0.01440622243616316
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,7168,32,0.003965333518054751
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,7168,32,0.014579556054539151
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,7168,1536,0.035869333479139544
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,7168,2048,0.037956443097856306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,7168,1024,0.033843557039896645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,7168,768,0.03322844372855292
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,6144,16384,0.07457066906823052
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,6144,16384,0.0881884429189894
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,6144,65536,0.3088577853308784
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,6144,12288,0.05556977457470364
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,6144,12288,0.07074399789174397
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,6144,65536,0.24368355009290907
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,6144,10240,0.04399289025200737
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,6144,10240,0.0653217766020033
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,7168,512,0.02867822183503045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,6144,8192,0.0399982233842214
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,7168,256,0.02823466724819607
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,6144,8192,0.05212355653444926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,6144,7168,0.04799733228153653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,7168,128,0.026389333936903212
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,6144,7168,0.03218222326702542
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,6144,6144,0.04314755400021871
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,6144,6144,0.027450667487250432
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,6144,5120,0.025369778275489807
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,6144,5120,0.037717332442601524
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,6144,16384,0.12342845069037543
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,6144,4096,0.022479999396536086
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,6144,4096,0.031712889671325684
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,6144,65536,0.4188222090403239
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,6144,3584,0.019326221611764696
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,6144,12288,0.1006622248225742
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,6144,3584,0.029890666405359905
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,6144,3072,0.016599110431141324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,6144,3072,0.02770044406255086
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,6144,10240,0.08829777770572239
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,6144,2560,0.01346933344999949
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,6144,2560,0.025568000144428674
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,6144,8192,0.0758462217119005
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,6144,2048,0.011753777662913004
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,6144,2048,0.023548444112141926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,6144,7168,0.06943200031916301
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,6144,1536,0.009621333744790819
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,6144,1536,0.021464889248212177
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,6144,6144,0.06410221921073066
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,6144,5120,0.05860800213283963
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,6144,1024,0.007855999800893996
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,6144,1024,0.019293333093325298
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,6144,768,0.0069875551594628235
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,6144,768,0.017076444294717576
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,6144,3584,0.04731111062897576
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,6144,512,0.005615111026499007
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,6144,512,0.016188444362746347
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,6144,4096,0.04998488889800178
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,6144,256,0.004456000195609199
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,6144,256,0.015245333313941956
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,6144,2560,0.04075110952059428
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,6144,128,0.003999111139112049
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,6144,128,0.014715555641386243
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,6144,768,0.03322488731808133
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,6144,64,0.003658666792843077
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,6144,64,0.014369777507252164
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,6144,32,0.0038897775941424896
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,6144,32,0.0144177774588267
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,6144,3072,0.04340888725386726
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,6144,1536,0.035535110367669
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,6144,2048,0.03756711218092177
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,5120,65536,0.3068648974100749
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,6144,1024,0.03223022156291538
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,5120,16384,0.06926221979988946
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,5120,16384,0.08575999736785889
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,5120,65536,0.22169600592719185
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,5120,12288,0.05362577901946174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,5120,12288,0.06927200158437093
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,5120,10240,0.06106488572226631
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,5120,10240,0.04288799895180596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,5120,8192,0.03888622257444594
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,5120,8192,0.0505528880490197
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,6144,128,0.02533688810136583
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,5120,7168,0.031354665756225586
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,6144,256,0.027928001350826685
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,5120,7168,0.04571288824081421
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,5120,6144,0.027059555053710938
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,6144,512,0.02832888894610935
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,5120,6144,0.042057779100206166
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,5120,5120,0.023507555325826008
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,5120,5120,0.03695733348528544
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,5120,16384,0.11863022380405003
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,5120,4096,0.020619556307792664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,5120,65536,0.39086400138007266
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,5120,4096,0.031559109687805176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,5120,3584,0.01868266695075565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,5120,3584,0.029262221521801416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,5120,10240,0.08532000250286526
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,5120,12288,0.09685689210891724
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,5120,3072,0.015356444650226168
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,5120,3072,0.027408000495698717
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,5120,2560,0.013581333061059317
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,5120,2560,0.02548800077703264
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,5120,8192,0.07323022021187676
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,5120,2048,0.010439111126793755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,5120,2048,0.023268444670571223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,5120,7168,0.06706311305363973
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,5120,1536,0.008944888909657797
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,5120,1536,0.021291555629836187
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,5120,6144,0.06178577740987142
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,5120,1024,0.007221333682537079
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,5120,1024,0.01811822255452474
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,5120,5120,0.05614933040406969
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,5120,768,0.006619555668698416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,5120,768,0.01682577696111467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,5120,3584,0.0454711119333903
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,5120,512,0.00527555536892679
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,5120,512,0.015981333123313058
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,5120,2560,0.04013155566321479
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,5120,256,0.004401777767472797
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,5120,256,0.015172445111804538
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,5120,4096,0.04846133457289802
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,5120,128,0.0038817777401871155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,5120,128,0.01442577772670322
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,5120,3072,0.0418177776866489
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,5120,64,0.003605333467324575
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,5120,64,0.014275555809338888
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,5120,32,0.003735110991530948
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,5120,768,0.03118844495879279
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,5120,32,0.014087110757827759
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,5120,2048,0.036701334847344294
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,4096,65536,0.305660433239407
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,5120,1536,0.03490044342146979
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,4096,16384,0.08343555529912312
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,4096,16384,0.0510168903403812
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,4096,12288,0.04169155491722954
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,4096,12288,0.06733422146903144
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,4096,65536,0.16836978329552543
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,5120,1024,0.032448000378078885
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,4096,10240,0.035331557194391884
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,4096,10240,0.05938577651977539
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,4096,8192,0.031168888012568157
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,4096,8192,0.049807111422220864
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,5120,128,0.024919110867712233
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,4096,7168,0.027629334065649245
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,4096,7168,0.044827553961012095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,5120,256,0.02755022214518653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,4096,6144,0.022295999858114455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,4096,6144,0.04056888818740845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,5120,512,0.027634667025672063
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,4096,5120,0.01998222205373976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,4096,5120,0.03659111261367798
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,4096,16384,0.10586844550238715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,4096,4096,0.016898666818936665
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,4096,4096,0.03131466772821214
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,4096,65536,0.36250845591227215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,4096,3584,0.015376889043384127
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,4096,12288,0.0858924455112881
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,4096,3584,0.02919911013709174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,4096,3072,0.012814222110642327
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,4096,3072,0.027273777458402846
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,4096,10240,0.07613066832224528
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,4096,2560,0.011167111496130625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,4096,2560,0.025281776984532673
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,4096,8192,0.0663599967956543
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,4096,2048,0.009580444130632613
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,4096,2048,0.023186667097939387
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,4096,7168,0.06075555748409695
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,4096,1536,0.008282666405042013
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,4096,1536,0.019931554794311523
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,4096,6144,0.05663910839292738
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,4096,1024,0.006785777707894643
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,4096,1024,0.01761511133776771
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,4096,5120,0.051817778084013194
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,4096,768,0.005791999813583162
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,4096,768,0.0166311115026474
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,4096,4096,0.045251554912990995
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,4096,512,0.004794666750563515
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,4096,512,0.015956444872750174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,4096,3584,0.04306133257018196
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,4096,256,0.004036444342798657
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,4096,256,0.015102222561836243
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,4096,3072,0.04023289018207126
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,4096,128,0.0036346668170558084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,4096,128,0.014518222875065274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,4096,2560,0.038387556870778404
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,4096,64,0.0033599999215867785
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,4096,64,0.014195554786258273
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,4096,32,0.003528888854715559
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,4096,32,0.014055111342006259
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,4096,2048,0.03554577628771464
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3584,65536,0.30560000737508136
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,4096,1536,0.033965332640541926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3584,16384,0.0469395551416609
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3584,65536,0.1502951118681166
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3584,16384,0.0832479993502299
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3584,12288,0.03750844465361701
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3584,12288,0.0666195551554362
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,4096,1024,0.03129244512981839
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,4096,768,0.029841777351167467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3584,10240,0.03327200147840712
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,4096,512,0.027154665854242113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,4096,256,0.027264000640975103
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3584,10240,0.05855288770463732
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3584,8192,0.048560000128216214
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,4096,128,0.024832000335057575
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3584,7168,0.023665777511066858
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3584,8192,0.026271111435360376
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3584,7168,0.044900443818834096
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3584,6144,0.04037955403327942
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3584,6144,0.02079288827048408
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3584,5120,0.017836444907718234
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3584,5120,0.03615199857287937
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,3584,16384,0.1067697803179423
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3584,4096,0.014547554983033074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3584,4096,0.03089600139194065
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,3584,65536,0.3649368815951877
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3584,3584,0.013036444783210754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3584,3584,0.029120001527998183
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,3584,12288,0.08650044600168864
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3584,3072,0.011678222152921887
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3584,3072,0.027225777506828308
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,3584,8192,0.06698133548100789
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,3584,10240,0.07650044229295519
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3584,2560,0.010926222635640038
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3584,2560,0.02532711128393809
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,3584,7168,0.06102666589948866
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3584,2048,0.00889066689544254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,3584,6144,0.056890666484832764
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3584,2048,0.02222844461599986
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3584,1536,0.007816889219813878
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3584,1536,0.01975022256374359
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3584,1024,0.006182222316662471
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3584,1024,0.017321777012613084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,3584,5120,0.05212444398138258
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3584,768,0.005389333185222413
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3584,768,0.016501333978441026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,3584,4096,0.04548177785343594
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3584,512,0.004483555754025777
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3584,512,0.015807999504937064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,3584,3584,0.04350489046838549
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3584,256,0.0038168889780839286
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3584,256,0.014977777997652689
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,3584,3072,0.04024089044994778
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3584,128,0.003432000055909157
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3584,128,0.014269333746698169
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,3584,2560,0.03849422269397312
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3584,64,0.0032586666444937387
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3584,64,0.013971555564138623
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3584,32,0.0033679999825027254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3584,32,0.013974222871992322
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,3584,1536,0.03187733226352268
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,3584,2048,0.035826666487587824
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,3584,768,0.03010666701528761
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,3584,1024,0.030322667625215318
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3072,16384,0.08259466621610853
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3072,16384,0.043643554051717125
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3072,12288,0.03442400031619602
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3072,65536,0.30360621876186794
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3072,12288,0.06618666648864746
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3072,10240,0.03156355685657925
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3072,10240,0.058424002594418
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3072,65536,0.12830577956305608
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3072,8192,0.02682133350107405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3072,8192,0.048693332407209605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,3584,512,0.02769244379467434
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,3584,256,0.027250667413075764
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3072,7168,0.021440888444582622
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3072,7168,0.044263998667399086
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,3584,128,0.02479733361138238
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3072,6144,0.018962666392326355
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3072,6144,0.040215111441082425
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3072,5120,0.016899555921554565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3072,5120,0.03602133194605509
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,3072,16384,0.10912622345818414
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3072,4096,0.014261333478821648
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3072,4096,0.030759112702475652
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,3072,12288,0.08908533387713963
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,3072,65536,0.3571137852138943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3072,3584,0.012334222594896952
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3072,3584,0.02905422117975023
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3072,3072,0.011113777756690979
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3072,3072,0.026829333768950567
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,3072,10240,0.07824799749586317
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3072,2560,0.00979200005531311
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3072,2560,0.024957334001859028
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,3072,8192,0.06686755683686998
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3072,2048,0.008624888956546783
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3072,2048,0.022099556194411382
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,3072,7168,0.06093511316511366
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3072,1536,0.007190222541491191
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3072,1536,0.019341333044899833
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,3072,6144,0.05600088834762573
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3072,1024,0.0058622223635514574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3072,1024,0.017307556337780423
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,3072,5120,0.051423999998304576
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3072,768,0.004856889032655292
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3072,768,0.016512889001104567
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,3072,4096,0.0443484452035692
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3072,512,0.004233777936961916
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3072,512,0.015740444262822468
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,3072,3584,0.04204977883232964
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3072,256,0.0037288889288902283
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3072,256,0.014908444550302295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,3072,2560,0.037224001354641385
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,3072,3072,0.03945422172546387
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3072,128,0.0033760000434186724
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3072,128,0.014197332991494073
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3072,64,0.0031591111587153543
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3072,64,0.01423288881778717
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,3072,32,0.0033555556502607134
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,3072,32,0.013850666582584381
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,3072,2048,0.033915556139416166
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2560,65536,0.30478045675489635
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2560,65536,0.12306755118899876
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2560,16384,0.04064622190263536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,3072,1536,0.031508445739746094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2560,16384,0.08195644617080688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2560,12288,0.032496889432271324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2560,12288,0.0656800005171034
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,3072,768,0.029172443681293066
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,3072,1024,0.029343999094433252
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2560,10240,0.057877335283491343
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2560,10240,0.02735377848148346
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2560,8192,0.025219556358125474
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2560,8192,0.04859022299448649
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,3072,128,0.024715556038750544
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,3072,256,0.027152000202072993
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,3072,512,0.027229333917299908
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2560,7168,0.043785777356889516
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2560,7168,0.01846311158604092
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2560,6144,0.016044444508022733
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2560,6144,0.03996710975964864
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2560,5120,0.013891556196742587
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2560,5120,0.035896887381871544
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,2560,65536,0.33493955930074054
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2560,4096,0.011891555454995898
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2560,4096,0.030488000975714788
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,2560,16384,0.09884088569217259
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2560,3584,0.010791999598344168
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2560,3584,0.02868088748719957
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,2560,12288,0.08116533358891805
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2560,3072,0.009902222288979424
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2560,3072,0.026712889472643535
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,2560,10240,0.07219822539223565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2560,2560,0.0090506664580769
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2560,2560,0.024361777636739943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,2560,8192,0.06240266561508179
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2560,2048,0.007791111038790808
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2560,2048,0.02214844524860382
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,2560,6144,0.05285066697332594
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2560,1536,0.00664533343580034
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,2560,7168,0.05716977516810099
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2560,1536,0.01927911076280806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2560,1024,0.005167110926575131
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2560,1024,0.017031111650996737
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,2560,5120,0.049207998646630176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2560,768,0.0046302220887608
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2560,768,0.016353777713245816
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,2560,4096,0.043826666143205434
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2560,512,0.0041306668685542215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2560,512,0.015519110692871941
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,2560,3584,0.04122399952676561
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2560,256,0.003622222277853224
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2560,256,0.014684443672498068
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,2560,3072,0.03858844439188639
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2560,128,0.0033902221669753394
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2560,128,0.014076444837782117
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,2560,2560,0.03684622380468581
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2560,64,0.003072888900836309
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2560,64,0.013975999421543546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,2560,2048,0.032772445016437106
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2560,32,0.0031999999450312722
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2560,32,0.01386666629049513
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,2560,1536,0.03074755602412754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2048,65536,0.3045208983951145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,2560,768,0.029370665550231934
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2048,65536,0.09173688623640273
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,2560,1024,0.029461334149042766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2048,16384,0.031488001346588135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2048,16384,0.08141511016421847
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2048,12288,0.06570666366153292
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2048,12288,0.027041777968406677
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2048,10240,0.05753511190414429
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2048,10240,0.02385777731736501
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2048,8192,0.02095111045572493
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2048,8192,0.04792444573508369
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,2560,512,0.02693866689999898
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2048,7168,0.01715555621518029
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,2560,256,0.027365333504146997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2048,7168,0.04338755541377597
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2048,6144,0.01440088947614034
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2048,6144,0.03952888978852166
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,2560,128,0.02462844385041131
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2048,5120,0.011874666644467248
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2048,5120,0.03535733289188809
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,2048,12288,0.0759448872672187
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,2048,16384,0.09310755464765762
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2048,4096,0.010288000106811523
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2048,4096,0.030751112434599135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,2048,10240,0.06788088877995808
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2048,3584,0.009472000102202097
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2048,3584,0.027481777800454035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2048,3072,0.008589333130253686
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,2048,8192,0.05870044231414795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2048,3072,0.026337777574857075
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,2048,65536,0.3095208803812663
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2048,2560,0.007840000092983246
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2048,2560,0.023402666052182514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2048,2048,0.007071110937330458
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2048,2048,0.02122933334774441
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,2048,7168,0.05377422107590569
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2048,1536,0.006134222365087933
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2048,1536,0.018974221414989896
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,2048,6144,0.05039733317163256
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2048,1024,0.004905777672926585
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2048,1024,0.017111111018392775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,2048,5120,0.04690666662322151
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2048,768,0.004496888981925117
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2048,768,0.016229333149062265
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,2048,4096,0.04152444336149428
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2048,512,0.004016889052258598
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,2048,3584,0.03993511199951172
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2048,512,0.015448888142903646
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2048,256,0.003555555724435382
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,2048,3072,0.03720533185535007
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2048,256,0.014720888601409065
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2048,128,0.0032879999942249725
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2048,128,0.014083556003040738
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,2048,2560,0.03547733359866672
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2048,64,0.003100444459252887
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2048,64,0.013923555612564087
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,2048,32,0.003248888792263137
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,2048,32,0.013735999663670858
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,2048,2048,0.033287998702791
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1536,65536,0.30441154374016655
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,2048,1536,0.030064000023735896
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1536,65536,0.07190489106708102
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1536,16384,0.08172444502512614
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,2048,1024,0.02818044357829624
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1536,16384,0.02587555514441596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1536,12288,0.06539733542336358
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1536,12288,0.021125333176718816
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1536,10240,0.02074044446150462
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,2048,768,0.028714666763941448
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1536,10240,0.05744889047410753
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1536,8192,0.018944889307022095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1536,8192,0.047469334469901196
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,2048,512,0.026649778087933857
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1536,7168,0.014924445086055331
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,2048,256,0.026847110854254827
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1536,7168,0.043475555049048535
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,2048,128,0.024352888266245525
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1536,6144,0.013382222089502545
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1536,6144,0.03913244605064392
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1536,5120,0.011790222591824003
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1536,5120,0.035306665632459856
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,1536,16384,0.09163555834028457
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1536,4096,0.010196444061067369
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,1536,65536,0.3046915531158447
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1536,4096,0.029870222012201946
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,1536,12288,0.07534488704469469
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1536,3584,0.009447111023796929
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1536,3584,0.028201778729756672
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1536,3072,0.00867555538813273
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,1536,10240,0.06685955656899346
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1536,3072,0.025469332933425903
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1536,2560,0.00793422261873881
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1536,2560,0.023167999254332647
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,1536,8192,0.057850665516323514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1536,2048,0.006990222467316522
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1536,2048,0.02083288960986667
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,1536,7168,0.053215109639697604
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1536,1536,0.005759999983840519
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,1536,6144,0.049419555399152965
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1536,1536,0.019023999571800232
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1536,1024,0.0047377778424157035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1536,1024,0.017157332764731515
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,1536,5120,0.04659644431538052
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1536,768,0.004351110921965705
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1536,768,0.016181333197487723
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,1536,4096,0.041096889310412936
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1536,512,0.003855111284388436
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1536,512,0.015338665909237333
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,1536,3584,0.03928177886539035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1536,256,0.0034435554924938413
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,1536,3072,0.036919110351138644
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1536,256,0.014519999424616495
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1536,128,0.003192000091075897
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1536,128,0.014074666632546319
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,1536,2560,0.03422133458985223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1536,64,0.003032888803217146
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1536,64,0.013839999834696451
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1536,32,0.003068444422549672
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1536,32,0.013762666947311826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1024,65536,0.05296355485916138
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,1536,2048,0.03133777777353922
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1024,65536,0.30316710472106934
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1024,16384,0.021204445097181533
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1024,16384,0.08185599909888373
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,1536,1536,0.02952444553375244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1024,12288,0.016419554750124615
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1024,12288,0.06532088915506999
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,1536,768,0.028728889094458684
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,1536,1024,0.028188443846172754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1024,10240,0.01608888970481025
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1024,10240,0.05637600024541219
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1024,8192,0.013728000223636627
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1024,8192,0.046784000264273755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,1536,512,0.02667466633849674
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,1536,256,0.02681688964366913
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1024,7168,0.012533333566453723
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1024,7168,0.04257066713439094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,1536,128,0.0244195560614268
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1024,6144,0.011805333197116852
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1024,6144,0.03864533371395535
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1024,5120,0.010830221904648675
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1024,5120,0.03487288951873779
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,1024,65536,0.29484444194369847
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,1024,16384,0.08897955550087823
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1024,4096,0.010484444598356882
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1024,4096,0.02907466557290819
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1024,3584,0.00889066689544254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1024,3584,0.027327999472618103
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,1024,12288,0.07303288910124037
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1024,3072,0.008085333638721043
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1024,3072,0.024838222397698298
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,1024,10240,0.06479733520083956
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1024,2560,0.007426666716734569
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,1024,8192,0.056141330136193164
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1024,2560,0.022943110929595098
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1024,2048,0.006507555643717448
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1024,2048,0.02088444431622823
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,1024,7168,0.051903112067116626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,1024,6144,0.04821066723929512
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1024,1536,0.005610666755172942
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1024,1536,0.018823110395007663
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1024,1024,0.004853333450025982
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1024,1024,0.01692800058258904
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,1024,5120,0.04527022110091316
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1024,768,0.004342222379313575
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1024,768,0.01612444387541877
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,1024,4096,0.04032889008522034
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1024,512,0.0037671112351947357
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,1024,3584,0.03861777649985419
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1024,512,0.015206222732861837
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1024,256,0.0032968889507982465
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1024,256,0.014422222971916199
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,1024,3072,0.03472888800832961
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1024,128,0.0030337776988744736
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1024,128,0.013799111048380533
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,1024,2560,0.03405688868628608
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1024,64,0.0028719999310043124
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,1024,2048,0.030424889591005113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1024,64,0.013819555441538492
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,1024,32,0.0029279999434947968
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,1024,32,0.0136835558546914
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,768,65536,0.3041999869876438
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,768,65536,0.042393777105543345
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,768,16384,0.01703733371363746
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,768,16384,0.08265955580605401
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,1024,1536,0.02915999955601162
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,768,12288,0.013726222018400827
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,768,12288,0.06544799937142266
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,1024,768,0.028259555498758953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,1024,1024,0.027951111396153767
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,768,10240,0.012278222375445895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,768,10240,0.05560177564620972
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,768,8192,0.013001778059535556
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,768,8192,0.04658755660057068
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,1024,512,0.02659911082850562
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,1024,256,0.026533333791626826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,768,7168,0.012134222520722283
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,768,7168,0.042832887834972806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,768,6144,0.011295999917719098
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,768,6144,0.03866133424970839
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,1024,128,0.02492800023820665
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,768,5120,0.010597333312034607
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,768,5120,0.03449866506788466
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,768,16384,0.08921511305703057
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,768,4096,0.0094355551732911
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,768,65536,0.2933902210659451
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,768,4096,0.028945777151319716
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,768,3584,0.008522666990756989
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,768,3584,0.026850667264726426
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,768,12288,0.07295911179648505
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,768,3072,0.007706666986147563
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,768,10240,0.06445244285795423
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,768,3072,0.024759110477235582
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,768,2560,0.006491555521885554
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,768,2560,0.022802667485343084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,768,8192,0.0556817783249749
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,768,2048,0.0058044443527857465
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,768,2048,0.020814221766259935
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,768,7168,0.05180266830656263
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,768,1536,0.005120888766315249
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,768,1536,0.018887110882335238
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,768,6144,0.04795822170045641
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,768,1024,0.004376888689067629
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,768,1024,0.016978666186332703
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,768,5120,0.04538844360245598
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,768,768,0.004024888906213972
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,768,768,0.016134222348531086
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,768,4096,0.03791200121243795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,768,512,0.003680888977315691
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,768,3584,0.037399109866884016
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,768,512,0.01533066729704539
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,768,256,0.0032755554550223877
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,768,256,0.014589332871966891
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,768,3072,0.033839109871122576
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,768,128,0.002986666642957263
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,768,128,0.014035556051466199
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,768,2560,0.03243822190496657
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,768,64,0.0028231110837724474
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,768,64,0.013869333598348828
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,768,32,0.0029075555503368378
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,768,32,0.013629333012633853
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,768,2048,0.030335999197430078
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,512,65536,0.2999848789638943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,512,65536,0.04535911149448819
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,512,16384,0.012896888785892062
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,512,16384,0.08166844314999051
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,768,1536,0.029314666986465454
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,512,12288,0.011519999967681037
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,512,12288,0.06398222181532118
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,768,1024,0.027708444330427382
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,512,10240,0.010386666489972008
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,512,10240,0.05574133661058214
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,768,768,0.028186665640936956
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,512,8192,0.010566222170988718
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,512,8192,0.04673066735267639
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,768,512,0.026188444760110643
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,512,7168,0.01018044435315662
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,768,256,0.026534222894244727
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,512,7168,0.04240355557865567
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,512,6144,0.009526222116417354
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,512,6144,0.038176000118255615
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,768,128,0.024280889166726008
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,512,5120,0.00905688852071762
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,512,5120,0.033569776349597506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,512,16384,0.08722488747702704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,512,4096,0.008426666259765625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,512,65536,0.28912533654106987
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,512,4096,0.02872266703181797
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,512,3584,0.007910221815109253
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,512,3584,0.026515555050637987
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,512,12288,0.07172711027993096
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,512,3072,0.007479111353556315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,512,3072,0.024633778466118708
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,512,10240,0.06324444214502971
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,512,2560,0.00647199981742435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,512,2560,0.022622222701708477
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,512,8192,0.055289778444502086
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,512,2048,0.0058115555180443665
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,512,2048,0.020781333247820537
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,512,7168,0.050979554653167725
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,512,1536,0.005014222115278244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,512,1536,0.01884622209601932
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,512,6144,0.04732977681689792
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,512,1024,0.004316444612211651
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,512,1024,0.016945777667893302
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,512,5120,0.04452000061670939
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,512,768,0.003934222377008862
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,512,768,0.016016888949606154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,512,4096,0.03794666793611314
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,512,512,0.0036346668170558084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,512,3584,0.03594577643606398
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,512,512,0.015110222829712762
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,512,256,0.0032302221904198327
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,512,256,0.014422222971916199
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,512,3072,0.03390489021937052
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,512,128,0.002965333354141977
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,512,128,0.01367111090156767
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,512,2560,0.03208533260557387
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,512,2048,0.030020445585250854
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,512,64,0.002806222273243798
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,512,32,0.002850666642189026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,512,64,0.013693333499961428
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,512,32,0.01350044459104538
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,256,65536,0.30029688941107857
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,256,65536,0.031810667779710554
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,256,16384,0.010804444551467896
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,256,16384,0.08153333266576131
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,512,1536,0.028928889168633357
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,256,12288,0.00979022185007731
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,256,12288,0.06415910853279962
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,512,1024,0.02769777841038174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,256,10240,0.008918222453859117
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,256,10240,0.055496888028250806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,512,768,0.028219554159376357
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,256,8192,0.009666666388511658
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,256,8192,0.045287112394968666
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,512,256,0.026650667190551758
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,512,512,0.026429333620601233
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,256,7168,0.009177777502271865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,256,7168,0.042118221521377563
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,256,6144,0.008404444489214156
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,256,6144,0.037857777542538114
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,512,128,0.02384800049993727
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,256,5120,0.007463110817803278
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,256,5120,0.0329146683216095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,256,65536,0.28779999415079754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,256,16384,0.08709155850940281
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,256,4096,0.00702311098575592
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,256,4096,0.028451555305057104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,256,3584,0.007426666716734569
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,256,3584,0.026523555318514507
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,256,12288,0.07128710879219903
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,256,3072,0.007017778025733099
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,256,3072,0.02459022237194909
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,256,10240,0.06293600135379367
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,256,2560,0.006284444282452266
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,256,2560,0.02252799934811062
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,256,8192,0.054670221275753446
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,256,2048,0.005543111099137201
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,256,2048,0.020555555820465088
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,256,7168,0.050675554407967456
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,256,1536,0.004914666629499859
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,256,6144,0.04703377683957418
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,256,1536,0.018605333235528734
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,256,1024,0.004292444636424382
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,256,1024,0.016731555263201397
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,256,5120,0.042436444097095065
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,256,768,0.00397511116332478
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,256,768,0.015810666812790763
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,256,4096,0.0368755559126536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,256,512,0.003605333467324575
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,256,512,0.030328889687856037
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,256,3584,0.035579555564456515
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,256,256,0.0032302221904198327
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,256,256,0.014325333966149224
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,256,3072,0.03319199879964193
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,256,128,0.0029279999434947968
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,256,128,0.013668444421556262
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,256,2560,0.031785776217778526
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,256,64,0.0027884443600972495
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,256,64,0.013563555147912769
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,256,32,0.0028515555378463534
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,256,32,0.013411555853154926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,128,65536,0.020806221498383414
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,128,65536,0.3045484489864773
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,256,2048,0.02968888812594944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,128,16384,0.00962755580743154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,128,16384,0.08076444599363539
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,256,1536,0.028607110182444256
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,128,12288,0.008344888687133789
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,256,1024,0.027415111660957336
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,128,12288,0.06367644336488512
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,128,10240,0.0075786668393347
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,128,10240,0.053782221343782216
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,256,768,0.028021332290437486
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,128,8192,0.007164444360468123
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,128,8192,0.045287112394968666
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,256,512,0.026319111386934917
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,256,256,0.025991110338105097
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,128,7168,0.00699733363257514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,128,7168,0.04043733411365085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,128,6144,0.00674311113026407
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,128,6144,0.03651911020278931
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,256,128,0.023919999599456787
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,128,5120,0.007184888753626082
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,128,5120,0.03260088960329691
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,128,65536,0.2883795632256402
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,128,4096,0.006647999915811751
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,128,4096,0.028241776757770117
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,128,16384,0.08664711316426595
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,128,3584,0.007148444652557373
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,128,3584,0.0262106673585044
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,128,12288,0.07095733616087171
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,128,3072,0.006843555718660355
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,128,10240,0.06280977858437432
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,128,3072,0.02425777746571435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,128,2560,0.006159111029571957
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,128,2560,0.022479999396536086
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,128,8192,0.05369866556591458
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,128,2048,0.005527110977305307
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,128,2048,0.02052533295419481
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,128,7168,0.04939555459552341
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,128,1536,0.004815111143721474
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,128,1536,0.018639999959203932
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,128,6144,0.04430933462248909
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,128,1024,0.004129777765936322
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,128,1024,0.016773333152135212
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,128,5120,0.04181955589188469
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,128,768,0.003740444365474913
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,128,768,0.015784000356992085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,128,4096,0.03696533375316196
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,128,512,0.0034017778105205963
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,128,512,0.014886221951908536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,128,3584,0.03531022204293145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,128,256,0.003031111011902491
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,128,256,0.014195554786258273
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,128,3072,0.03316088848643833
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,128,128,0.0028239999794297745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,128,128,0.013575110998418597
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,128,2560,0.031665778822369046
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,128,64,0.0026702222724755607
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,128,64,0.013501333693663279
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,128,32,0.0026808888134029177
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,128,32,0.013281777501106262
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,64,65536,0.016932444439993966
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,64,65536,0.3045368989308675
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,64,16384,0.007393777370452881
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,64,16384,0.08123022317886353
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,64,12288,0.0069013333155049225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,64,12288,0.06383911106321546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,64,10240,0.007102222078376346
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,64,10240,0.05556355582343208
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,64,8192,0.007017778025733099
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,64,8192,0.04470488760206434
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,64,7168,0.0068195557428730865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,64,7168,0.0399528874291314
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,64,6144,0.006602666858169768
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,128,2048,0.029684444268544514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,64,5120,0.00702133360836241
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,64,6144,0.03609866566128201
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,64,4096,0.00654755574133661
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,64,5120,0.03243555625279745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,64,4096,0.02815822097990248
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,64,3584,0.007027555671003129
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,64,3072,0.006729777902364731
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,128,1536,0.02847911251915826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,64,3584,0.02630399995379978
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,64,2560,0.006059555543793573
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,64,3072,0.02442755632930332
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,64,2048,0.005392000079154968
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,64,2560,0.022463111413849726
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,64,1536,0.0047360000511010485
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,64,2048,0.020561777883105807
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,64,1024,0.004028444488843282
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,64,1536,0.018580444984965853
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,64,768,0.003677333394686381
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,64,1024,0.01668711172209846
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,64,512,0.0033528889632887314
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,64,768,0.01571999986966451
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,64,256,0.002987555538614591
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,64,512,0.014935111006100973
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,64,128,0.0028133332315418455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,64,256,0.014195554786258273
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,64,128,0.013516444298956128
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,64,64,0.0026693333768182327
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,64,32,0.002711999954448806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,64,64,0.013405332962671915
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,64,32,0.013210666676362356
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,32,65536,0.014922666880819531
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,32,16384,0.006760889043410619
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,32,12288,0.0069306666652361555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,32,16384,0.08129688766267565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,32,65536,0.3043217658996582
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,128,1024,0.027003554834259882
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,32,10240,0.007052444749408298
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,32,12288,0.0630906687842475
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,32,8192,0.00684533350997501
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,32,10240,0.05414311091105143
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,32,7168,0.006714666469229593
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,32,8192,0.04403911034266154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,32,6144,0.006452444526884291
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,32,7168,0.0402124457889133
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,32,5120,0.0069057775868309875
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,32,6144,0.03607111175855001
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,32,4096,0.0064497776329517365
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,32,5120,0.03233244352870517
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,32,3584,0.006940444310506185
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,32,4096,0.028246220615175035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,32,3072,0.006573333508438534
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,32,3584,0.026570666167471144
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,32,2560,0.006133333262470033
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,32,3072,0.024341333243581984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,32,2048,0.005492444253630108
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,32,2560,0.022450667288568284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,32,1536,0.0047422221137417685
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,32,2048,0.020604444874657523
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,32,1024,0.004063111212518481
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,32,1536,0.01870133313867781
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,128,768,0.02779910961786906
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,32,1024,0.0166640000210868
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,32,768,0.0036959999965296853
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,32,512,0.003342222215400802
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,32,768,0.01574577722284529
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,32,512,0.014914666612943014
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,32,256,0.0030248889492617715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,32,128,0.0028328889360030494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,32,256,0.014174222946166992
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,32,64,0.0026568888376156488
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,32,128,0.01349511080318027
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,32,64,0.013403555585278405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,256,32,32,0.0026435556097163092
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,256,32,32,0.013251555462678274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,128,512,0.025794666674402025
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,128,256,0.026429333620601233
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,256,128,128,0.024008888337347243
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,65536,12288,0.3719200028313531
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,65536,10240,0.31551021999782985
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,65536,12288,0.40894577238294816
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,65536,16384,0.4894791179233127
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,65536,16384,0.520759105682373
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,65536,12288,0.424690670437283
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,65536,8192,0.26313866509331596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,65536,10240,0.34343645307752824
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,65536,10240,0.35741599400838214
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,65536,8192,0.28773511780632866
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,65536,7168,0.2309804492526584
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,65536,16384,0.5577493243747288
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,65536,8192,0.28583023283216685
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,65536,7168,0.25185688336690265
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,65536,6144,0.19983377721574572
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,65536,7168,0.2560906675126818
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,65536,5120,0.16854755083719888
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,65536,6144,0.22231554985046387
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,65536,6144,0.21641955110761854
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,65536,5120,0.1893084446589152
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,65536,4096,0.1373680035273234
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,65536,5120,0.18269066015879312
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,65536,3584,0.12117511696285671
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,65536,4096,0.15414488315582275
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,65536,4096,0.1543955538007948
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,65536,3584,0.13858489195505777
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,65536,2560,0.09033866723378499
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,65536,3584,0.13789955774943033
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,65536,3072,0.10555999808841282
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,65536,2560,0.10790933503044976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,65536,3072,0.12396978007422553
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,65536,2560,0.10421599944432576
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,65536,3072,0.12148088879055446
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,65536,2048,0.07747289207246569
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,65536,1536,0.06116266383065117
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,65536,2048,0.09256088733673096
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,65536,2048,0.08607288863923813
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,65536,1024,0.046216888560189136
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,65536,1024,0.06214844518237644
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,65536,768,0.038104888465669416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,65536,512,0.03349511159790887
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,65536,1536,0.07587644126680163
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,65536,1024,0.04991199904017978
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,65536,768,0.05598844422234429
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,65536,768,0.04160177707672119
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,65536,1536,0.06850044594870673
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,65536,256,0.02013066742155287
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,65536,256,0.02832711074087355
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,65536,128,0.014755555325084262
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,65536,256,0.04232800006866455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,65536,128,0.025584888127115037
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,65536,512,0.05021689004368252
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,65536,512,0.029693331983354356
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,65536,128,0.03641688823699951
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,65536,64,0.01182044463025199
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,65536,32,0.013074666261672974
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,65536,64,0.027423999375767175
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,65536,32,0.029584887954923842
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,16384,12288,0.1163182258605957
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,16384,16384,0.15170133113861084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,16384,16384,0.18046489026811388
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,16384,12288,0.14075999789767796
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,16384,12288,0.11271022425757514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,16384,10240,0.09883289204703437
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,16384,10240,0.09527289205127293
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,16384,16384,0.1464257770114475
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,16384,8192,0.08250044451819526
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,16384,10240,0.1221182213889228
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,16384,8192,0.10537155469258626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,16384,7168,0.07461333274841309
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,16384,7168,0.094541331132253
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,16384,8192,0.07671378056208293
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,16384,6144,0.06635466549131605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,16384,7168,0.07056622372733222
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,16384,6144,0.08405866887834336
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,16384,6144,0.059529774718814425
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,16384,5120,0.05648799737294515
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,16384,65536,0.5549253357781304
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,16384,4096,0.04855555627081129
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,16384,5120,0.051111999485227794
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,16384,5120,0.07369510995017158
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,16384,4096,0.06273600127961901
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,16384,4096,0.046089778343836464
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,16384,3584,0.044216887818442456
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,16384,3584,0.03760622276200189
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,16384,3072,0.03984888725810581
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,16384,3584,0.05936533212661743
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,16384,3072,0.033428443802727595
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,16384,3072,0.05502133236991035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,16384,2560,0.03578844335344102
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,16384,2048,0.03142755561404758
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,16384,2560,0.028676443629794653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,16384,1536,0.027660444378852844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,16384,2560,0.0510115557246738
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,16384,2048,0.045759999089770846
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,16384,65536,0.6424391004774306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,16384,2048,0.02405244443151686
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,16384,1536,0.04223555657598707
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,16384,1024,0.02277066641383701
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,16384,1536,0.01923377811908722
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,16384,768,0.01129244433508979
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,16384,512,0.008798222574922774
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,16384,768,0.020844444632530212
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,16384,768,0.03588355580965678
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,16384,512,0.019838222199016146
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,16384,1024,0.013623110949993134
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,16384,256,0.0069617778062820435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,16384,128,0.00508977762526936
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,16384,512,0.0346942212846544
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,16384,256,0.016999999682108562
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,16384,128,0.015796444482273526
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,16384,64,0.004550222307443619
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,16384,256,0.03132355544302199
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,16384,32,0.00480000012450748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,16384,1024,0.03884622123506334
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,16384,128,0.02649066680007511
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,16384,64,0.016034666034910414
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,16384,32,0.016597333881590102
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,12288,16384,0.13702578014797634
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,12288,16384,0.13959556155734593
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,12288,12288,0.1042871144082811
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,12288,12288,0.0862355563375685
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,12288,16384,0.11347288555569118
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,12288,12288,0.11183288362291123
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,12288,10240,0.08788800239562988
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,12288,10240,0.0721084475517273
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,12288,10240,0.09702222214804755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,12288,8192,0.05935644441180759
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,12288,8192,0.07344710826873779
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,16384,65536,0.5979475445217556
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,12288,8192,0.08439733584721883
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,12288,7168,0.0671022203233507
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,12288,7168,0.0521031121412913
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,12288,7168,0.07837955819235908
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,12288,6144,0.05931466817855835
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,12288,6144,0.04536088969972399
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,12288,6144,0.0699146654870775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,12288,5120,0.039183109998703
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,12288,5120,0.050966223080952965
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,12288,5120,0.06255288918813069
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,12288,4096,0.04417422082689074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,12288,4096,0.03292177783118354
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,12288,4096,0.05407111181153191
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,12288,65536,0.47568713294135195
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,12288,3072,0.03712800145149231
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,12288,3584,0.04075555668936835
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,12288,3584,0.05145155390103658
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,12288,65536,0.5088835292392307
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,12288,2560,0.033032000064849854
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,12288,2560,0.045660446087519325
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,12288,2048,0.029297779003779095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,12288,3072,0.04887911015086704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,12288,3584,0.029890666405359905
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,12288,3072,0.027622222900390625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,12288,2560,0.0222044438123703
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,12288,2048,0.041045334604051374
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,12288,1536,0.025644444757037695
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,12288,1024,0.010779555473062726
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,12288,1024,0.022247999906539917
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,12288,1536,0.03877244393030802
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,12288,768,0.009160888691743216
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,12288,1024,0.03619733452796936
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,12288,2048,0.018464888135592144
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,12288,1536,0.014483556151390076
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,12288,768,0.020657777786254883
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,12288,512,0.0075128889746136135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,12288,512,0.0184568895234002
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,12288,768,0.03404444456100464
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,12288,256,0.005724444571468566
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,12288,512,0.033187554942237005
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,12288,256,0.016960889101028442
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,12288,128,0.004472889006137848
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,12288,256,0.02831022275818719
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,12288,128,0.015816888875431485
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,12288,64,0.004018666843573253
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,12288,32,0.004414222306675381
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,12288,128,0.026565333207448322
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,12288,64,0.01587199999226464
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,12288,32,0.016215999921162922
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,10240,16384,0.13096888860066733
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,10240,16384,0.132006221347385
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,10240,16384,0.10708889034059312
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,10240,12288,0.08067555560006036
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,10240,12288,0.10268800126181708
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,12288,65536,0.42755556106567383
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,10240,10240,0.08670933379067315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,10240,12288,0.10725777679019505
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,10240,10240,0.09159821934170193
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,10240,8192,0.0734951098759969
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,10240,8192,0.060630222161610924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,10240,10240,0.06860444280836317
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,10240,8192,0.08037866486443414
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,10240,65536,0.49581866794162327
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,10240,7168,0.0655004448360867
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,10240,7168,0.04890311095449659
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,10240,6144,0.05856533182991875
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,10240,7168,0.07611199882295397
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,10240,5120,0.04988266693221199
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,10240,4096,0.04324000080426534
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,10240,6144,0.06691022051705255
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,10240,4096,0.051755554146236844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,10240,5120,0.059752000702752
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,10240,6144,0.04211466511090597
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,10240,5120,0.03946311275164286
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,10240,4096,0.03040800160831875
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,10240,3584,0.03973244296179877
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,10240,3072,0.03563377923435635
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,10240,2560,0.02070844504568312
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,10240,3072,0.046359111865361534
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,10240,2560,0.03202133377393087
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,10240,3584,0.04961155520545112
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,10240,3584,0.0270480000310474
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,10240,2560,0.0432951119210985
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,10240,2048,0.028263999356163874
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,10240,1536,0.01371022231049008
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,10240,2048,0.03979555434650845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,10240,1536,0.0247013337082333
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,10240,3072,0.02383733292420705
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,10240,1024,0.010185778141021729
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,10240,2048,0.018598222070270114
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,10240,1536,0.03728533453411526
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,10240,768,0.008615111311276754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,10240,1024,0.021672889590263367
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,10240,768,0.02015999952952067
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,10240,1024,0.035056889057159424
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,10240,512,0.006983111302057902
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,10240,512,0.01793866687350803
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,10240,768,0.03363733490308126
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,10240,256,0.005305777821275923
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,10240,512,0.03194577826393975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,10240,256,0.016748444901572335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,10240,128,0.004524444540341695
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,10240,256,0.027878221538331773
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,10240,128,0.015596444408098856
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,10240,64,0.004026666697528628
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,10240,128,0.025599110457632277
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,10240,32,0.0041404445138242514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,10240,64,0.015698666373888653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,10240,32,0.016015110744370353
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,10240,65536,0.4607253074645996
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,8192,16384,0.09629333019256592
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,8192,12288,0.07456711265775892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,8192,12288,0.06723822487725152
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,8192,16384,0.08650844626956516
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,10240,65536,0.3820168972015381
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,8192,10240,0.06319110923343234
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,8192,65536,0.34241512086656356
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,8192,10240,0.052367998494042285
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,8192,8192,0.05576355589760674
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,8192,8192,0.04303644431961907
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,8192,7168,0.04974044362703959
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,8192,7168,0.0380604432688819
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,8192,6144,0.03340177734692892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,8192,6144,0.044572446081373424
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,8192,65536,0.3040124575297038
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,8192,5120,0.028757333755493164
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,8192,5120,0.03961155480808682
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,8192,16384,0.11654310756259495
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,8192,4096,0.024217777782016333
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,8192,4096,0.03420977791150411
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,8192,12288,0.09430844253963894
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,8192,3584,0.02142222225666046
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,8192,10240,0.08651110861036514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,8192,3584,0.031791110833485924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,8192,3072,0.019119110372331407
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,8192,3072,0.02906222144762675
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,8192,65536,0.3908382256825765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,8192,8192,0.07326400279998779
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,8192,2560,0.015434667468070984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,8192,2560,0.026272889640596177
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,8192,2048,0.012609778179062737
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,8192,2048,0.02390577726893955
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,8192,7168,0.06856622298558553
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,8192,1536,0.01036800030204985
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,8192,1536,0.021513778302404616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,8192,6144,0.06215911441379123
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,8192,1024,0.008383110993438298
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,8192,1024,0.019754666421148512
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,8192,5120,0.056290666262308754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,8192,768,0.007350222104125553
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,8192,768,0.018174222773975797
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,8192,3584,0.0463066664006975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,8192,512,0.006207110981146495
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,8192,512,0.01645866698688931
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,8192,4096,0.04847466614511278
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,8192,256,0.004440000073777305
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,8192,256,0.015455999308162265
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,8192,3072,0.04344977935155233
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,8192,128,0.003917333152559069
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,8192,128,0.014442667365074158
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,8192,1536,0.035855998595555626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,8192,64,0.003583999971548716
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,8192,64,0.01444355481200748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,8192,32,0.003918222255176968
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,8192,2560,0.04141777753829956
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,8192,32,0.014639111028777229
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,8192,2048,0.037666665183173284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,8192,768,0.032264888286590576
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,8192,1024,0.03421511252721151
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,7168,65536,0.32115112410651314
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,7168,16384,0.09282488955391778
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,7168,16384,0.07651644282870822
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,7168,16384,0.12051378356085883
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,7168,12288,0.07268089056015015
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,7168,12288,0.08412355846828884
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,7168,65536,0.2760213216145833
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,7168,65536,0.42585688167148167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,7168,12288,0.09612355629603068
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,7168,10240,0.06106311082839966
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,7168,10240,0.051065779394573636
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,7168,10240,0.08231200112236871
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,7168,8192,0.05411644445525276
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,7168,8192,0.07185777690675524
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,7168,7168,0.047982222504085965
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,7168,8192,0.040038221412234835
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,7168,7168,0.03400444322162204
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,7168,6144,0.04311199982961019
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,7168,7168,0.0663715534740024
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,7168,6144,0.029263999727037217
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,7168,6144,0.06004533502790663
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,7168,5120,0.026328888204362657
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,7168,5120,0.03802311089303758
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,7168,4096,0.033526221911112465
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,7168,4096,0.047225776645872325
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,7168,3584,0.030660443835788306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,7168,5120,0.05384622348679436
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,7168,4096,0.02132622235351139
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,7168,3584,0.04552533229192098
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,7168,3584,0.019111999207072787
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,7168,3072,0.027911110056771174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,8192,128,0.025070221887694463
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,7168,3072,0.04342933495839437
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,7168,2560,0.025734222597546045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,7168,2048,0.012401777837011548
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,7168,2048,0.023600000474188063
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,7168,2560,0.04055466585689121
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,7168,3072,0.016964443855815463
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,7168,2560,0.01517688896920946
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,7168,2048,0.03708622190687392
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,7168,1536,0.010398222340477837
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,7168,1536,0.021167111065652635
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,8192,256,0.02701688806215922
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,7168,1024,0.008101333346631791
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,7168,1536,0.035791999763912626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,7168,768,0.017239999439981248
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,7168,512,0.006151999864313338
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,7168,1024,0.019259555472267997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,7168,768,0.007181333170996771
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,7168,1024,0.034527109728919134
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,7168,768,0.03179644544919332
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,7168,512,0.016237333416938782
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,7168,256,0.004676444249020683
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,7168,512,0.02975733412636651
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,7168,256,0.015273777974976433
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,7168,128,0.004168888760937585
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,7168,256,0.027465777264700994
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,7168,128,0.01442311041884952
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,7168,64,0.003777777983082665
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,7168,128,0.025582222474945918
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,7168,32,0.004003555410438114
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,7168,64,0.014319111903508505
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,7168,32,0.014300444059901767
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,6144,16384,0.08529155784183079
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,8192,512,0.02991466720898946
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,6144,12288,0.05459377831882901
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,6144,12288,0.06778755452897814
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,6144,16384,0.06682044267654419
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,6144,10240,0.057542224725087486
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,6144,10240,0.048473778698179454
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,6144,8192,0.049844443798065186
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,6144,8192,0.04738666613896688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,6144,65536,0.30739111370510525
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,6144,7168,0.04597066508399116
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,6144,7168,0.03128533230887519
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,6144,6144,0.02940711047914293
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,6144,6144,0.041182223293516375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,6144,65536,0.22176178296407065
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,6144,5120,0.023853333459960088
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,6144,5120,0.03511111272705926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,6144,16384,0.10541155603196885
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,6144,12288,0.08497066630257501
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,6144,4096,0.022015111313925848
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,6144,4096,0.03134311238924662
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,6144,10240,0.07515377468532987
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,6144,8192,0.06593422094980876
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,6144,3584,0.01770666738351186
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,6144,3584,0.028912888632880315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,6144,3072,0.026886221435334947
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,6144,3072,0.015855999456511605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,6144,2560,0.013574221895800697
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,6144,2560,0.024824889169798955
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,6144,7168,0.06289777490827772
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,6144,2048,0.01110577748881446
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,6144,65536,0.3498835563659668
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,6144,2048,0.02276177704334259
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,6144,1536,0.010089777410030365
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,6144,1536,0.020711110697852243
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,6144,6144,0.05717955695258247
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,6144,1024,0.007505777809354994
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,6144,1024,0.01865688959757487
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,6144,5120,0.05175644490453932
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,6144,768,0.006461333483457565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,6144,768,0.017158221867349412
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,6144,4096,0.04545333319240146
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,6144,512,0.005491555564933353
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,6144,512,0.016318221886952717
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,6144,3584,0.044144001271989614
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,6144,256,0.004319111092223061
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,6144,256,0.015263110399246216
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,6144,3072,0.04218488931655884
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,6144,128,0.0038204445607132385
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,6144,2560,0.03993511199951172
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,6144,128,0.014223110344674854
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,6144,64,0.0034773333205117118
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,6144,64,0.0141004439857271
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,6144,32,0.0037093332244290244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,6144,32,0.014125333891974555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,6144,2048,0.03670577870474921
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,6144,1536,0.03530399998029073
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,6144,1024,0.03410133388307359
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,5120,16384,0.058818664815690785
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,6144,768,0.03175466590457492
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,5120,16384,0.08074844545788236
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,5120,65536,0.2845582167307536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,5120,12288,0.04872622092564901
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,5120,12288,0.06500799788369073
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,5120,65536,0.1958844396803114
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,5120,10240,0.05751111110051473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,5120,10240,0.0370053317811754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,5120,8192,0.0341635545094808
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,5120,8192,0.047784888082080416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,6144,256,0.026759111218982275
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,6144,512,0.029275556405385334
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,6144,128,0.025166221790843542
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,5120,7168,0.02716355522473653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,5120,7168,0.04292000002331204
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,5120,6144,0.038650665018293597
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,5120,6144,0.023552889625231426
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,5120,5120,0.020807999703619216
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,5120,5120,0.03383288780848185
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,5120,16384,0.11412710613674587
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,5120,4096,0.019395555059115093
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,5120,65536,0.39001422458224827
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,5120,4096,0.03007733490731981
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,5120,3584,0.015545777148670621
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,5120,3584,0.028078221612506445
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,5120,12288,0.09117511245939468
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,5120,3072,0.013359111216333179
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,5120,3072,0.026241777671708003
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,5120,10240,0.07861333423190646
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,5120,2560,0.01181244436237547
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,5120,2560,0.02408355474472046
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,5120,8192,0.06968533330493502
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,5120,2048,0.00996088898844189
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,5120,2048,0.022430222895410325
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,5120,7168,0.06470133198632134
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,5120,1536,0.008472888833946651
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,5120,1536,0.020510221521059673
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,5120,6144,0.05805866585837471
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,5120,1024,0.007466666400432587
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,5120,1024,0.018002667360835604
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,5120,5120,0.05150399936570061
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,5120,768,0.006495999793211619
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,5120,768,0.0170168893204795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,5120,4096,0.045365333557128906
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,5120,512,0.005377777748637729
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,5120,3584,0.04233333468437195
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,5120,512,0.016241777274343703
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,5120,256,0.004187555362780889
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,5120,256,0.015114666687117683
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,5120,3072,0.04120622078577677
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,5120,128,0.0036764442920684814
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,5120,128,0.014362666341993542
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,5120,2560,0.03868710994720459
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,5120,64,0.003350222276316749
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,5120,64,0.014016888207859464
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,5120,32,0.0034702221552530923
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,5120,32,0.01405244403415256
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,5120,2048,0.0354746679464976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,5120,1536,0.034055110481050276
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,5120,1024,0.0328693323665195
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,5120,768,0.029595557186338637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,4096,65536,0.27562223540412056
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,4096,65536,0.15735910998450384
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,4096,16384,0.051186667548285596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,4096,16384,0.07856177621417575
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,4096,12288,0.06271466943952772
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,4096,12288,0.05269777774810791
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,4096,10240,0.05387022097905477
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,4096,10240,0.042697777350743614
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,4096,8192,0.03786399960517883
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,5120,256,0.02677511175473531
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,5120,512,0.029126223590638902
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,4096,8192,0.04647199975119697
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,4096,7168,0.042394667863845825
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,4096,7168,0.030193779203626845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,4096,6144,0.02533955540921953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,4096,6144,0.03818044397566054
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,5120,128,0.025314667158656653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,4096,5120,0.02068800065252516
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,4096,5120,0.03323288758595785
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,4096,16384,0.1005751093228658
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,4096,4096,0.017044444878896076
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,4096,4096,0.029461334149042766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,4096,12288,0.08205511172612508
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,4096,3584,0.015311110350820752
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,4096,65536,0.3101946777767605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,4096,3584,0.027815110153622095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,4096,10240,0.07247466511196561
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,4096,3072,0.012671999633312225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,4096,3072,0.025939555631743536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,4096,2560,0.010418666733635796
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,4096,2560,0.02418311105834113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,4096,8192,0.06286844280030993
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,4096,2048,0.008854221966531541
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,4096,2048,0.022319111559126113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,4096,7168,0.05937688880496555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,4096,1536,0.007925333248244392
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,4096,1536,0.01978488928741879
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,4096,6144,0.05304977628919813
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,4096,1024,0.006224888894293044
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,4096,1024,0.01761777698993683
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,4096,5120,0.04822399881150988
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,4096,768,0.005261333452330695
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,4096,768,0.016952888833151925
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,4096,4096,0.042034665743509926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,4096,512,0.004652444273233414
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,4096,512,0.01605333387851715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,4096,3584,0.04077333211898804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,4096,3072,0.03894666830698649
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,4096,256,0.003887999802827835
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,4096,256,0.015096000499195524
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,4096,128,0.003576000117593341
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,4096,128,0.014249778456158109
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,4096,2560,0.03706399930848016
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,4096,64,0.0032764443506797156
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,4096,64,0.013943110903104147
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,4096,32,0.0035217776894569397
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,4096,32,0.013850666582584381
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,4096,768,0.028039111031426325
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,4096,1536,0.031027555465698242
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,4096,2048,0.03434755404790243
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,4096,1024,0.03125866585307651
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3584,16384,0.0459537771013048
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3584,65536,0.2674764527214898
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3584,16384,0.07605599694781832
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3584,12288,0.06178400251600477
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3584,65536,0.1441004408730401
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3584,12288,0.04823199907938639
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3584,10240,0.05292977889378866
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3584,10240,0.03897510965665182
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3584,8192,0.023728888895776536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3584,8192,0.04593422346644931
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,4096,128,0.024525332782003615
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,4096,256,0.026353778110610113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3584,7168,0.041639112763934664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3584,7168,0.023932443724738225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3584,6144,0.023710222707854375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,4096,512,0.028254220883051556
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3584,6144,0.037666665183173284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3584,5120,0.01814844376511044
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3584,5120,0.03293244375122918
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,3584,16384,0.09874844551086426
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3584,4096,0.015673778123325773
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3584,4096,0.02934488985273573
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,3584,12288,0.08089777496125963
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3584,3584,0.012568888564904531
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,3584,65536,0.30677154329088
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3584,3584,0.027499554885758296
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,3584,10240,0.07145688931147258
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3584,3072,0.01088444474670622
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3584,3072,0.025848888688617285
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3584,2560,0.01148444414138794
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3584,2560,0.023953777220514085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,3584,8192,0.062217778629726835
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3584,2048,0.009662222531106737
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3584,2048,0.022142221530278523
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,3584,7168,0.058357331487867564
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3584,1536,0.007999999655617608
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3584,1536,0.01922044489118788
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,3584,6144,0.05232888791296217
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,3584,5120,0.04716711242993673
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3584,1024,0.0063946665161185795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3584,1024,0.017575111654069688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3584,768,0.005151999907361137
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3584,768,0.016781333420011733
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,3584,3584,0.04026577870051066
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3584,512,0.004515555583768421
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3584,512,0.015994666351212394
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,3584,4096,0.04157333241568671
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3584,256,0.0039022221333450745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3584,256,0.014920888675583733
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,3584,2560,0.03695111142264472
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3584,128,0.0034328889515664843
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3584,128,0.014164444473054675
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,3584,3072,0.03875199953715006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3584,64,0.003145777723855443
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3584,64,0.014069333672523499
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3584,32,0.0032186667538351486
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3584,32,0.013762666947311826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,3584,2048,0.033949332104788885
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3072,65536,0.2678248882293701
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3072,65536,0.12155021561516656
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3072,16384,0.04102755586306254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3072,16384,0.0750560031996833
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,3584,1536,0.03278133273124695
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3072,12288,0.03135644396146139
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,3584,1024,0.029677334758970473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,3584,768,0.027895109521018133
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3072,12288,0.0612915555636088
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3072,10240,0.05272355675697327
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3072,10240,0.02686311139000787
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3072,8192,0.022579555710156757
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3072,8192,0.04522933231459724
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,3584,128,0.025041777226659987
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,3584,256,0.026294221480687458
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3072,7168,0.019853333632151287
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3072,7168,0.040538665321138166
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3072,6144,0.017650667164060805
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3072,6144,0.03700000047683716
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,3584,512,0.02800977892345852
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3072,5120,0.015057777365048727
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3072,5120,0.032474666833877563
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,3072,65536,0.36781954765319824
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3072,4096,0.012975999878512489
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3072,4096,0.029221332735485498
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,3072,16384,0.10650222169028388
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3072,3584,0.011525332927703857
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3072,3584,0.027440889014138117
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,3072,12288,0.08466133144166733
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3072,3072,0.010371555884679159
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,3072,10240,0.07489599784215291
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3072,3072,0.0258933338854048
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3072,2560,0.010397333237859938
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3072,2560,0.023823999696307715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,3072,8192,0.06600177950329252
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3072,2048,0.00852711084816191
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3072,2048,0.02095288866096073
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,3072,7168,0.061306668652428516
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3072,1536,0.007668444679843054
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3072,1536,0.019141332970725167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,3072,6144,0.05528088741832309
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3072,1024,0.00582311095462905
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3072,1024,0.01755555636352963
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,3072,5120,0.04946400059594048
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3072,768,0.005048000150256687
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3072,768,0.01665422154797448
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,3072,4096,0.04340177774429321
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3072,512,0.00442755553457472
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,3072,3584,0.04085955686039395
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3072,512,0.01591022147072686
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3072,256,0.0038515557017591265
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3072,256,0.014919110470347934
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,3072,3072,0.040431999497943454
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3072,128,0.003547555456558863
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3072,128,0.013966222604115805
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,3072,2560,0.03722044494416978
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3072,64,0.0031351111829280853
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3072,64,0.013838222457302941
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,3072,32,0.0032248888164758682
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,3072,32,0.013823111024167804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,3072,2048,0.0332640012105306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2560,65536,0.2650257746378581
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,3072,1536,0.03261333372857835
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,3072,768,0.028389334678649902
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,3072,1024,0.030221333106358845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2560,16384,0.03758488761054145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2560,65536,0.10392710897657607
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2560,16384,0.07429511017269559
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2560,12288,0.061034666167365186
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2560,12288,0.030307554536395605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2560,10240,0.05223555697335137
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2560,10240,0.02667288813326094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2560,8192,0.01853155593077342
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2560,8192,0.04483111037148369
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,3072,512,0.02883111106024848
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2560,7168,0.01627200014061398
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,3072,256,0.026535110341178045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2560,7168,0.041026665104760066
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2560,6144,0.014129777749379476
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,3072,128,0.02482755482196808
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2560,6144,0.037583112716674805
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2560,5120,0.012223111258612739
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2560,5120,0.03264000018437704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,2560,12288,0.0678044425116645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,2560,16384,0.0818551116519504
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2560,4096,0.012401777837011548
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,2560,10240,0.060589333375295006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2560,4096,0.029283556673261855
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,2560,65536,0.2605404324001736
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2560,3584,0.011026666396194033
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2560,3584,0.027319111757808264
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2560,3072,0.00887822194231881
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2560,3072,0.025677333275477093
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2560,2560,0.008426666259765625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2560,2560,0.023015111684799194
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,2560,8192,0.05329422156016032
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2560,2048,0.007499555746714274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2560,2048,0.020848888489935134
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,2560,7168,0.05076977941724989
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2560,1536,0.006534222099516127
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2560,1536,0.01899111105336083
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,2560,6144,0.04748444424735176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2560,1024,0.005136000199450387
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2560,1024,0.017124444246292114
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,2560,5120,0.04343555702103508
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2560,768,0.004435555388530095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2560,768,0.016474665866957772
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,2560,4096,0.03840444485346476
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2560,512,0.003991999973853429
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2560,512,0.015649777319696214
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,2560,3584,0.03787111242612203
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,2560,3072,0.036729776197009616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2560,256,0.003499555504984326
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2560,256,0.014674666855070325
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2560,128,0.0032879999942249725
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2560,128,0.013984888792037964
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,2560,2560,0.03512444429927402
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2560,64,0.003061333257291052
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2560,64,0.013923555612564087
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2560,32,0.003119999956753519
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2560,32,0.013609777722093793
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,2560,2048,0.03151911165979173
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,2560,1536,0.030642665094799463
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2048,65536,0.26815112431844074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2048,65536,0.08769066466225518
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2048,16384,0.07469511032104492
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2048,16384,0.030545777744717065
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2048,12288,0.029567999972237483
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2048,12288,0.06094666984346178
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,2560,1024,0.02879200047916836
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2048,10240,0.025056888659795124
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,2560,768,0.02751200066672431
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2048,10240,0.05165511038568285
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2048,8192,0.023208889696333144
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2048,8192,0.04440711273087395
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,2560,512,0.027715555495686
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2048,7168,0.015319110618697273
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,2560,256,0.025994666748576697
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,2560,128,0.02466222147146861
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2048,7168,0.04044444362322489
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2048,6144,0.01351999988158544
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2048,6144,0.036978665325376726
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2048,5120,0.011930666863918304
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2048,5120,0.03230488962597317
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,2048,16384,0.08061955372492473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2048,4096,0.010486221975750394
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2048,4096,0.029030223687489826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,2048,65536,0.25507020950317383
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2048,3584,0.009429333110650381
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,2048,12288,0.06705422533882989
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2048,3584,0.027119111683633592
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2048,3072,0.008652444514963362
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2048,3072,0.024615110622511968
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,2048,10240,0.059952000776926674
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2048,2560,0.007896000312434303
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2048,2560,0.02247200078434414
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,2048,8192,0.053258667389551796
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2048,2048,0.007111999723646376
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2048,2048,0.020423110988405015
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,2048,6144,0.046296000480651855
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2048,1536,0.006275555739800136
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,2048,5120,0.04250311189227634
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2048,1536,0.018917333748605516
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,2048,7168,0.05029955506324768
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2048,1024,0.004872888740566042
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2048,1024,0.01714666684468587
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2048,768,0.004341333276695675
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2048,768,0.01632000009218852
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,2048,4096,0.03820177912712097
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2048,512,0.0038817777401871155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2048,512,0.015578667322794596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,2048,3584,0.03745155533154806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2048,256,0.0035146665241983202
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2048,256,0.01461955573823717
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,2048,3072,0.036361777120166354
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2048,128,0.0032355555643637977
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2048,128,0.013919111755159167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,2048,2560,0.035022222333484225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2048,64,0.003031111011902491
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,2048,2048,0.030528889762030706
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2048,64,0.013763555222087435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,2048,32,0.00309955556359556
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,2048,32,0.013549333645237817
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1536,65536,0.07212799787521362
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1536,65536,0.2674044503106011
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1536,16384,0.027148443791601393
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1536,16384,0.0750017762184143
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,2048,1536,0.02938488788074917
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,2048,1024,0.02861600120862325
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,2048,768,0.02720622221628825
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1536,12288,0.060177776548597545
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1536,12288,0.02163377735349867
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1536,10240,0.019296889503796894
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1536,10240,0.050833778248892895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1536,8192,0.016336888074874878
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1536,8192,0.043859554661644824
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,2048,256,0.025891555680169
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,2048,512,0.02791911032464769
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1536,7168,0.015030221806632148
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1536,7168,0.04043911231888665
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,2048,128,0.024107555548350017
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1536,6144,0.013861333330472311
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1536,6144,0.03681866659058465
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1536,5120,0.012608889076444836
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1536,5120,0.03254400028122796
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,1536,65536,0.2531893253326416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,1536,16384,0.07958666483561198
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1536,4096,0.010242666635248397
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1536,4096,0.028140445550282795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1536,3584,0.009337777892748514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1536,3584,0.026855111122131348
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,1536,12288,0.06625511248906453
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,1536,10240,0.059008002281188965
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1536,3072,0.008551110823949179
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1536,3072,0.024516445067193773
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1536,2560,0.007690666450394525
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1536,2560,0.02235200007756551
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,1536,8192,0.051626665724648364
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1536,2048,0.006872889068391588
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1536,2048,0.020586666133668687
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,1536,7168,0.050119111935297646
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1536,1536,0.0058017778727743365
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,1536,6144,0.046610666645897754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1536,1536,0.018817777434984844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1536,1024,0.004652444273233414
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,1536,5120,0.04288799895180596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1536,1024,0.017088888419999015
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1536,768,0.004290666845109728
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1536,768,0.016297777493794758
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,1536,4096,0.03837066557672288
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1536,512,0.003801777958869934
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1536,512,0.015263110399246216
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,1536,3584,0.037314666642083064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1536,256,0.003375111147761345
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1536,256,0.014479110638300577
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,1536,3072,0.03601333167817857
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1536,128,0.003088888815707631
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1536,128,0.01389422184891171
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,1536,2560,0.034119112624062434
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1536,64,0.003015111097031169
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1536,64,0.013779555757840475
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,1536,2048,0.030136889881557886
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1536,32,0.003083555648724238
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1536,32,0.01349066694577535
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1024,65536,0.2664746708340115
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1024,65536,0.05144977900716993
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1024,16384,0.02256177696916792
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1024,16384,0.0742844475640191
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,1536,1536,0.029394666353861492
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,1536,1024,0.028746667835447524
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,1536,768,0.02760533326201969
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1024,12288,0.016578666037983365
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1024,12288,0.05951466825273302
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1024,10240,0.01481599940194024
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1024,10240,0.05085511008898417
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1024,8192,0.01312355531586541
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1024,8192,0.04348355531692505
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,1536,512,0.02758488886886173
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1024,7168,0.01218933363755544
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1024,7168,0.03969866699642605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,1536,256,0.02590222160021464
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1024,6144,0.011395555403497485
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,1536,128,0.024097777075237695
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1024,6144,0.036644445525275335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1024,5120,0.010500444306267632
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1024,5120,0.03152711192766825
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,1024,16384,0.09766222370995416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1024,4096,0.00960622231165568
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,1024,65536,0.3192604382832845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1024,4096,0.027507555153634813
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1024,3584,0.008918222453859117
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1024,3584,0.025829333398077223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,1024,12288,0.07714933156967163
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1024,3072,0.008060444560315874
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1024,3072,0.024141333169407312
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,1024,10240,0.06740799877378675
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1024,2560,0.007338666253619724
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1024,2560,0.02218933403491974
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,1024,8192,0.05965955389870537
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1024,2048,0.0059262220230367445
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1024,2048,0.020390222469965618
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,1024,7168,0.056926223966810435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1024,1536,0.005205333232879639
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1024,1536,0.018246221873495314
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,1024,6144,0.05064444409476387
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1024,1024,0.004378666480382283
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,1024,5120,0.04639644424120585
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1024,1024,0.017087999317381117
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1024,768,0.004104888687531154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1024,768,0.01627111103799608
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,1024,4096,0.04127466678619385
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1024,512,0.0036968886852264404
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,1024,3584,0.0373564428753323
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1024,512,0.015436444017622205
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1024,256,0.00332622230052948
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1024,256,0.014616888430383472
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,1024,3072,0.036215109957589045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1024,128,0.00305600009030766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1024,128,0.013711999687883588
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,1024,2560,0.03401599989997016
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1024,64,0.0028897778441508612
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1024,64,0.013639999760521783
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,1024,32,0.0028453332682450614
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,1024,32,0.01348533315791024
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,768,65536,0.04151555564668443
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,768,65536,0.26819554964701336
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,1024,2048,0.03137955400678847
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,768,16384,0.014953777194023132
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,768,16384,0.07430044147703382
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,1024,1536,0.03074399961365594
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,768,12288,0.013036444783210754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,768,12288,0.05924266576766968
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,1024,1024,0.029234667619069416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,768,10240,0.011922666596041786
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,1024,768,0.027907556957668726
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,768,10240,0.05061866508589851
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,768,8192,0.01110311100880305
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,768,8192,0.043486220969094165
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,1024,512,0.027668444646729365
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,1024,256,0.026330666409598455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,768,7168,0.010668444136778513
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,768,7168,0.039855112632115684
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,768,6144,0.009992000129487779
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,768,6144,0.035862220658196345
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,1024,128,0.024098666177855596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,768,5120,0.009324444664849175
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,768,5120,0.032002667586008705
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,768,65536,0.2944284545050727
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,768,4096,0.009477333062224919
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,768,16384,0.09276266892751057
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,768,4096,0.02776799930466546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,768,3584,0.008532444636027018
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,768,3584,0.025898666845427618
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,768,12288,0.07449155383639865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,768,10240,0.06528266933229235
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,768,3072,0.007671111159854465
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,768,3072,0.02383733292420705
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,768,2560,0.006620444357395172
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,768,2560,0.022134222918086584
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,768,8192,0.057856890890333385
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,768,2048,0.00574311117331187
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,768,2048,0.02049333353837331
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,768,7168,0.05484177668889364
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,768,1536,0.005110222018427319
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,768,1536,0.018820444742838543
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,768,6144,0.04943822158707512
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,768,1024,0.004327110946178436
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,768,1024,0.017072889539930556
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,768,5120,0.045143110884560474
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,768,768,0.003983111017280155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,768,768,0.01625422139962514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,768,4096,0.03892711136076186
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,768,512,0.0035955554081334006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,768,512,0.01553422212600708
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,768,3584,0.03696444299485948
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,768,256,0.003220444545149803
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,768,256,0.01457333399189843
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,768,3072,0.03542133172353109
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,768,128,0.002990222225586573
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,768,128,0.013567999833159976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,768,2560,0.0332293344868554
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,768,64,0.0027946666296985415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,768,2048,0.03085244364208645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,768,64,0.013763555222087435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,768,32,0.0028400001012616684
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,768,32,0.013479111095269522
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,512,65536,0.2641226715511746
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,512,65536,0.03289955523278978
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,512,16384,0.012830221818553077
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,512,16384,0.07393422391679552
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,768,1536,0.029902223083708022
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,512,12288,0.011063111325105032
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,768,1024,0.02882844540807936
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,512,12288,0.058223108450571694
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,512,10240,0.010239111052619087
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,512,10240,0.050219555695851646
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,768,768,0.02756533357832167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,512,8192,0.0094355551732911
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,512,8192,0.04296177625656128
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,768,512,0.02769511110252804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,512,7168,0.008904889225959778
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,512,7168,0.039108445247014366
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,768,256,0.025984888275464375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,512,6144,0.008488000267081791
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,512,6144,0.03541600041919284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,768,128,0.023950222465727065
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,512,5120,0.009707555174827576
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,512,5120,0.031014220582114324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,512,16384,0.07532888650894165
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,512,65536,0.23752710554334852
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,512,4096,0.008374222450786168
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,512,4096,0.026808889375792608
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,512,3584,0.007705777883529663
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,512,3584,0.02529777752028571
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,512,12288,0.0627795524067349
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,512,3072,0.007314666277832455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,512,10240,0.055748442808787026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,512,3072,0.023602666126357183
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,512,2560,0.006344888773229387
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,512,2560,0.0218231115076277
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,512,8192,0.04923911227120293
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,512,2048,0.005671111245950063
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,512,2048,0.020043555233213637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,512,7168,0.04755555589993795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,512,1536,0.004945777770545748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,512,1536,0.01828177769978841
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,512,6144,0.04374044471316867
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,512,1024,0.0042053332759274375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,512,1024,0.016792888442675274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,512,5120,0.039056890540652804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,512,768,0.003903999924659729
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,512,768,0.015726221932305228
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,512,4096,0.0343191126982371
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,512,512,0.0035208890007601846
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,512,3584,0.03351644343800015
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,512,512,0.015236443943447538
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,512,256,0.003167111012670729
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,512,256,0.014380445082982382
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,512,3072,0.03391111228201125
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,512,128,0.0029315555261241067
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,512,128,0.013779555757840475
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,512,2560,0.03156800071398417
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,512,64,0.0027582223216692605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,512,64,0.013624889155228933
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,512,32,0.002785777880085839
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,512,32,0.013383111192120446
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,256,65536,0.030245333909988403
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,256,65536,0.2669146590762668
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,512,2048,0.029285334878497656
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,256,16384,0.009320889082219865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,256,16384,0.0727928876876831
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,512,1536,0.028852442900339764
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,512,1024,0.02824444572130839
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,256,12288,0.01019200020366245
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,256,12288,0.058231108718448214
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,256,10240,0.009264888862768808
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,256,10240,0.049901333120134145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,512,768,0.02689244515366024
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,256,8192,0.008193777667151557
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,256,8192,0.04170133339034187
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,512,512,0.027403554982609216
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,256,7168,0.007734222544564142
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,256,7168,0.038141333394580416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,512,256,0.025588444537586633
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,512,128,0.02423733307255639
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,256,6144,0.007446222007274628
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,256,6144,0.034480889638264976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,256,5120,0.0075191110372543335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,256,5120,0.030230220821168687
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,256,65536,0.23656177520751953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,256,4096,0.007075555622577667
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,256,4096,0.026675555441114638
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,256,16384,0.07428799735175239
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,256,3584,0.007311999797821045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,256,3584,0.0251528885629442
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,256,12288,0.06195200151867337
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,256,3072,0.0069777775141927935
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,256,3072,0.02331377731429206
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,256,10240,0.05518577827347649
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,256,2560,0.006267555471923616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,256,2560,0.02163377735349867
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,256,8192,0.0481333335240682
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,256,2048,0.005573333137565189
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,256,2048,0.019817777805858187
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,256,7168,0.04550133479966057
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,256,1536,0.004831111265553368
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,256,1536,0.01812266641192966
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,256,6144,0.04215555720859104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,256,1024,0.0041448887851503165
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,256,1024,0.01641688909795549
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,256,5120,0.03860444492763943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,256,768,0.003825777934657203
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,256,768,0.015672889020707872
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,256,4096,0.03394666645261977
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,256,512,0.0034764446318149567
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,256,512,0.014902222487661572
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,256,3584,0.03345511025852627
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,256,256,0.003101333354910215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,256,256,0.014290667242474027
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,256,3072,0.0326222214433882
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,256,128,0.0028657778683635923
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,256,128,0.013565333353148567
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,256,2560,0.03126044405831231
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,256,64,0.002700444517864121
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,256,64,0.013487999637921652
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,256,32,0.002741333304180039
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,256,32,0.01328000012371275
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,128,65536,0.021707556313938562
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,256,2048,0.02905333373281691
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,128,65536,0.26846665806240505
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,128,16384,0.007853333320882585
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,128,16384,0.07292266686757405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,256,1536,0.02853155467245314
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,128,12288,0.006894222150246303
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,128,12288,0.0564791096581353
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,256,768,0.02660177813635932
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,256,1024,0.027906666199366253
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,128,10240,0.007314666277832455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,128,10240,0.04906488789452446
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,128,8192,0.007116444408893585
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,128,8192,0.040976001156700984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,256,512,0.02693333393997616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,128,7168,0.007015110717879401
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,128,7168,0.03735288977622986
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,256,256,0.02551288902759552
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,128,6144,0.006689777804745569
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,128,6144,0.034817778401904635
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,256,128,0.023826667004161414
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,128,5120,0.007101332975758447
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,128,5120,0.030167112747828167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,128,65536,0.2373510996500651
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,128,4096,0.006678222368160884
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,128,4096,0.026721777187453374
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,128,16384,0.0743724438879225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,128,3584,0.007122666471534305
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,128,3584,0.025007999605602686
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,128,12288,0.061553776264190674
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,128,3072,0.006735999965005451
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,128,3072,0.023112888137499493
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,128,10240,0.05528622203403049
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,128,2560,0.006175111151403851
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,128,2560,0.021405332618289526
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,128,8192,0.046543111403783165
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,128,2048,0.005406222409672207
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,128,2048,0.019807111885812547
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,128,7168,0.045838223563300244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,128,1536,0.004754666652944353
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,128,1536,0.018016889691352844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,128,6144,0.04189777705404493
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,128,1024,0.004108444270160464
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,128,1024,0.01637066735161675
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,128,5120,0.03812355465359158
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,128,768,0.003681777666012446
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,128,768,0.015623110863897534
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,128,4096,0.03404622276624044
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,128,512,0.0034017778105205963
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,128,3584,0.033286223808924355
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,128,512,0.014799111419253878
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,128,256,0.0030462222380770575
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,128,256,0.013971555564138623
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,128,3072,0.03259999884499444
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,128,128,0.002846222163902389
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,128,128,0.01348088930050532
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,128,2560,0.03120000163714091
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,128,64,0.0027022223091787765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,128,2048,0.028796444336573284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,128,64,0.013384888569513956
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,128,32,0.0026488889836602737
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,128,1536,0.028159111738204956
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,128,32,0.013232000172138214
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,64,65536,0.013953778478834363
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,64,16384,0.00778399987353219
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,64,12288,0.007078222102589077
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,64,16384,0.07268266545401679
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,64,65536,0.26690488391452366
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,64,10240,0.00701155596309238
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,64,12288,0.05702755848566691
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,64,8192,0.0068986668354935115
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,64,10240,0.048120889398786754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,64,7168,0.00675999994079272
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,64,8192,0.040503111150529646
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,64,6144,0.006600889066855113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,64,7168,0.03716444306903415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,64,5120,0.0069413334131240845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,64,6144,0.0338986681567298
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,64,4096,0.006522666662931442
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,64,5120,0.02998488810327318
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,64,3584,0.006978666616810693
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,64,4096,0.026521777113278706
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,64,3072,0.006659555352396435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,64,3584,0.024903999434577093
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,64,2560,0.006224000205596288
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,64,3072,0.02313422163327535
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,64,2560,0.021480000681347314
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,64,2048,0.005408000200986862
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,64,1536,0.0047404443224271136
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,64,2048,0.019648889700571697
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,64,1024,0.0040755557517210645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,64,1536,0.01796000036928389
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,64,768,0.003649777836269803
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,64,1024,0.016321778297424316
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,64,512,0.0033440000067154565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,64,768,0.01553066737122006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,64,256,0.0029893333299292457
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,64,512,0.01478488908873664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,64,128,0.002809777855873108
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,64,256,0.0141004439857271
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,64,128,0.013393777940008374
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,64,64,0.0026835555003749
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,64,64,0.013258666627936892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,64,32,0.0026613333159022858
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,64,32,0.013175110850069257
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,32,65536,0.013630222115251752
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,32,16384,0.007894222107198503
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,32,65536,0.26690223481920033
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,32,16384,0.07175288597742717
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,32,12288,0.007335999773608313
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,32,10240,0.0069324444565508105
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,32,12288,0.05788533555136787
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,32,8192,0.006800000038411882
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,32,10240,0.04741155438952976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,32,7168,0.00665244460105896
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,32,8192,0.040491554472181536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,32,6144,0.006478222293986215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,32,7168,0.03714222378200955
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,32,5120,0.006828444285525217
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,32,6144,0.033907555871539645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,32,4096,0.0064560001095136
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,32,5120,0.029940442906485662
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,32,4096,0.026520889666345384
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,32,3584,0.0069217777086628815
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,32,3072,0.006703111032644908
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,32,3584,0.025050666597154405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,32,3072,0.023183999790085688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,32,2560,0.006060444646411472
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,32,2048,0.005381333331267039
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,32,2560,0.0214355554845598
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,32,1536,0.0047004446387290955
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,32,2048,0.019714666737450492
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,32,1536,0.018018666240904067
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,32,1024,0.004026666697528628
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,32,1024,0.016220443778567843
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,32,768,0.003664888855483797
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,32,512,0.0032968889507982465
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,32,768,0.015440889530711703
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,32,256,0.0030097777230872046
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,32,512,0.014780445231331719
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,32,128,0.0027991111079851785
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,32,256,0.01404266721672482
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,32,64,0.0026515554636716843
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,32,128,0.013423999978436364
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,192,32,32,0.0026817777090602448
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,32,64,0.013247999880048962
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,192,32,32,0.01310222182008955
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,128,1024,0.02793333265516493
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,128,768,0.026758222116364375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,128,512,0.02702577743265364
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,128,256,0.02517866591612498
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,65536,12288,0.4290026558770074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,65536,16384,0.5294604301452637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,65536,16384,0.5870310995313857
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,65536,12288,0.44085245662265354
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,65536,10240,0.36098843150668675
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,65536,10240,0.3417199982537164
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,65536,10240,0.3886257807413737
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,65536,8192,0.30442576938205296
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,65536,12288,0.40968622101677793
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,65536,8192,0.3304568926493327
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,65536,7168,0.26861243777804905
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,65536,8192,0.2763208813137478
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,65536,7168,0.2890302340189616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,65536,7168,0.24281599786546496
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,65536,16384,0.5471386909484863
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,65536,6144,0.2291004392835829
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,65536,5120,0.19458222389221191
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,65536,6144,0.22754311561584473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,65536,4096,0.1558613379796346
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,65536,5120,0.20927911334567598
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,192,128,128,0.023747555083698694
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,65536,6144,0.21107021967569986
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,65536,5120,0.19158666663699678
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,65536,4096,0.17237155967288545
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,65536,3584,0.13813155227237278
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,65536,4096,0.14623822106255427
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,65536,3584,0.1542017724778917
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,65536,3584,0.12846399678124323
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,65536,3072,0.12286667029062907
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,65536,3072,0.13698932859632704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,65536,2560,0.102801779905955
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,65536,3072,0.1128728919559055
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,65536,2048,0.08694933520423041
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,65536,1536,0.06720711125267877
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,65536,2560,0.0963840021027459
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,65536,1536,0.0821217762099372
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,65536,2048,0.07971733146243624
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,65536,2560,0.11107555362913345
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,65536,2048,0.10163288646274143
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,65536,1024,0.04437333345413208
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,65536,1536,0.0627066691716512
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,65536,768,0.03786044319470724
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,65536,1024,0.0637155572573344
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,65536,1024,0.04540444413820902
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,65536,512,0.03506133291456435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,65536,256,0.028965334097544353
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,65536,768,0.055952886740366616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,65536,256,0.04128444525930617
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,65536,768,0.03583911061286926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,65536,128,0.01164533363448249
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,65536,512,0.026844445202085707
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,65536,512,0.04895111256175571
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,65536,64,0.009641778137948778
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,65536,256,0.017141333884663053
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,65536,128,0.025642666551801894
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,65536,64,0.028388443920347426
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,65536,32,0.009648000200589498
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,65536,128,0.03546311126814948
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,65536,32,0.0299893319606781
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,16384,16384,0.19043289290534127
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,16384,16384,0.1657279994752672
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,16384,12288,0.12927644782596165
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,16384,12288,0.10492266549004449
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,16384,16384,0.1374088923136393
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,16384,10240,0.10906133386823867
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,16384,12288,0.14860799577501085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,16384,10240,0.12906932830810547
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,16384,8192,0.09336088763342963
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,16384,10240,0.08767733309004043
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,16384,8192,0.11169955465528701
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,16384,8192,0.07156088617112902
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,16384,7168,0.08431555827458699
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,16384,7168,0.06323644187715319
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,16384,6144,0.07219644387563069
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,16384,6144,0.08647911416159736
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,16384,7168,0.10070844491322835
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,16384,6144,0.055252446068657764
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,16384,5120,0.06212800078921848
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,16384,5120,0.04751822352409363
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,16384,4096,0.052256001366509326
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,16384,5120,0.0777991082933214
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,16384,65536,0.6014284557766384
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,16384,4096,0.06548533174726698
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,16384,3584,0.048114667336146034
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,16384,4096,0.040604445669386126
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,16384,3584,0.0613022247950236
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,16384,3584,0.0351875556839837
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,16384,3072,0.04345600141419304
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,16384,3072,0.033561779393090144
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,16384,3072,0.05710666709476047
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,16384,2560,0.03839911023775736
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,16384,2560,0.05127822359402975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,16384,2048,0.03352800011634827
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,16384,2560,0.0269804447889328
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,16384,1536,0.028957333829667833
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,16384,2048,0.04758755697144402
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,16384,2048,0.022498667240142822
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,16384,768,0.010528888967302112
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,16384,1024,0.023338667220539514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,16384,1536,0.04354755414856804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,16384,1024,0.03767022159364488
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,16384,1536,0.018128888474570382
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,16384,768,0.02086755633354187
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,16384,512,0.008481777376598781
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,16384,1024,0.012889777620633444
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,16384,512,0.019718221492237516
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,16384,768,0.03654844562212626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,16384,256,0.006734222173690796
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,16384,256,0.017119111286269296
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,16384,512,0.03412355648146735
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,16384,128,0.004849777867396672
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,16384,256,0.030326220724317763
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,16384,64,0.004297777596447202
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,16384,128,0.015758223003811307
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,16384,128,0.0270506673389011
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,16384,32,0.004654222064548069
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,16384,64,0.01609777741962009
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,16384,32,0.01625688870747884
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,12288,16384,0.142048888736301
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,16384,65536,0.6802586449517144
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,12288,16384,0.14152088430192736
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,12288,16384,0.11919466654459636
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,12288,12288,0.0947164429558648
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,12288,12288,0.1103022230996026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,12288,12288,0.11204532782236735
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,12288,10240,0.08234400219387479
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,12288,10240,0.0935022234916687
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,16384,65536,0.5661199887593588
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,12288,10240,0.09869066874186198
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,12288,8192,0.08073244492212932
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,12288,8192,0.08533689048555161
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,12288,65536,0.5326622327168783
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,12288,8192,0.05853066841761271
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,12288,7168,0.0730373329586453
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,12288,7168,0.07777600155936347
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,12288,6144,0.061258667045169406
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,12288,7168,0.05705510907702976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,12288,6144,0.06899911165237427
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,12288,5120,0.05418577790260315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,12288,6144,0.04540888799561394
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,12288,5120,0.06157244576348198
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,12288,65536,0.4909502135382758
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,12288,4096,0.04642044504483541
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,12288,4096,0.053391112221611865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,12288,4096,0.032272001107533775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,12288,3584,0.04389422138532003
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,12288,5120,0.039164443810780845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,12288,3584,0.02892355455292596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,12288,3072,0.04865511258443197
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,12288,2560,0.03502310978041755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,12288,3072,0.03902933332655165
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,12288,3584,0.05133688780996534
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,12288,2560,0.044179555442598134
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,12288,2048,0.030368000268936157
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,12288,3072,0.02573777735233307
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,12288,1536,0.01383733335468504
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,12288,2048,0.041470223002963595
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,12288,1536,0.02590311070283254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,12288,2560,0.021039111746682063
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,12288,1024,0.010278222461541494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,12288,1536,0.03947110970815023
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,12288,2048,0.017530666457282174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,12288,1024,0.021762667430771723
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,12288,768,0.008781332936551835
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,12288,1024,0.035218665997187294
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,12288,768,0.02072000006834666
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,12288,512,0.007337777978844113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,12288,512,0.018720888429217868
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,12288,768,0.03510844376352098
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,12288,256,0.005342222336265776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,12288,128,0.004320000194840961
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,12288,512,0.03310755557484097
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,12288,256,0.017072889539930556
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,12288,128,0.015651555524932016
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,12288,64,0.0038613333470291565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,12288,256,0.028373334142896865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,12288,128,0.02568977740075853
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,12288,32,0.004056888735956616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,12288,64,0.016068443655967712
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,12288,32,0.01591822173860338
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,10240,16384,0.13482133547465006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,10240,16384,0.14013155301411948
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,10240,12288,0.08640978071424697
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,10240,16384,0.10784977674484253
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,10240,12288,0.10644355747434829
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,12288,65536,0.4164382351769342
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,10240,12288,0.10776888661914402
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,10240,10240,0.09132355451583862
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,10240,10240,0.07341777616076998
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,10240,10240,0.09408089187410142
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,10240,8192,0.07695200045903523
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,10240,8192,0.0662133362558153
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,10240,8192,0.08221422301398383
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,10240,7168,0.06950133376651339
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,10240,6144,0.05987822347217136
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,10240,7168,0.04885777831077576
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,10240,7168,0.07418400049209595
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,10240,65536,0.5150373246934679
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,10240,6144,0.0652737749947442
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,10240,6144,0.04696711235576206
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,10240,5120,0.05234044459131029
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,10240,4096,0.04484000139766269
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,10240,5120,0.04154577851295471
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,10240,5120,0.059122668372260205
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,10240,4096,0.05147022340032789
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,10240,4096,0.033997333712048
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,10240,3584,0.04142577780617608
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,10240,3584,0.04977955420811971
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,10240,3072,0.03745244608985053
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,10240,3072,0.04670755399598015
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,10240,3584,0.027134221461084154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,10240,2560,0.033004442850748696
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,10240,2560,0.042359110381868154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,10240,2048,0.028935111231274072
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,10240,2048,0.040151112609439425
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,10240,1536,0.012464000119103326
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,10240,2560,0.018982221682866413
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,10240,65536,0.4757848845587836
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,10240,1536,0.025232000483406916
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,10240,3072,0.023992889457278784
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,10240,2048,0.015863110621770222
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,10240,1024,0.009787555370065901
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,10240,1024,0.021553777986102637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,10240,1536,0.03852799865934584
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,10240,768,0.008485332959228093
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,10240,768,0.02017688916789161
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,10240,1024,0.035024000538720026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,10240,512,0.007251555720965068
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,10240,768,0.03460355599721273
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,10240,512,0.01796800063716041
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,10240,128,0.004279110994603899
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,10240,256,0.005116444494989184
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,10240,128,0.015583111180199517
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,10240,512,0.03121688961982727
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,10240,256,0.016744888491100736
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,10240,256,0.027664888236257765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,10240,64,0.0042168887125121225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,10240,128,0.025412445267041523
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,10240,64,0.015768888923856948
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,10240,32,0.004414222306675381
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,10240,32,0.015758223003811307
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,8192,16384,0.10134400261773004
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,8192,16384,0.11051733626259698
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,8192,16384,0.07277333074145846
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,8192,12288,0.07920889059702556
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,8192,12288,0.055233776569366455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,8192,12288,0.08914133575227524
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,8192,10240,0.06835022237565783
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,8192,10240,0.046682665745417275
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,8192,10240,0.08045866754319933
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,8192,8192,0.059177776177724205
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,8192,8192,0.07059199942482842
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,8192,8192,0.03840800126393636
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,8192,65536,0.35236355993482804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,8192,7168,0.05372000071737501
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,8192,7168,0.033634665939542986
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,8192,6144,0.04574666751755608
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,8192,7168,0.0653360022438897
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,8192,6144,0.058767113420698375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,8192,6144,0.029419554604424372
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,8192,5120,0.04056088791953193
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,8192,4096,0.03583555420239767
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,8192,5120,0.05370577838685778
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,8192,65536,0.3643671141730414
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,8192,5120,0.02590222160021464
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,10240,65536,0.3851946724785699
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,8192,3584,0.033709334002600774
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,8192,4096,0.022299556268586054
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,8192,3584,0.04571200079388089
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,8192,4096,0.047003553973303906
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,8192,3072,0.0309804446167416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,8192,3584,0.019122666782803006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,8192,3072,0.04272977842224968
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,8192,2560,0.027217777238951787
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,8192,3072,0.017118222183651395
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,8192,2560,0.03970311085383097
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,8192,2048,0.024324445260895625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,8192,1536,0.00997866690158844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,8192,2560,0.014895111322402954
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,8192,2048,0.03800088829464383
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,8192,1536,0.022082666556040447
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,8192,1536,0.036421332094404436
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,8192,1024,0.007997333175606197
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,8192,1024,0.019351111518012155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,8192,1024,0.03383466601371765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,8192,768,0.007076444725195567
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,8192,2048,0.012286221815480126
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,8192,768,0.017492444978819955
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,8192,512,0.006021333237489064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,8192,768,0.0311119986904992
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,8192,512,0.016428444120619033
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,8192,256,0.004307555655638377
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,8192,512,0.031088001198238794
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,8192,256,0.015526221858130561
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,8192,128,0.0038399998512532976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,8192,256,0.02681066592534383
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,8192,128,0.01440177857875824
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,8192,64,0.0035235554807715942
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,8192,128,0.02474755545457204
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,8192,32,0.0038728887836138406
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,8192,64,0.014544889330863953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,8192,32,0.014343111051453484
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,7168,16384,0.09517066346274482
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,7168,16384,0.1313191122478909
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,7168,12288,0.04959555466969808
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,7168,16384,0.0635511080423991
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,7168,12288,0.07538666990068224
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,7168,12288,0.10020533535215591
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,7168,10240,0.06507733133104113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,7168,10240,0.042324443658192955
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,7168,8192,0.05637511279847887
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,7168,10240,0.08803555700514053
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,7168,8192,0.03776444329155816
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,7168,8192,0.07683999670876397
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,7168,65536,0.3426204522450765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,8192,65536,0.29253511958652073
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,7168,7168,0.05128799875577291
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,7168,7168,0.06983200046751234
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,7168,7168,0.030939555830425684
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,7168,6144,0.0442826681666904
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,7168,6144,0.02734755476315816
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,7168,5120,0.03945866558286879
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,7168,4096,0.03472444415092468
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,7168,6144,0.06095288859473335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,7168,5120,0.05563822057512072
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,7168,4096,0.049026668071746826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,7168,3584,0.03197777933544583
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,7168,5120,0.023898666103680927
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,7168,65536,0.4285573429531521
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,7168,3584,0.04625599914126926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,7168,4096,0.02012888921631707
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,7168,3072,0.02964977754486932
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,7168,2560,0.026494221554862127
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,7168,3072,0.04267466730541653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,7168,3584,0.017872000734011333
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,7168,2560,0.03974755605061849
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,7168,2048,0.011394666300879585
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,7168,3072,0.015948444604873657
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,7168,2048,0.02398666739463806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,7168,1536,0.00959911114639706
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,7168,2560,0.014710222681363424
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,7168,2048,0.03795822130309211
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,7168,1536,0.021553777986102637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,7168,1024,0.00777688870827357
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,7168,1536,0.036453333165910505
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,7168,1024,0.019202666150199044
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,7168,768,0.0069004446268081665
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,7168,1024,0.03379199902216593
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,7168,768,0.017263111140992906
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,7168,512,0.005983110931184556
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,7168,768,0.03308088911904229
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,7168,512,0.016360888878504436
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,7168,256,0.004249777644872665
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,7168,512,0.02930755416552226
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,7168,256,0.015257777439223396
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,7168,128,0.0037946667936113146
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,7168,256,0.027573333846198186
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,7168,128,0.014302222265137566
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,7168,64,0.003439999909864532
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,7168,128,0.025829333398077223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,7168,64,0.01440711153878106
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,7168,32,0.0035822221802340615
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,7168,32,0.014297778407732645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,6144,16384,0.0886346697807312
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,7168,65536,0.26979467603895396
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,6144,16384,0.09858222140206231
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,6144,16384,0.06679733594258626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,6144,12288,0.07123644484413995
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,6144,12288,0.04615288972854614
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,6144,12288,0.08074222008387248
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,6144,10240,0.06151911285188463
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,6144,10240,0.07224266396628486
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,6144,10240,0.03927733169661628
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,6144,8192,0.05487377776039971
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,6144,8192,0.063573333952162
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,6144,8192,0.03224533465173509
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,6144,7168,0.028763555818133887
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,6144,7168,0.04881155490875244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,6144,7168,0.06070578098297119
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,6144,6144,0.04169599877463447
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,6144,5120,0.03834311167399088
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,6144,6144,0.025419554776615564
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,6144,6144,0.05425866776042514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,6144,5120,0.04964444372389051
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,6144,4096,0.03259199857711792
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,6144,4096,0.04460710949367947
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,6144,3584,0.0306640002462599
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,6144,5120,0.022316444251272414
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,6144,4096,0.01861422260602315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,6144,3584,0.043549332353803843
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,6144,3072,0.028120001157124836
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,6144,65536,0.3235235479142931
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,6144,3584,0.016743999388482835
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,6144,3072,0.0415013333161672
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,6144,65536,0.3164702256520589
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,6144,2560,0.012823111481136747
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,6144,3072,0.014707555373509726
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,6144,2560,0.025567111041810777
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,6144,2048,0.010952889091438718
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,6144,2048,0.023385778069496155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,6144,2560,0.03860177927547031
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,6144,1536,0.009273777405420939
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,6144,2048,0.03727822171317206
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,6144,1536,0.021066667305098638
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,6144,1024,0.007603555917739868
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,6144,1536,0.036333332459131874
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,6144,1024,0.01794577803876665
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,6144,768,0.006736889067623351
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,6144,1024,0.03380622135268317
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,6144,768,0.017304889029926725
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,6144,512,0.005493333356248007
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,6144,768,0.030994666947258845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,6144,512,0.01633066601223416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,6144,256,0.004226666771703296
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,6144,512,0.02873777680926853
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,6144,256,0.015188443991872998
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,6144,128,0.003765333443880081
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,6144,256,0.026400888959566753
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,6144,128,0.014289778139856128
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,6144,64,0.0034213333080212274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,6144,128,0.024681778417693243
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,6144,32,0.003516444315512975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,6144,64,0.01404088901148902
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,6144,32,0.014513777361975776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,5120,16384,0.08173333273993598
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,5120,16384,0.12199200524224176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,6144,65536,0.21609155337015787
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,5120,12288,0.045681777927610606
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,5120,12288,0.0644399987326728
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,5120,10240,0.05563111437691582
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,5120,12288,0.09562489059236313
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,5120,16384,0.05621866385142008
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,5120,10240,0.04182666540145874
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,5120,10240,0.08306488725874159
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,5120,8192,0.04998133248753018
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,5120,7168,0.044821331898371376
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,5120,8192,0.07302221986982557
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,5120,8192,0.034987555609809026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,5120,6144,0.03787022166781955
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,5120,7168,0.06874044736226399
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,5120,7168,0.027052443888452318
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,5120,6144,0.0593093302514818
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,5120,65536,0.2940888934665256
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,5120,5120,0.035382221142450966
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,5120,6144,0.02379200028048621
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,5120,5120,0.05383377936151293
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,5120,4096,0.030432889858881634
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,5120,4096,0.046741333272722035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,5120,3584,0.02868710954984029
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,5120,4096,0.017062221964200337
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,5120,5120,0.020709334148301017
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,5120,3584,0.04515999886724684
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,5120,3072,0.016368889146380954
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,5120,3584,0.015365334020720588
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,5120,3072,0.025977777110205755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,5120,3072,0.04216000106599596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,5120,2560,0.011959111524952782
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,5120,2560,0.024097777075237695
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,5120,65536,0.3981191052330865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,5120,2048,0.010466666685210334
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,5120,2560,0.03863199883037143
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,5120,2048,0.02233955595228407
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,5120,1536,0.008999110923873054
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,5120,1024,0.007432888779375289
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,5120,2048,0.037384000089433454
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,5120,1536,0.01963200006220076
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,5120,1536,0.03516088922818502
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,5120,1024,0.017808889349301655
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,5120,768,0.006724444528420766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,5120,1024,0.030253334177864924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,5120,768,0.017113778326246474
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,5120,512,0.005570666657553778
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,5120,768,0.03049422303835551
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,5120,512,0.016215110818545025
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,5120,256,0.0041982221106688184
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,5120,512,0.028928889168633357
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,5120,256,0.015117333994971381
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,5120,128,0.003914666672547658
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,5120,256,0.026571555270089045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,5120,128,0.014275555809338888
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,5120,64,0.003452444449067116
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,5120,128,0.025066667132907446
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,5120,32,0.0035386664999855887
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,5120,64,0.014111111561457316
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,5120,32,0.013929777675204806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,4096,16384,0.07546933492024739
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,4096,16384,0.0934257772233751
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,4096,16384,0.0447982218530443
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,4096,12288,0.059023108747270375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,4096,12288,0.0362426671716902
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,4096,10240,0.05264711048867968
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,4096,12288,0.0768773357073466
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,4096,8192,0.04474400149451362
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,4096,10240,0.06911822160085042
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,4096,10240,0.032179554303487144
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,4096,8192,0.027772444817754958
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,4096,7168,0.04063644342952304
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,4096,65536,0.259622229470147
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,4096,8192,0.060488886303371854
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,4096,7168,0.056700441572401256
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,5120,65536,0.18783556090460884
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,4096,6144,0.03594755464129978
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,4096,7168,0.025191111697090998
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,4096,5120,0.020625778370433383
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,4096,6144,0.05058844553099739
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,4096,5120,0.03293244375122918
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,4096,6144,0.02316355539692773
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,4096,5120,0.045443554719289146
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,4096,4096,0.028627554575602215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,4096,3584,0.01683466633160909
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,4096,4096,0.04117777612474229
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,4096,3584,0.027366222606764898
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,4096,3072,0.014426666829321118
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,4096,4096,0.019356444478034973
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,4096,3584,0.0398960014184316
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,4096,65536,0.28679466247558594
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,4096,3072,0.02536533276240031
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,4096,2560,0.01012711144155926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,4096,3072,0.03799022237459818
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,4096,2560,0.023541332946883306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,4096,2048,0.008547555241319869
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,4096,2560,0.03555555476082696
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,4096,2048,0.021703110800849065
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,4096,1536,0.007357333269384172
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,4096,2048,0.03481511274973551
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,4096,1536,0.01919555498494042
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,4096,1024,0.0060159998635451
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,4096,1536,0.03310222095913357
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,4096,1024,0.017681777477264404
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,4096,768,0.005016889009210798
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,4096,768,0.01696711116366916
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,4096,1024,0.029109332296583388
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,4096,512,0.004422222160630756
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,4096,768,0.029211554262373183
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,4096,256,0.01498311095767551
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,4096,512,0.015927111109097798
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,4096,256,0.0038622220357259116
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,4096,512,0.027879112296634253
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,4096,128,0.0035413333939181436
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,4096,256,0.025815111067559984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,4096,128,0.014198222094111972
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,4096,64,0.0032888888898822996
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,4096,128,0.02425333360830943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,4096,32,0.0034835553831524322
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,4096,64,0.013980444934633044
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,4096,32,0.013835555149449242
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3584,16384,0.07345689005321927
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,3584,16384,0.09239821963840061
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3584,16384,0.04054488738377889
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3584,12288,0.05811466773351034
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,4096,65536,0.14774400658077663
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3584,12288,0.03344444433848063
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,3584,12288,0.07602133353551228
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3584,10240,0.05130222108629015
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,3584,10240,0.06822755601671007
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3584,10240,0.029839999145931665
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3584,65536,0.2534524334801568
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3584,8192,0.043997334109412305
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3584,8192,0.02574577762020959
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3584,7168,0.03990044527583652
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,3584,8192,0.05947111050287882
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,3584,7168,0.05602755811479357
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3584,6144,0.0355697770913442
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3584,5120,0.03176000052028232
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,3584,6144,0.04983199967278374
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3584,7168,0.023736000061035156
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3584,4096,0.012130666938092975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,3584,5120,0.04502933224042257
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3584,6144,0.021338666478792827
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3584,4096,0.028536889288160536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3584,5120,0.014455111490355598
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3584,3584,0.010935111178292168
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,3584,4096,0.04077422287729051
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3584,3584,0.027154665854242113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3584,3072,0.009906666974226633
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,3584,3584,0.03935911258061727
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3584,3072,0.02514311174551646
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3584,2560,0.008850666383902231
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,3584,3072,0.03783822059631348
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3584,2560,0.023559111687872145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3584,2048,0.007831110722488826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3584,2048,0.021176000436147053
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,3584,2560,0.03552800085809495
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3584,1536,0.006788444601827198
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,3584,65536,0.2825351026323107
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,3584,2048,0.03456622362136841
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3584,1536,0.019139554765489366
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3584,1024,0.00543200017677413
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,3584,1536,0.0322035551071167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3584,1024,0.017533333765135873
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3584,768,0.00453955555955569
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,3584,1024,0.028693334923850164
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3584,768,0.01687111126052009
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3584,512,0.004026666697528628
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,3584,768,0.02912799848450555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3584,512,0.015983111328548856
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3584,256,0.0035706667436493766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,3584,512,0.02794577678044637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3584,256,0.014899555179807874
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3584,128,0.0032906666811969546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,3584,256,0.026398221651713055
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3584,65536,0.14407644006941053
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3584,128,0.014087999860445658
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3584,64,0.003125333330697484
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,3584,128,0.02514133354028066
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3584,32,0.0032115555885765287
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3584,64,0.013959111438857185
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3584,32,0.013768889009952545
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3072,12288,0.057293335596720375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3072,16384,0.07185777690675524
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,3072,16384,0.11045422156651814
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3072,12288,0.030989332331551447
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3072,10240,0.05051911208364698
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,3072,10240,0.07675288783179389
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,3072,12288,0.0870506697230869
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3072,16384,0.03932444585694207
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3072,8192,0.0433555543422699
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,3072,8192,0.0676106678114997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3072,10240,0.027481777800454035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3072,8192,0.02388711108101739
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3072,7168,0.039734221167034574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3072,7168,0.021716444028748408
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,3072,7168,0.06307555569542779
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3072,6144,0.03512710995144314
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3072,5120,0.016916443904240925
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3072,65536,0.24892711639404297
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3072,6144,0.019623999794324238
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,3072,6144,0.05402222275733948
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3072,5120,0.031543999910354614
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3072,4096,0.014277334014574686
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,3072,5120,0.049714667929543384
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3072,4096,0.02846755584081014
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3072,3584,0.026752889156341553
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3072,3584,0.01200977795653873
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,3072,65536,0.3599982261657715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,3072,4096,0.04447022080421448
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3072,3072,0.011006222003036074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,3072,3584,0.04314133193757799
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3072,3072,0.025008888708220586
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3072,2560,0.008542222281297049
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,3072,3072,0.04041688972049289
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3072,2560,0.02327288852797614
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3072,2048,0.007735110819339752
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3072,2048,0.0210817770825492
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,3072,2560,0.03741777936617533
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3072,1536,0.006725333217117522
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,3072,2048,0.036397334602144026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3072,1536,0.019117333822780185
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3072,1024,0.005665777872006099
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,3072,1536,0.03213511241806878
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3072,1024,0.017439999514155917
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3072,768,0.0047120000753137796
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,3072,1024,0.029598222838507757
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3072,768,0.016752888758977253
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3072,512,0.003999999827808804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,3072,768,0.02955555584695604
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3072,512,0.015862221519152325
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3072,256,0.003546666767862108
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,3072,512,0.028008000718222723
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3072,256,0.014880888991885714
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3072,128,0.0033164444482988785
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3072,65536,0.11703466044531928
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,3072,256,0.026157332791222468
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3072,128,0.014099554883109199
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3072,64,0.0031439999325407874
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,3072,128,0.024460444847742718
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3072,64,0.013897778259383308
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,3072,32,0.003185777821474605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,3072,32,0.013749333719412485
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2560,16384,0.0716480016708374
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2560,12288,0.02626488937271966
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2560,12288,0.05695911248524984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,2560,16384,0.07761955261230469
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2560,16384,0.03171022070778741
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2560,10240,0.050275554259618126
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,2560,12288,0.06504533025953504
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,2560,10240,0.058409776952531606
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2560,8192,0.04302488764127096
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2560,10240,0.023413333627912734
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2560,7168,0.03947822252909342
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2560,8192,0.01978400018480089
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,2560,8192,0.05360355642106798
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2560,6144,0.01641688909795549
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2560,6144,0.03491910960939195
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,2560,7168,0.05025777882999844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2560,7168,0.018400889303949144
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2560,5120,0.015086222026083203
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,2560,6144,0.044416887892617114
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2560,5120,0.03143999973932902
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2560,4096,0.013201778133710226
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,2560,5120,0.04138222336769104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2560,4096,0.02849422229660882
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2560,3584,0.011512888802422417
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,2560,4096,0.03780622283617655
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2560,3584,0.02685333291689555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2560,3072,0.010716444088353051
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2560,3072,0.024852444728215534
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,2560,3584,0.03750400079621209
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2560,65536,0.2482924461364746
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,2560,65536,0.2545253435770671
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2560,2560,0.008388444781303406
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,2560,3072,0.036143110858069524
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2560,2560,0.022081777453422546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2560,2048,0.007517333659860823
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,2560,2560,0.03369244601991441
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2560,2048,0.020435555113686454
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2560,1536,0.006578666468461354
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,2560,2048,0.033174223370022245
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2560,1536,0.018886221779717337
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2560,1024,0.005154666801293691
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,2560,1536,0.0306915541489919
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2560,1024,0.01718311177359687
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2560,768,0.004471999903519948
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,2560,1024,0.02881866693496704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2560,512,0.004002666721741359
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2560,768,0.016364443633291457
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,2560,768,0.028836445675955877
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2560,256,0.014632888966136508
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2560,512,0.015579554769727917
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2560,256,0.0037333332002162933
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,2560,512,0.027501333090994094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2560,128,0.0033573332346147964
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,2560,128,0.024808888634045918
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2560,65536,0.10031111372841729
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,2560,256,0.025807110799683466
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2560,128,0.013902222116788229
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2560,64,0.0030542222989930045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2560,32,0.0031146665828095544
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2560,64,0.013863999810483722
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2560,32,0.013533333109484779
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2048,12288,0.05682488944795397
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2048,16384,0.0714284446504381
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,2048,12288,0.06436622142791748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,2048,16384,0.07755466964509752
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2048,16384,0.02690222197108799
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2048,12288,0.022180444664425317
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2048,10240,0.04941599898868137
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,2048,10240,0.05785955323113335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2048,10240,0.019677332705921598
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2048,8192,0.042585776911841504
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,2048,8192,0.05202310946252611
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2048,7168,0.015279110934999255
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2048,8192,0.01676266723208957
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2048,7168,0.0392506652408176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2048,6144,0.013879111243618859
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,2048,7168,0.04983199967278374
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2048,6144,0.03479377759827508
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2048,5120,0.013914666242069669
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,2048,6144,0.04430133435461256
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2048,65536,0.24793688456217447
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2048,5120,0.03148355417781406
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2048,4096,0.01165511127975252
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,2048,5120,0.04112622141838074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2048,4096,0.02821777926550971
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2048,3584,0.010204444328943888
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2048,3584,0.026001777913835313
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,2048,4096,0.0373288889726003
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,2048,65536,0.2541484567854139
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2048,3072,0.009218666288587783
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,2048,3584,0.03704444567362467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2048,3072,0.02482844392458598
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2048,2560,0.007905777957704332
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,2048,3072,0.035572446054882474
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2048,2560,0.021892444954978094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2048,2048,0.00701422244310379
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,2048,2560,0.03233777814441257
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2048,2048,0.020242666204770405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2048,1536,0.005934222290913264
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,2048,2048,0.031974222924974226
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2048,1536,0.01872977779971229
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2048,1024,0.017101332545280457
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2048,1024,0.004947555561860402
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,2048,1536,0.030369778474171955
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2048,768,0.004408888932731417
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,2048,1024,0.02831911047299703
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2048,768,0.016357334123717416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2048,512,0.0038577777643998465
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,2048,768,0.028558221128251817
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2048,512,0.015594666202863058
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2048,256,0.0034186666210492453
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,2048,512,0.02734044525358412
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,2048,256,0.025586666332350835
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2048,256,0.014607111612955729
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2048,128,0.0031893334041039148
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2048,128,0.013751111096805997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2048,64,0.002987555538614591
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,2048,128,0.02402399977048238
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2048,64,0.013787555197874704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2048,32,0.003142222141226133
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,2048,32,0.013535999589496188
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,2048,65536,0.08027733034557767
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1536,16384,0.07173511054780748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1536,12288,0.056087997224595815
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,1536,16384,0.07440355751249525
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1536,10240,0.015849777393870883
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,1536,12288,0.06159733401404487
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1536,16384,0.022778666681713525
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1536,12288,0.018426666657129925
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1536,10240,0.04900000161594815
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1536,8192,0.014375999569892883
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,1536,10240,0.05583022038141886
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1536,8192,0.04222755630811056
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1536,7168,0.01328000012371275
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,1536,8192,0.050082667006386646
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1536,7168,0.03917600048912896
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1536,6144,0.015069334043396844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,1536,7168,0.04857866631613838
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1536,6144,0.03489689032236735
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1536,5120,0.013321778012646569
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,1536,6144,0.04402222235997518
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1536,5120,0.031253334548738264
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1536,4096,0.011275555524561139
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1536,65536,0.24708710776435006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,1536,5120,0.04067644476890564
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1536,4096,0.02823466724819607
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1536,3584,0.010405333505736457
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,1536,4096,0.037124445041020714
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1536,3584,0.02645333276854621
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1536,3072,0.009359111388524374
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,1536,3584,0.03637689020898607
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,1536,65536,0.2278533246782091
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1536,3072,0.02367111047108968
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1536,2560,0.007676444119877285
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,1536,2560,0.03146044413248698
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,1536,3072,0.035261332988739014
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1536,2560,0.02186577849917942
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1536,2048,0.006539555473460092
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1536,2048,0.020201777418454487
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1536,1536,0.005695111221737332
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,1536,2048,0.030576000610987347
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1536,1536,0.01851288808716668
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1536,1024,0.004471999903519948
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1536,65536,0.06660088565614489
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,1536,1536,0.03022222386466132
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1536,1024,0.01699733402993944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1536,768,0.004173333446184794
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,1536,1024,0.028209778997633193
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1536,768,0.016184000505341422
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,1536,512,0.02713955607679155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1536,512,0.0037671112351947357
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,1536,256,0.02572533322705163
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,1536,768,0.028582221931881372
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1536,512,0.015692444311247934
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1536,256,0.003375111147761345
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1536,256,0.014358222484588623
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1536,128,0.0030586665703190696
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1536,128,0.013819555441538492
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1536,64,0.002934222213096089
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,1536,128,0.023881778120994568
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1536,64,0.013808888693650564
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1536,32,0.0030088888274298776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1536,32,0.013403555585278405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1024,16384,0.07077866792678833
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1024,12288,0.014675555957688225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,1024,16384,0.09747999906539917
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1024,12288,0.0550782216919793
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1024,16384,0.016569778323173523
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1024,10240,0.014891554911931356
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,1024,12288,0.07776800129148695
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1024,10240,0.04859200119972229
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1024,8192,0.014098667436175875
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1024,8192,0.04212444358401828
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,1024,10240,0.0701084468099806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1024,65536,0.24876801172892252
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1024,7168,0.012845333251688214
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,1024,8192,0.06111466884613037
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1024,7168,0.0388026667965783
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1024,6144,0.012438221938080259
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,1024,7168,0.05763022104899088
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1024,6144,0.034259554412629865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1024,5120,0.011041777829329172
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,1024,6144,0.050761779149373375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1024,5120,0.030305776331159804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1024,4096,0.009569777382744683
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,1024,65536,0.3161626656850179
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,1024,5120,0.046784000264273755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1024,4096,0.02663822140958574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1024,65536,0.04574133290184868
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1024,3584,0.008928000099129146
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,1024,4096,0.0418488879998525
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1024,3584,0.02500711050298479
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1024,3072,0.008162666526105668
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,1024,3584,0.03934489025010003
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1024,3072,0.0229759994480345
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1024,2560,0.007119999991522894
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1024,2560,0.02162577708562215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,1024,3072,0.036778665251202054
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1024,2048,0.00638755535085996
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,1024,2560,0.03340177734692892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1024,2048,0.020075556304719713
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1024,1536,0.005240889059172736
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,1024,2048,0.03282489048110114
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1024,1536,0.018532445033391316
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1024,1024,0.00479288895924886
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,1024,1536,0.03142222099834018
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1024,768,0.0042071110672420925
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1024,1024,0.016903999778959487
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,1024,1024,0.02937244375546773
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1024,768,0.016167999969588388
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1024,512,0.003777777983082665
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,1024,768,0.028982222080230713
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1024,512,0.015413333972295126
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1024,256,0.0033173333439562055
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,1024,512,0.027480888697836135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1024,256,0.014456889695591398
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1024,128,0.0029599999801980127
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,1024,256,0.02610844373703003
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1024,128,0.013600889179441663
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1024,64,0.0028133332315418455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,1024,128,0.02455288834042019
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,1024,32,0.002926222152180142
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1024,64,0.013736888766288757
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,1024,32,0.013448889056841532
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,768,16384,0.01423555612564087
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,768,16384,0.07083733214272393
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,768,12288,0.016837333639462788
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,768,16384,0.09106044636832343
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,768,12288,0.055471999777687915
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,768,10240,0.014519999424616495
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,768,10240,0.048432889911863536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,768,12288,0.07343111435572307
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,768,8192,0.012917333179050021
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,768,8192,0.041542222102483116
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,768,10240,0.06406044297748141
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,768,7168,0.012246222131782107
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,768,8192,0.05916621949937609
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,768,65536,0.24750044610765246
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,768,7168,0.03828533490498861
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,768,6144,0.011378666592968835
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,768,6144,0.033803555700514056
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,768,7168,0.05615022447374132
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,768,5120,0.010089777410030365
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,768,6144,0.04871911141607496
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,768,65536,0.039055999782350324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,768,65536,0.2872382269965278
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,768,5120,0.029919111066394385
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,768,4096,0.009242666264375051
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,768,4096,0.026574222577942744
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,768,5120,0.0454364452097151
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,768,3584,0.008776889079146916
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,768,4096,0.038400888442993164
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,768,3584,0.025144888295067683
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,768,3072,0.008309333688682979
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,768,3584,0.037250667810440063
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,768,3072,0.023388443721665278
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,768,2560,0.006773333168692059
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,768,3072,0.03526488939921061
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,768,2560,0.021669333179791767
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,768,2048,0.0057884446448749965
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,768,2560,0.032778667079077825
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,768,2048,0.020000000794728596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,768,1536,0.005098666581842634
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,768,2048,0.03212622139188979
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,768,1536,0.01811911165714264
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,768,1024,0.004300444490379757
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,768,1536,0.031593776411480375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,768,1024,0.017083555459976196
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,768,768,0.003972444269392225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,768,1024,0.028598222467634413
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,768,768,0.016201777590645682
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,768,512,0.003617777592606015
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,768,768,0.02861600120862325
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,768,512,0.015423110789722867
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,768,256,0.003237333355678452
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,768,512,0.0273991111252043
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,768,256,0.014469333820872836
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,768,128,0.0029537777105967202
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,768,256,0.025621333056026038
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,768,128,0.013616888887352414
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,768,64,0.002797333316670524
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,768,128,0.02478933334350586
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,768,64,0.013741333451535968
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,768,32,0.00286311118139161
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,768,32,0.01349955548842748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,512,16384,0.013880000346236758
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,512,16384,0.06966311401791044
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,512,12288,0.011720889144473605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,512,12288,0.05458577805095249
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,512,16384,0.06907377640406291
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,512,10240,0.010622222390439777
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,512,12288,0.057651552889082164
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,512,10240,0.047896001074049205
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,512,65536,0.24867110782199434
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,512,8192,0.010330666270520952
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,512,10240,0.05295822355482313
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,512,8192,0.04093688726425171
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,512,7168,0.009736888938479954
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,512,65536,0.21230755911933052
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,512,8192,0.04749777913093567
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,512,7168,0.03741511040263706
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,512,6144,0.00906133320596483
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,512,65536,0.030579557021458943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,512,7168,0.046237332953347095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,512,5120,0.037167112032572426
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,512,6144,0.033379554748535156
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,512,5120,0.008584000170230865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,512,4096,0.03363822235001458
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,512,6144,0.0415653321478102
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,512,5120,0.029807110627492268
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,512,4096,0.008228444390826756
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,512,4096,0.026289777623282537
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,512,3584,0.008290666672918532
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,512,3584,0.024534222152498033
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,512,3072,0.0073902226156658595
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,512,3584,0.033258666594823204
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,512,3072,0.022845332821210224
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,512,2560,0.006360000206364526
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,512,3072,0.03201244605912103
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,512,2560,0.021328000558747187
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,512,2048,0.005589333259397083
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,512,2560,0.030284444491068523
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,512,2048,0.019630221856964957
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,512,1536,0.004933333231343163
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,512,2048,0.03013777732849121
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,512,1536,0.0181048893266254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,512,1024,0.027847111225128174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,512,1024,0.004237333519591225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,512,1536,0.029351999362309773
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,512,1024,0.016417778200573392
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,512,768,0.003850666599141227
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,512,768,0.015822221835454304
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,512,512,0.0035235554807715942
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,512,768,0.027979556057188246
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,512,512,0.01496266656451755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,512,256,0.003134222287270758
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,512,512,0.026680888401137456
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,512,256,0.014203555054134794
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,512,128,0.0029297777348094513
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,512,256,0.02572000026702881
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,512,128,0.01367555558681488
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,512,64,0.0027457777824666765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,512,128,0.023583999938435022
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,512,32,0.0027795556104845474
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,512,64,0.013535999589496188
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,512,32,0.013340444200568728
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,256,16384,0.01092977739042706
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,256,16384,0.06944533189137776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,256,12288,0.00960799968904919
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,256,65536,0.2473448912302653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,256,16384,0.06823377476798163
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,256,65536,0.21126666333940294
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,256,12288,0.05376088950369093
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,256,65536,0.028510222832361858
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,256,10240,0.008914666871229807
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,256,12288,0.05704889032575819
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,256,10240,0.04656444324387444
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,256,8192,0.007792889244026608
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,256,10240,0.05244444476233589
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,256,8192,0.03978666663169861
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,256,7168,0.007592000067234039
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,256,8192,0.045575112104415894
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,256,7168,0.03648799988958571
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,256,6144,0.0070186663005087115
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,256,7168,0.045174221197764076
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,256,6144,0.032580445210138954
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,256,5120,0.00794311116139094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,256,6144,0.03924177752600776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,256,5120,0.029155555698606703
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,256,4096,0.007258666886223688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,256,5120,0.03650755683581034
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,256,3584,0.007303111255168915
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,256,4096,0.02606755495071411
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,256,3584,0.032920890384250216
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,256,3584,0.024263999528355066
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,256,4096,0.03339644604259067
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,256,3072,0.006912888752089606
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,256,3072,0.022709333234363135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,256,2560,0.006222222414281633
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,256,3072,0.03171822097566392
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,256,2048,0.005782222168313132
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,256,2560,0.021006221572558086
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,256,2560,0.029839999145931665
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,256,2048,0.01956355571746826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,256,1024,0.0041457778877682155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,256,1536,0.004836444639497333
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,256,1536,0.029409776131312054
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,256,2048,0.02968177861637539
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,256,1536,0.017910222212473553
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,256,1024,0.016208888755904306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,256,768,0.0038115556041399636
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,256,1024,0.027317333552572463
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,256,768,0.015652444627549913
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,256,512,0.0034479999707804788
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,256,512,0.0148026661740409
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,256,768,0.028006222512986925
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,256,256,0.0031351111829280853
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,256,512,0.026681777503755357
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,256,256,0.014136888914638095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,256,128,0.0028782222006056043
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,256,256,0.025031111306614343
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,256,128,0.01346933344999949
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,256,64,0.0027280000762807
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,256,128,0.024381332927280005
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,256,32,0.0027831111931138565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,256,64,0.013471999930010902
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,256,32,0.01329955541425281
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,128,65536,0.020954666866196528
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,128,16384,0.008180444439252218
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,128,65536,0.250270234213935
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,128,16384,0.06874044736226399
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,128,12288,0.0074728892909155945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,128,65536,0.21070755852593315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,128,12288,0.05722577704323662
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,128,16384,0.06816711028416951
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,128,12288,0.053411556614769824
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,128,10240,0.007303111255168915
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,128,8192,0.007252444823582967
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,128,10240,0.045457777049806386
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,128,10240,0.052001777622434825
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,128,7168,0.0069679998689227635
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,128,8192,0.03903466794225905
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,128,6144,0.03254666593339708
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,128,8192,0.04447999927732679
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,128,7168,0.036244445376926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,128,6144,0.00673333348499404
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,128,7168,0.043268445465299815
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,128,5120,0.007116444408893585
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,128,6144,0.039150221480263606
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,128,5120,0.02904711167017619
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,128,4096,0.006695111178689533
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,128,5120,0.036291556225882635
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,128,4096,0.0258568889564938
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,128,3584,0.007107555866241455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,128,4096,0.03312266535229153
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,128,3584,0.02420444455411699
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,128,3072,0.006782222125265334
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,128,3072,0.02260533306333754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,128,3584,0.032912890116373696
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,128,2560,0.006117333140638139
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,128,3072,0.03166666626930237
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,128,2560,0.020976889464590285
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,128,2048,0.005408888889683618
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,128,2560,0.029898666673236426
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,128,2048,0.019307555423842538
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,128,1536,0.004791111167934206
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,128,2048,0.029325332906511094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,128,1536,0.017768000562985737
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,128,1024,0.004057777838574516
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,128,1536,0.02910311023394267
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,128,1024,0.01605955594115787
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,128,768,0.003682666768630346
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,128,1024,0.027432888746261597
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,128,768,0.015419556034935845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,128,512,0.0033608888172441055
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,128,768,0.02771200074089898
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,128,512,0.014704888065656027
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,128,256,0.0030008889734745026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,128,512,0.026678222748968337
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,128,256,0.013953778478834363
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,128,128,0.0028213332924577924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,128,256,0.02492444382773505
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,128,128,0.013535999589496188
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,128,64,0.002667555585503578
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,160,128,128,0.02342133389578925
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,128,32,0.0026942222482628296
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,128,64,0.013341333303186627
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,128,32,0.013205333716339536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,64,65536,0.013973333769374423
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,64,16384,0.007175111108356052
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,64,12288,0.006999111009968652
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,64,16384,0.06754044691721599
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,64,65536,0.2495804362826877
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,64,10240,0.007074666519959767
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,64,12288,0.05383200115627713
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,64,8192,0.006847111301289664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,64,10240,0.045647998650868736
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,64,7168,0.006741333338949416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,64,5120,0.00692622239391009
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,64,8192,0.039472887913386025
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,64,7168,0.03600710961553786
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,64,6144,0.006535111202134027
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,64,6144,0.03228088882234361
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,64,4096,0.006609777609507243
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,64,5120,0.029100444581773546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,64,4096,0.025659556190172832
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,64,3072,0.022608000371191237
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,64,3584,0.007048889166778988
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,64,3584,0.024216888679398432
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,64,3072,0.006708444406588872
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,64,2560,0.006057777752478917
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,64,2048,0.005385777602593104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,64,2560,0.02091288897726271
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,64,2048,0.019298666053348117
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,64,1536,0.0047288888858424295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,64,1536,0.017629333668284945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,64,1024,0.004019555532270008
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,64,1024,0.016031111280123394
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,64,768,0.003624000069167879
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,64,512,0.003346666693687439
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,64,768,0.01536177761024899
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,64,256,0.002982222164670626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,64,512,0.014680888917711047
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,64,128,0.0027760000278552375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,64,256,0.013999111122555204
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,64,64,0.00267999991774559
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,64,128,0.013447110851605734
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,64,64,0.013276444541083442
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,64,32,0.0026488889836602737
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,64,32,0.013000888956917657
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,32,65536,0.013191111385822296
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,32,16384,0.00776977754301495
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,32,16384,0.06735289096832275
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,32,10240,0.0069057775868309875
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,32,12288,0.05282666616969638
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,32,65536,0.24973066647847494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,32,12288,0.0070471109615431885
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,32,8192,0.006810666786299811
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,32,10240,0.04583200150065952
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,32,7168,0.006551110910044775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,32,8192,0.03901066713862949
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,32,7168,0.03603911068704393
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,32,6144,0.006452444526884291
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,32,5120,0.006866666591829724
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,32,6144,0.03242666522661845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,32,4096,0.006534222099516127
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,32,5120,0.029073778125974867
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,32,2560,0.006056889063782162
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,32,3584,0.0069875551594628235
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,32,2560,0.020995555652512446
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,32,4096,0.02567377852069007
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,32,3584,0.02418400016095903
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,32,3072,0.006679111056857639
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,32,3072,0.02253155575858222
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,32,2048,0.005384888913896348
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,32,2048,0.019277332557572257
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,32,1536,0.0046951112647851305
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,32,1536,0.017645332548353408
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,32,1024,0.004004444513056013
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,32,256,0.0029973333908451926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,32,1024,0.01611555616060893
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,32,768,0.003657777690225177
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,32,768,0.015389333168665567
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,32,512,0.003310222178697586
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,32,512,0.014675555957688225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,32,256,0.013954665925767688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,32,128,0.002806222273243798
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,32,128,0.013384000294738345
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,32,64,0.002611555573013094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,160,32,32,0.0026293332791990707
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,32,64,0.013260444833172692
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,160,32,32,0.01312711089849472
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,65536,12288,0.23177689976162383
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,65536,10240,0.1987137794494629
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,65536,16384,0.3079964319864909
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,65536,12288,0.4032951196034749
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,65536,8192,0.16283110777537027
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,65536,10240,0.3401946756574843
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,65536,8192,0.271762662463718
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,65536,16384,0.5367751121520996
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,65536,7168,0.15077688958909777
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,65536,7168,0.2373510996500651
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,65536,6144,0.12486578358544244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,65536,6144,0.20590221881866455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,65536,5120,0.11331466833750407
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,65536,5120,0.17453955279456246
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,65536,12288,0.2673368983798557
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,65536,16384,0.3299520015716553
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,65536,4096,0.14164711369408503
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,65536,4096,0.08771377801895142
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,65536,10240,0.22959287961324057
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,65536,3584,0.07944622304704455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,65536,3584,0.12482310665978326
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,65536,8192,0.18889333142174614
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,65536,3072,0.07513244284523858
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,65536,3072,0.10942400164074367
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,65536,7168,0.16999377144707573
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,65536,2560,0.061787552303738065
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,65536,2560,0.09275466865963405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,65536,2048,0.07661244604322645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,65536,2048,0.05372800098525154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,65536,6144,0.15079555246565077
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,65536,1536,0.04502755403518677
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,65536,1536,0.06012800004747179
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,65536,1024,0.04327733318010966
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,65536,1024,0.03677777780426873
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,65536,5120,0.1277057727177938
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,65536,768,0.03357155455483331
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,65536,768,0.03260533346070184
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,65536,3584,0.10567377673255073
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,65536,4096,0.11129600471920437
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,65536,512,0.02514133354028066
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,65536,512,0.028296000427669946
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,65536,256,0.015389333168665567
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,65536,256,0.024455111887719896
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,65536,3072,0.09171466694937812
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,65536,128,0.010496888723638324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,65536,128,0.021048888564109802
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,65536,2560,0.08634666601816814
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,65536,64,0.008619555168681675
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,65536,64,0.022470222579108343
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,65536,32,0.00811644477976693
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,65536,32,0.02305511136849721
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,65536,1536,0.06369511286417644
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,65536,2048,0.07218222485648261
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,65536,256,0.037631998459498085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,65536,768,0.04936800069279141
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,65536,1024,0.052741332186592944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,65536,512,0.04421600037150913
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,16384,16384,0.10055910878711277
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,65536,128,0.0326204432381524
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,16384,16384,0.1319813330968221
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,16384,12288,0.07985155449973212
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,16384,12288,0.10170666376749675
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,16384,65536,0.36115466223822695
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,16384,10240,0.07050489054785834
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,16384,10240,0.08575822247399224
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,16384,8192,0.05755644374423557
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,16384,8192,0.07019288672341241
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,16384,7168,0.05491377909978231
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,16384,7168,0.06120444668663872
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,16384,65536,0.5522933536105686
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,16384,6144,0.05334044496218363
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,16384,6144,0.04923466510242886
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,16384,5120,0.04576889011594984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,16384,5120,0.04158488909403483
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,16384,16384,0.12766844696468777
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,16384,12288,0.10111288891898261
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,16384,4096,0.035742223262786865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,16384,4096,0.03848000036345588
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,16384,10240,0.09193066755930583
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,16384,3584,0.03369333346684774
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,16384,3584,0.033350222640567355
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,16384,8192,0.07842044697867499
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,16384,3072,0.032476445039113365
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,16384,3072,0.030815998713175457
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,16384,2560,0.025927111506462097
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,16384,2560,0.028251555230882432
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,16384,7168,0.06885155704286364
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,16384,65536,0.4018835491604275
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,16384,2048,0.02163377735349867
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,16384,2048,0.026195555925369263
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,16384,1536,0.017608000172509086
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,16384,1536,0.02257866660753886
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,16384,6144,0.062289781040615506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,16384,1024,0.013433777623706393
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,16384,1024,0.01984177695380317
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,16384,5120,0.056589331891801625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,16384,768,0.010121777653694153
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,16384,768,0.018890667292806838
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,16384,4096,0.05081599950790405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,16384,512,0.008052444292439355
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,16384,3584,0.04929777648713854
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,16384,512,0.017934223016103108
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,16384,256,0.006100444330109491
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,16384,256,0.015124445160230001
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,16384,3072,0.04567822151713901
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,16384,128,0.004487110922733943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,16384,128,0.01422133379512363
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,16384,2560,0.043756445248921715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,16384,64,0.004129777765936322
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,16384,64,0.01461155547036065
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,16384,32,0.004552000098758274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,16384,32,0.014583999911944071
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,16384,2048,0.04064266549216376
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,16384,1536,0.038042667839262224
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,16384,1024,0.0336497790283627
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,16384,768,0.03386933273739285
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,12288,16384,0.09117511245939468
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,12288,16384,0.1004142231411404
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,16384,512,0.03245066603024801
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,12288,65536,0.31725777520073783
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,16384,256,0.027929776244693335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,16384,128,0.02528800070285797
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,12288,12288,0.07684355311923556
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,12288,10240,0.06401511033376057
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,12288,12288,0.07294311126073201
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,12288,10240,0.0657786660724216
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,12288,8192,0.052412443690829806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,12288,8192,0.05288444293869866
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,12288,7168,0.04735555582576328
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,12288,7168,0.04715111189418369
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,12288,65536,0.4143599934048123
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,12288,6144,0.04131644301944309
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,12288,6144,0.044715556833479136
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,12288,5120,0.03548444310824076
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,12288,5120,0.03775733378198411
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,12288,16384,0.10947910944620769
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,12288,4096,0.029094222519132826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,12288,4096,0.03328355484538608
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,12288,12288,0.09463200304243301
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,12288,3584,0.026287110315428838
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,12288,3584,0.031195554468366835
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,12288,10240,0.08484711249669392
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,12288,3072,0.023226666781637404
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,12288,3072,0.028577778074476454
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,12288,8192,0.06853866577148438
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,12288,65536,0.3695555528004964
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,12288,2560,0.020205333828926086
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,12288,2560,0.026354667213228013
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,12288,2048,0.01683111157682207
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,12288,2048,0.023947555157873366
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,12288,7168,0.06404978036880493
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,12288,1536,0.013334222137928009
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,12288,1536,0.021363554729355708
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,12288,6144,0.05773955583572388
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,12288,1024,0.009922666682137383
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,12288,1024,0.019508444600635104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,12288,5120,0.05300977826118469
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,12288,768,0.008384000096056197
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,12288,768,0.018357333209779527
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,12288,3584,0.04684177703327603
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,12288,512,0.00692622239391009
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,12288,512,0.016728889611032274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,12288,4096,0.048473778698179454
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,12288,256,0.004838222016890843
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,12288,256,0.015102222561836243
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,12288,2560,0.041947556866539844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,12288,128,0.004055110944641961
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,12288,128,0.014208000567224292
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,12288,3072,0.043319112724728055
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,12288,64,0.0036275556517971885
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,12288,64,0.014334221680959066
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,12288,32,0.003975999852021535
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,12288,1536,0.03632266653908624
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,12288,32,0.01440177857875824
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,12288,768,0.033352000845803156
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,12288,2048,0.03915289044380188
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,12288,1024,0.03300889001952277
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,10240,16384,0.08817244238323635
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,10240,16384,0.08700711197323269
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,10240,65536,0.30135377248128253
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,10240,12288,0.07344533337487115
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,10240,12288,0.0664880010816786
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,12288,256,0.025810667210155066
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,12288,128,0.024638222323523626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,10240,10240,0.060602664947509766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,10240,10240,0.056602663464016385
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,10240,65536,0.357826656765408
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,12288,512,0.031389332479900785
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,10240,8192,0.05041955577002632
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,10240,8192,0.04775644342104594
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,10240,7168,0.04589510957400004
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,10240,7168,0.04054666558901469
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,10240,6144,0.043306665288077466
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,10240,6144,0.035498665438758
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,10240,5120,0.030653334326214258
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,10240,5120,0.03703377644220988
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,10240,16384,0.10542044374677871
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,10240,4096,0.02569155560599433
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,10240,4096,0.03229866756333245
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,10240,65536,0.3426471021440294
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,10240,12288,0.08575110965304905
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,10240,3584,0.023028444912698533
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,10240,3584,0.030003554291195337
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,10240,3072,0.0203484445810318
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,10240,3072,0.0273991111252043
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,10240,10240,0.07730666796366374
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,10240,2560,0.017301332619455125
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,10240,2560,0.025556445121765137
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,10240,7168,0.06117866436640421
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,10240,2048,0.014259555273585849
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,10240,2048,0.02253244486120012
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,10240,8192,0.06786933210160997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,10240,1536,0.011301333705584208
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,10240,1536,0.020768000019921195
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,10240,6144,0.05590133534537422
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,10240,1024,0.008755555583371056
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,10240,1024,0.01901955571439531
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,10240,5120,0.05125688844256931
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,10240,768,0.007458666960398356
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,10240,768,0.017861333158281114
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,10240,4096,0.04663733310169644
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,10240,3584,0.045046223534478076
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,10240,512,0.006268444574541516
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,10240,512,0.015697777271270752
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,10240,256,0.004223111189073986
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,10240,256,0.014892444014549255
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,10240,2560,0.03993511199951172
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,10240,128,0.00378577783703804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,10240,128,0.014162666267818876
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,10240,768,0.03315644462903341
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,10240,64,0.003590222034189436
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,10240,64,0.014271999398867289
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,10240,32,0.0037173334923055437
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,10240,32,0.014260444376203748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,10240,1536,0.0355386667781406
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,10240,3072,0.04113599989149306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,10240,2048,0.03686044282383389
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,10240,1024,0.03257688879966736
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,8192,16384,0.0693111088540819
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,8192,65536,0.2192364401287503
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,8192,16384,0.06745155652364095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,8192,12288,0.05508888761202494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,8192,12288,0.053279111782709755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,8192,10240,0.05021244287490845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,8192,10240,0.04903644323348999
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,10240,128,0.02444355520937178
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,10240,256,0.026151999831199646
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,8192,8192,0.03692355420854356
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,8192,8192,0.040680888626310564
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,8192,65536,0.286131567425198
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,8192,7168,0.03706399930848016
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,8192,7168,0.03275555703375075
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,8192,6144,0.02880355715751648
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,8192,6144,0.033823112646738686
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,10240,512,0.02937955657641093
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,8192,5120,0.025026667449209426
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,8192,5120,0.030584888325797185
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,8192,16384,0.08930399682786729
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,8192,4096,0.021151999632517498
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,8192,65536,0.27853600184122723
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,8192,4096,0.02700977855258518
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,8192,3584,0.01869155632125007
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,8192,3584,0.02541333271397485
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,8192,12288,0.07440088854895698
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,8192,3072,0.01661600006951226
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,8192,3072,0.023382221659024555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,8192,10240,0.06689866383870442
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,8192,2560,0.014418666561444601
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,8192,2560,0.022403554783927068
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,8192,8192,0.057760000228881836
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,8192,2048,0.01202311118443807
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,8192,2048,0.02072266737620036
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,8192,7168,0.05508977837032742
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,8192,1536,0.00965244405799442
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,8192,1536,0.018890667292806838
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,8192,6144,0.049201776583989464
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,8192,1024,0.007745777567227681
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,8192,1024,0.017807111144065857
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,8192,5120,0.045440001620186694
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,8192,768,0.006678222368160884
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,8192,768,0.015543111496501498
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,8192,4096,0.042450666427612305
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,8192,512,0.0058222222659322954
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,8192,512,0.014977777997652689
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,8192,3584,0.0403511126836141
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,8192,256,0.004095111042261124
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,8192,256,0.014228444960382251
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,8192,3072,0.03772711091571384
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,8192,128,0.003718222181002299
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,8192,128,0.01350488927629259
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,8192,2560,0.037081778049468994
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,8192,64,0.0033582221302721235
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,8192,64,0.013487999637921652
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,8192,32,0.0037457777394188773
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,8192,32,0.013336000343163809
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,8192,2048,0.034797334008746676
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,8192,1536,0.03451733456717597
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,8192,1024,0.031805333164003156
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,8192,768,0.032131556007597185
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,7168,65536,0.20726666185590956
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,7168,16384,0.06948000192642212
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,7168,16384,0.06363644202550252
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,8192,256,0.025281776984532673
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,8192,512,0.027860442797342937
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,8192,128,0.02405422263675266
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,7168,12288,0.051947553952534996
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,7168,65536,0.26090312004089355
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,7168,10240,0.046720888879564076
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,7168,12288,0.05520977907710605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,7168,10240,0.041491554843054876
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,7168,8192,0.03885689046647813
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,7168,8192,0.03380266825358073
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,7168,7168,0.035657776726616755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,7168,7168,0.030571556753582422
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,7168,6144,0.03274844421280755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,7168,6144,0.026797332697444495
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,7168,5120,0.023388443721665278
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,7168,5120,0.029285334878497656
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,7168,16384,0.08749066458808051
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,7168,12288,0.07202488846249051
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,7168,4096,0.01945955554644267
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,7168,4096,0.025991999440722998
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,7168,65536,0.27466222974989146
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,7168,10240,0.06536089049445258
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,7168,3584,0.024458666642506916
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,7168,3584,0.017457778255144756
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,7168,3072,0.02233244478702545
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,7168,3072,0.015397333436542086
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,7168,8192,0.056341330210367836
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,7168,2560,0.021207110749350652
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,7168,2560,0.012678222523795234
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,7168,7168,0.053545776340696544
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,7168,2048,0.010537777509954242
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,7168,2048,0.01984355515903897
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,7168,1536,0.008847110801272923
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,7168,1536,0.018607111440764535
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,7168,6144,0.04794666502210829
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,7168,1024,0.007146666447321574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,7168,1024,0.017128000656763714
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,7168,5120,0.04476444588767158
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,7168,768,0.006224888894293044
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,7168,768,0.015372445185979208
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,7168,3584,0.04081155525313483
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,7168,512,0.0052382221652401825
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,7168,4096,0.040968888335757785
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,7168,512,0.014832889040311178
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,7168,256,0.0038764443662431505
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,7168,256,0.014167999227841696
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,7168,3072,0.03776622149679396
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,7168,128,0.0034684443639384378
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,7168,128,0.013493333425786761
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,7168,2560,0.03728533453411526
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,7168,64,0.0033111110743549135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,7168,64,0.013482666677898832
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,7168,32,0.0033893332713180115
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,7168,2048,0.03494666682349311
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,7168,32,0.013179555535316467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,7168,1536,0.03455911080042521
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,7168,1024,0.03193777799606323
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,7168,768,0.030811554855770532
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,6144,16384,0.06037422021230062
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,6144,16384,0.0609111123614841
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,6144,12288,0.04187822341918945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,6144,65536,0.19549777772691515
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,6144,12288,0.05040622088644239
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,6144,10240,0.03569422165552775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,6144,10240,0.04476444588767158
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,6144,65536,0.21572177939944795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,7168,256,0.02589599953757392
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,7168,512,0.02809600035349528
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,7168,128,0.02390577726893955
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,6144,8192,0.03751377926932441
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,6144,8192,0.030202666918436687
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,6144,7168,0.03421955638461643
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,6144,7168,0.025620443953408137
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,6144,6144,0.03123289015558031
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,6144,6144,0.022691556149058874
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,6144,5120,0.019852444529533386
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,6144,5120,0.02834133307139079
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,6144,16384,0.09069155322180854
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,6144,65536,0.2707688808441162
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,6144,4096,0.02479822271400028
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,6144,4096,0.017257778180970084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,6144,3584,0.014736000034544202
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,6144,12288,0.0753600001335144
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,6144,3584,0.023568888505299885
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,6144,3072,0.01296533313062456
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,6144,3072,0.022032888399230108
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,6144,10240,0.06799022356669109
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,6144,2560,0.011228444675604502
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,6144,2560,0.020850666695170935
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,6144,8192,0.05817066960864597
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,6144,2048,0.009630222287442949
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,6144,2048,0.019704888264338177
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,6144,7168,0.05428444345792135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,6144,1536,0.00812177773978975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,6144,1536,0.018400000201331247
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,6144,6144,0.04936888813972473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,6144,1024,0.006531555619504716
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,6144,1024,0.015763555963834126
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,6144,5120,0.04523377617200216
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,6144,768,0.005865777946180767
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,6144,768,0.015280000037617154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,6144,4096,0.041165331999460854
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,6144,512,0.0046640001237392426
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,6144,3584,0.03951644566324022
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,6144,512,0.014674666855070325
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,6144,256,0.00371111101574368
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,6144,256,0.0140684445699056
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,6144,3072,0.036899556716283165
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,6144,128,0.0034053333931499054
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,6144,128,0.01333688861793942
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,6144,2560,0.03609955641958449
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,6144,64,0.003126222226354811
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,6144,64,0.013336000343163809
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,6144,32,0.003363555504216088
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,6144,32,0.013235555754767524
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,6144,1536,0.03362222181426154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,6144,2048,0.033952001068327166
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,6144,768,0.03000088863902622
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,6144,1024,0.02994399931695726
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,5120,16384,0.058931556012895375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,5120,16384,0.054029332266913525
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,5120,65536,0.18797600269317627
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,5120,65536,0.1904026667277018
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,5120,12288,0.04210933380656772
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,5120,12288,0.05089511142836677
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,5120,10240,0.04359111189842224
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,5120,10240,0.03700977894994948
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,5120,8192,0.025776000486479864
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,6144,256,0.02531911101606157
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,6144,128,0.023988443944189284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,5120,8192,0.03590044379234314
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,5120,7168,0.022943110929595098
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,5120,7168,0.03297777639495002
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,6144,512,0.027665777338875666
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,5120,6144,0.0303111109468672
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,5120,6144,0.02056711084312863
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,5120,5120,0.017969777186711628
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,5120,5120,0.026878222823143005
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,5120,16384,0.07775200075573392
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,5120,4096,0.015217777755525378
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,5120,4096,0.024420445164044697
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,5120,65536,0.23945156733194986
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,5120,3584,0.013337777720557319
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,5120,12288,0.06511377625995211
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,5120,3584,0.02327466673321194
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,5120,3072,0.011886222494973077
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,5120,3072,0.02188355558448368
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,5120,10240,0.05959644582536486
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,5120,2560,0.00996888925631841
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,5120,2560,0.020482665962643094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,5120,8192,0.051795555485619434
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,5120,2048,0.008647110727098253
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,5120,2048,0.019466666711701285
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,5120,7168,0.04907200071546766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,5120,1536,0.00738666703303655
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,5120,6144,0.044575111733542554
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,5120,1536,0.01717333330048455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,5120,1024,0.006113777971929974
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,5120,1024,0.015375110838148328
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,5120,5120,0.04161689016554091
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,5120,768,0.005240889059172736
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,5120,768,0.01518311103185018
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,5120,4096,0.038225776619381376
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,5120,512,0.0041724443435668945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,5120,512,0.014711999230914645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,5120,3584,0.03786933422088623
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,5120,256,0.003657777690225177
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,5120,256,0.014095111025704278
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,5120,3072,0.0353999998834398
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,5120,128,0.0033484444850020935
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,5120,128,0.013226666384273104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,5120,2560,0.035319109757741295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,5120,64,0.003097777772280905
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,5120,64,0.013252444565296173
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,5120,32,0.0031759999692440033
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,5120,32,0.013069333301650153
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,5120,2048,0.03331466515858968
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,5120,1536,0.033324443631702
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,5120,1024,0.02931200133429633
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,5120,768,0.02972444560792711
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,4096,16384,0.04208533300293816
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,4096,16384,0.05661511421203613
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,4096,65536,0.1861413319905599
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,4096,12288,0.04699466625849406
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,4096,12288,0.034033778640959
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,4096,65536,0.14464622073703343
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,4096,10240,0.030096888542175293
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,4096,10240,0.04274844461017185
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,4096,8192,0.02645155621899499
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,4096,8192,0.03449599941571554
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,5120,512,0.027543110979927912
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,5120,256,0.024876443876160517
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,4096,7168,0.020800888538360596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,4096,7168,0.032249778509140015
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,5120,128,0.023817777633666992
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,4096,6144,0.029042667812771265
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,4096,6144,0.018551111221313477
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,4096,5120,0.01589777734544542
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,4096,5120,0.02658222284581926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,4096,16384,0.071743110815684
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,4096,4096,0.013970666461520724
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,4096,4096,0.023072888453801472
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,4096,12288,0.06047822369469536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,4096,3584,0.012069332930776807
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,4096,3584,0.022579555710156757
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,4096,65536,0.21310755941602919
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,4096,3072,0.010556444525718689
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,4096,3072,0.021409778131379023
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,4096,10240,0.055200000603993736
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,4096,2560,0.009473777479595607
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,4096,2560,0.020287111401557922
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,4096,8192,0.048319111267725624
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,4096,2048,0.008095111283991072
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,4096,2048,0.01924888955222236
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,4096,7168,0.04518577787611219
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,4096,1536,0.007181333170996771
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,4096,1536,0.01647111111217075
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,4096,6144,0.04195288817087809
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,4096,1024,0.005488889084921942
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,4096,1024,0.015517334143320719
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,4096,5120,0.03963199920124478
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,4096,768,0.004872000051869286
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,4096,768,0.015256888336605497
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,4096,4096,0.03620533479584588
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,4096,512,0.004280000097221798
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,4096,512,0.014396443963050842
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,4096,3584,0.03639199998643663
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,4096,256,0.003726222034957674
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,4096,256,0.014009777042600842
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,4096,3072,0.0340648889541626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,4096,128,0.0035413333939181436
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,4096,128,0.013329777452680798
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,4096,2560,0.034103999535242714
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,4096,64,0.003152888889114062
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,4096,64,0.013225778109497495
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,4096,32,0.003383999897374047
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,4096,32,0.01293155550956726
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,4096,2048,0.032413333654403687
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3584,65536,0.17920888794793022
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,4096,1536,0.03209066722128127
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,4096,1024,0.02753066685464647
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3584,16384,0.040059556563695274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3584,65536,0.12884444660610622
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3584,16384,0.055881778399149575
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3584,12288,0.04578311244646708
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3584,12288,0.03274844421280755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3584,10240,0.028592887851927016
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3584,10240,0.040919999281565346
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,4096,768,0.02822222312291463
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3584,8192,0.019520888725916546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3584,8192,0.03371289041307237
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,4096,512,0.02678844498263465
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,4096,256,0.024925332930352952
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3584,7168,0.030887110365761652
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3584,7168,0.01789688898457421
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,4096,128,0.023333332604832117
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3584,6144,0.015867556134859722
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3584,6144,0.028173334068722192
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3584,5120,0.013654222091039022
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3584,5120,0.02552177839808994
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,3584,16384,0.07054933574464586
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3584,4096,0.011720889144473605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3584,4096,0.023220444718996685
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,3584,65536,0.2081973287794325
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3584,3584,0.011187555889288584
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,3584,12288,0.05995733208126492
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3584,3584,0.022295110755496558
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3584,3072,0.009510222408506606
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3584,3072,0.021073778470357258
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,3584,10240,0.05450755688879225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3584,2560,0.008789333204428354
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3584,2560,0.019792889555295307
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,3584,8192,0.047473778327306114
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3584,2048,0.007853333320882585
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3584,2048,0.0188755558596717
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,3584,7168,0.044666667779286705
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3584,1536,0.006867555694447623
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3584,1536,0.01629333363638984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,3584,6144,0.04163022173775567
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3584,1024,0.005721777677536011
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3584,1024,0.015309333801269531
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,3584,5120,0.03939022289382087
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3584,768,0.004839999808205498
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3584,768,0.014848889576064216
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,3584,4096,0.03650310966703627
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3584,512,0.0041857775714662345
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3584,512,0.014527999692493014
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,3584,3584,0.035828444692823626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3584,256,0.003585777762863371
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3584,256,0.013890667094124688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,3584,3072,0.03403644429312812
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3584,128,0.003272888975010978
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3584,128,0.013304000099500021
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,3584,2560,0.03417155477735732
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3584,64,0.0031795555518733132
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3584,64,0.013167999684810638
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3584,32,0.0032319999817344877
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3584,32,0.01294133315483729
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,3584,2048,0.032264888286590576
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,3584,1536,0.030101332399580214
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,3584,1024,0.027280888623661462
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,3584,768,0.028047111299302842
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3072,16384,0.05401688814163208
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3072,16384,0.03501244386037191
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3072,65536,0.17686311403910318
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3072,12288,0.02755111124780443
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3072,12288,0.044416887892617114
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3072,65536,0.11734399530622695
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3072,10240,0.03947733177079095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3072,10240,0.024338665935728285
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3072,8192,0.021584888299306233
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3072,8192,0.03242222136921353
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,3584,512,0.026929777529504564
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3072,7168,0.016711999972661335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,3584,256,0.02481777800454034
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3072,7168,0.030416889323128596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,3584,128,0.023177777727444965
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3072,6144,0.014793778459231058
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3072,6144,0.02791022260983785
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3072,5120,0.012711111042234631
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3072,5120,0.025332444243960913
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,3072,16384,0.06948088937335543
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3072,4096,0.011015999648306104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3072,4096,0.023036445180575054
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,3072,65536,0.2062293291091919
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3072,3584,0.009922666682137383
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,3072,12288,0.05920888980229696
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3072,3584,0.022074666288163926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3072,3072,0.009110222260157267
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,3072,10240,0.05395022365781996
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3072,3072,0.020817778176731534
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3072,2560,0.008198222352398766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3072,2560,0.0192239996459749
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,3072,8192,0.04693333307902018
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3072,2048,0.007344000041484833
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3072,2048,0.01722666621208191
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,3072,7168,0.04453511039415995
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,3072,6144,0.04128355450100369
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3072,1536,0.006461333483457565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3072,1536,0.016334222422705758
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3072,1024,0.005418666534953647
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3072,1024,0.01533333294921451
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,3072,5120,0.03863111138343811
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3072,768,0.004338666796684265
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3072,768,0.01493866741657257
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,3072,3584,0.03595022360483805
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3072,512,0.0038951109680864546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,3072,4096,0.03570666578080919
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3072,512,0.014334221680959066
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3072,256,0.0034595556143257353
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3072,256,0.013728888498412238
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,3072,3072,0.03378222386042277
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3072,128,0.0031831111345026228
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3072,128,0.013208000196350945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,3072,2560,0.033766223324669734
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3072,64,0.003003555453485913
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3072,64,0.013169777890046438
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,3072,32,0.0030542222989930045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,3072,32,0.012889777620633444
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,3072,2048,0.03199822372860379
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2560,65536,0.16851022508409289
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2560,65536,0.09806044234169854
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,3072,1536,0.03108711043993632
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,3072,1024,0.027472888429959614
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2560,16384,0.05273155702484978
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,3072,768,0.02812355425622728
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2560,12288,0.04337244563632541
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2560,16384,0.03089244498146905
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2560,12288,0.025324443976084392
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2560,10240,0.02392888896995121
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2560,10240,0.03749777873357137
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2560,8192,0.019145778483814664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2560,8192,0.032046222024493747
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,3072,512,0.027072888281610277
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,3072,128,0.023258666197458904
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2560,7168,0.029226667351192896
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,3072,256,0.02512888941499922
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2560,7168,0.01756088932355245
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2560,6144,0.013974222871992322
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2560,6144,0.026751110951105755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2560,5120,0.01204177737236023
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2560,5120,0.024878222081396315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,2560,65536,0.2016151083840264
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,2560,12288,0.058279110325707324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,2560,16384,0.06877688566843669
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2560,4096,0.010463111102581024
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2560,4096,0.022650667362742957
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2560,3584,0.009649777577983009
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2560,3584,0.021756443712446425
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2560,3072,0.00887022250228458
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2560,3072,0.020604444874657523
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,2560,10240,0.05330311258633932
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2560,2560,0.008009778128729926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2560,2560,0.01830844415558709
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,2560,8192,0.04664177695910136
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2560,2048,0.007169778148333232
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2560,2048,0.01738933391041226
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,2560,7168,0.043775998883777194
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,2560,6144,0.04085866610209147
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2560,1536,0.0064560001095136
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,2560,5120,0.038253333833482527
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2560,1536,0.015903110305468243
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2560,1024,0.005024000174469418
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2560,1024,0.015105777316623263
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2560,768,0.00442755553457472
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2560,768,0.014688889185587565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,2560,4096,0.03580888774659898
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,2560,3584,0.03552177879545424
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2560,512,0.00388711111413108
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2560,512,0.014230221509933472
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2560,256,0.003383111208677292
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2560,256,0.013705777625242868
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,2560,3072,0.03356444504525926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2560,128,0.003083555648724238
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2560,128,0.013096888860066732
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,2560,2560,0.03353244397375319
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2560,64,0.0029377777957253983
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2560,64,0.013099555340078143
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2560,32,0.00305600009030766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2560,32,0.012787555654843649
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,2560,2048,0.031726221243540444
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2048,65536,0.16730844974517822
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2048,65536,0.0786951118045383
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2048,16384,0.02578488820128971
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2048,16384,0.051536887884140015
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,2560,1536,0.02978133492999607
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,2560,1024,0.027208000421524048
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2048,12288,0.04227733280923632
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2048,12288,0.02079288827048408
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2048,10240,0.0188426673412323
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2048,10240,0.03615999884075589
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,2560,768,0.0280693338976966
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2048,8192,0.01756977703836229
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2048,8192,0.03136266602410211
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,2560,512,0.026738666825824316
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2048,7168,0.015248888068728976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,2560,256,0.024772443705134924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2048,7168,0.028823110792371962
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,2560,128,0.02308533257908291
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2048,6144,0.013518222504191928
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2048,6144,0.0266311119000117
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2048,5120,0.012659555508030785
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2048,5120,0.02460355559984843
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,2048,65536,0.19165778160095215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2048,4096,0.010151999692122141
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2048,4096,0.022263111339675054
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,2048,16384,0.0661626656850179
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2048,3584,0.00923733330435223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2048,3584,0.021499555971887376
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,2048,12288,0.05635288688871595
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2048,3072,0.008605333666006723
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2048,3072,0.019186667270130582
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,2048,10240,0.051431109507878624
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2048,2560,0.007788444558779399
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2048,2560,0.017810665898852877
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,2048,8192,0.045018666320376925
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2048,2048,0.007113777928882175
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2048,2048,0.016855110724767048
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,2048,7168,0.04253422220547994
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,2048,6144,0.039641777674357094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2048,1536,0.005703111075692707
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2048,1536,0.01585155559910668
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2048,1024,0.004560889055331548
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2048,1024,0.015003555350833468
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,2048,5120,0.03746755586730109
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2048,768,0.0041893331540955436
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2048,768,0.014714666538768344
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,2048,4096,0.034733331865734525
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2048,512,0.00369422220521503
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2048,512,0.014152000347773233
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,2048,3584,0.03485777642991807
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2048,256,0.0032871110985676446
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2048,256,0.013619555367363824
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,2048,3072,0.03294666608174642
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2048,128,0.003020444470975134
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2048,128,0.01311822235584259
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,2048,2560,0.032650666104422674
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2048,64,0.0028924445311228433
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,2048,2048,0.029467556211683486
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2048,64,0.013040000365840064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,2048,32,0.0029768887907266617
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,2048,32,0.012773333324326409
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1536,65536,0.16391021675533718
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1536,65536,0.06074577569961548
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,2048,1536,0.029277334610621136
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1536,16384,0.05030666788419088
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1536,16384,0.02072000006834666
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,2048,1024,0.026877333720525105
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1536,12288,0.04045777850680881
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,2048,768,0.027698665857315063
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1536,10240,0.014697778556081982
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1536,12288,0.016349333855840895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1536,10240,0.035868446032206215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1536,8192,0.01369866645998425
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1536,8192,0.030214223596784804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,2048,512,0.026787555880016748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1536,7168,0.012853333519564735
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1536,7168,0.028350220786200628
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,2048,256,0.02458311120669047
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,2048,128,0.02349866595533159
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1536,6144,0.011974222130245633
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1536,6144,0.02637422250376807
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1536,5120,0.011066666907734342
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1536,5120,0.024382222029897902
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,1536,65536,0.18925866815778944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1536,4096,0.009706666900051964
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1536,4096,0.021926222576035395
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,1536,12288,0.054530666934119336
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,1536,16384,0.06524888674418132
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1536,3584,0.008940444224410588
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1536,3584,0.0203493336836497
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,1536,10240,0.0508284436331855
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1536,3072,0.008169777691364288
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1536,3072,0.018930666976504855
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1536,2560,0.00720266666677263
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1536,2560,0.017886221408843994
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,1536,8192,0.044624000787734985
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1536,2048,0.006351999938488007
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1536,2048,0.01665244499842326
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,1536,7168,0.04201777776082357
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1536,1536,0.005165333135260476
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1536,1536,0.01574577722284529
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,1536,6144,0.03937155670589871
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1536,1024,0.004331555631425646
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1536,1024,0.014944000376595391
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,1536,5120,0.03716355562210083
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1536,768,0.004015111260943943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1536,768,0.014488000008794995
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,1536,3584,0.0348053342766232
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,1536,4096,0.034803556071387395
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1536,512,0.003702222059170405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,1536,3072,0.03216444452603658
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1536,512,0.013959111438857185
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1536,256,0.0032906666811969546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1536,256,0.013600000076823764
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1536,128,0.0030088888274298776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1536,128,0.012972444295883179
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,1536,2560,0.03070666723781162
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1536,64,0.0029253332565228143
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1536,64,0.012988444831636218
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1536,32,0.002941333378354708
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1536,32,0.012716444830099741
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,1536,2048,0.02881866693496704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1024,65536,0.16354133023156062
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1024,65536,0.04361688759591845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1024,16384,0.016572443975342643
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1024,16384,0.048156443569395274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,1536,1536,0.029158221350775823
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1024,12288,0.013351110948456658
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1024,12288,0.03874933388498094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,1536,1024,0.026705778307384912
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1024,10240,0.01201333353916804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1024,10240,0.03515200151337518
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,1536,768,0.02771200074089898
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1024,8192,0.011142222417725457
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1024,8192,0.030174222257402208
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,1536,512,0.026497777965333726
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1024,7168,0.01073066641887029
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1024,7168,0.027871112028757732
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,1536,128,0.023149333066410486
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,1536,256,0.024625778198242188
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1024,6144,0.010013333625263637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1024,6144,0.025948445002237957
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1024,5120,0.009463110731707679
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1024,5120,0.023111111587948267
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,1024,65536,0.18584888511233857
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1024,4096,0.008816888762844933
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1024,4096,0.020940444535679285
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,1024,16384,0.06361511018541124
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1024,3584,0.008781332936551835
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1024,3584,0.019641778535313077
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,1024,12288,0.05295644534958733
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1024,3072,0.00812000036239624
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1024,3072,0.018383999665578205
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,1024,10240,0.04939377639028761
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1024,2560,0.006878222028414409
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1024,2560,0.01756444407833947
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,1024,8192,0.04340888725386726
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1024,2048,0.005966222120655908
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1024,2048,0.01664622293578254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,1024,7168,0.041788445578681103
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1024,1536,0.005125333451562458
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,1024,6144,0.03870933254559835
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1024,1536,0.015826667348543804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1024,1024,0.004403555558787452
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1024,1024,0.015056888262430826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,1024,5120,0.03667555583847894
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1024,768,0.003963555726740095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1024,768,0.014498665928840637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,1024,4096,0.03244888782501221
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1024,512,0.0035742223262786865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1024,512,0.014084445105658637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,1024,3584,0.03296444482273526
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1024,256,0.0032720000793536506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1024,256,0.01348977784315745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,1024,3072,0.029954665237002905
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1024,128,0.002951111023624738
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1024,128,0.0129120002190272
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,1024,2560,0.030349334081013996
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,1024,2048,0.028541333145565454
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1024,64,0.002871111035346985
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1024,64,0.012837332983811697
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,1024,32,0.0029093333416514923
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,1024,32,0.01273333364062839
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,768,65536,0.16234310468037924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,768,65536,0.04122311207983229
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,768,16384,0.013570666313171387
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,768,16384,0.04727200004789564
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,1024,1536,0.028869334194395278
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,768,12288,0.011549333731333414
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,768,12288,0.03808177842034234
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,1024,768,0.027930667002995808
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,1024,1024,0.026605332891146343
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,768,10240,0.010602666272057427
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,768,10240,0.03467822074890137
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,768,8192,0.009911999934249455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,768,8192,0.029716445340050593
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,1024,512,0.02667733364635044
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,768,7168,0.009386666946940953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,768,7168,0.02776088813940684
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,1024,256,0.024381332927280005
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,768,6144,0.0088746663596895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,768,6144,0.02502577834659153
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,1024,128,0.023248889380031165
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,768,5120,0.009787555370065901
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,768,5120,0.023056889573733013
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,768,16384,0.0628711117638482
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,768,65536,0.18435822592841256
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,768,4096,0.008442666795518663
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,768,4096,0.020440000626775954
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,768,3584,0.007817777494589487
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,768,3584,0.01941155559486813
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,768,12288,0.05295288893911573
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,768,3072,0.0075351107451650835
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,768,3072,0.018617777360810172
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,768,10240,0.04951999915970696
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,768,2560,0.006526222245560751
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,768,2560,0.017474666237831116
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,768,8192,0.04347999890645345
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,768,2048,0.005655999812814925
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,768,2048,0.016725333200560678
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,768,7168,0.041181332535213895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,768,6144,0.038005332152048744
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,768,1536,0.004942222187916438
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,768,1536,0.015778667396969266
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,768,1024,0.004247111164861255
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,768,1024,0.01498133275243971
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,768,5120,0.03454400102297465
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,768,768,0.0038622220357259116
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,768,768,0.014440000057220459
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,768,4096,0.031874666611353554
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,768,3584,0.03162755568822225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,768,512,0.0035146665241983202
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,768,512,0.01404888927936554
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,768,256,0.003120888852410846
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,768,256,0.013546666337384118
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,768,3072,0.030016889174779255
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,768,128,0.0029279999434947968
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,768,128,0.013023110727469126
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,768,2560,0.03009422289000617
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,768,64,0.002760888801680671
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,768,2048,0.02812355425622728
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,768,64,0.012975999878512489
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,768,32,0.002790222151411904
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,768,32,0.012637333737479316
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,512,65536,0.15973244773017034
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,512,65536,0.031288888719346784
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,512,16384,0.011305777562989129
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,512,16384,0.04659644431538052
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,768,1536,0.028780443800820246
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,768,1024,0.026540444956885442
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,512,12288,0.010189333723651038
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,512,12288,0.038091556893454656
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,512,10240,0.009737778041097853
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,512,10240,0.03388533327314589
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,768,768,0.02702755563788944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,512,8192,0.009961778091059791
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,512,8192,0.02940088841650221
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,768,512,0.026608000199000042
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,512,7168,0.009494221872753566
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,512,7168,0.02725155486000909
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,768,256,0.024461333950360615
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,512,6144,0.008812444077597724
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,512,6144,0.025036444266637165
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,768,128,0.02316711180739933
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,512,5120,0.007806222471925948
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,512,5120,0.022282666630215112
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,512,65536,0.18448621696896025
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,512,4096,0.007043555378913879
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,512,4096,0.020278222031063504
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,512,16384,0.06224266688028971
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,512,3584,0.0074275558193524676
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,512,3584,0.01938222183121575
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,512,12288,0.05283200078540378
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,512,3072,0.0070088886552386815
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,512,3072,0.018536888890796237
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,512,10240,0.049364444282319814
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,512,2560,0.0064426664676931165
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,512,2560,0.017507554756270517
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,512,8192,0.04272977842224968
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,512,2048,0.005667555663320754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,512,2048,0.01648355523745219
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,512,7168,0.04028355413013034
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,512,1536,0.004869333157936732
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,512,1536,0.01568266749382019
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,512,6144,0.037046220567491316
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,512,1024,0.004237333519591225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,512,1024,0.014660444524553088
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,512,5120,0.034487111700905695
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,512,768,0.003929777691761653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,512,768,0.014308444327778287
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,512,4096,0.031565331750445895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,512,512,0.0035964445107513
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,512,3584,0.03183022141456604
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,512,512,0.01391377713945177
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,512,256,0.0032275555034478507
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,512,256,0.013357333011097379
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,512,3072,0.029935999049080744
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,512,128,0.0029395555870400537
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,512,128,0.012899555265903473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,512,2560,0.029955555995305378
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,512,64,0.0027822222974565295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,512,64,0.012832889126406776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,512,32,0.0028719999310043124
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,512,32,0.012606222596433429
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,256,65536,0.022970666488011677
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,256,65536,0.15762221813201904
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,512,2048,0.028156446086035833
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,256,16384,0.010204444328943888
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,256,16384,0.04636355572276645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,512,1536,0.028472890456517536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,256,12288,0.009024000002278222
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,512,1024,0.02631644407908122
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,256,12288,0.037091556522581316
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,256,10240,0.00814577771557702
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,256,10240,0.03327733278274536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,512,768,0.027401778433057997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,256,8192,0.007497777541478475
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,256,8192,0.028107557031843398
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,512,512,0.02626488937271966
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,256,7168,0.007189333438873291
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,512,256,0.024480000138282776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,256,7168,0.026111111044883728
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,256,6144,0.006814222368929122
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,256,6144,0.02424977719783783
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,512,128,0.022924444741672937
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,256,5120,0.007096889118353526
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,256,5120,0.022256000174416438
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,256,65536,0.18437422646416557
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,256,4096,0.006741333338949416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,256,4096,0.0203840004073249
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,256,16384,0.061991108788384326
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,256,3584,0.007168889045715332
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,256,3584,0.019489778412712943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,256,12288,0.05231377813551161
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,256,3072,0.006753777878151999
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,256,3072,0.018295110927687753
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,256,10240,0.04881155490875244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,256,2560,0.006157333238257303
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,256,2560,0.017323555217848886
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,256,8192,0.04243910974926419
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,256,2048,0.005474666754404704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,256,2048,0.016358221570650738
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,256,7168,0.039445334010654025
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,256,1536,0.004795555439260271
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,256,6144,0.03747555613517761
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,256,1536,0.015480000111791821
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,256,1024,0.004055110944641961
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,256,1024,0.014668444792429606
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,256,5120,0.03384088807635837
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,256,768,0.003699555579158995
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,256,768,0.014289778139856128
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,256,4096,0.03135733471976386
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,256,512,0.003400000019205941
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,256,512,0.013946667313575745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,256,3584,0.03158044483926561
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,256,256,0.0030257778449190986
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,256,256,0.013295999831623502
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,256,3072,0.029779556724760268
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,256,128,0.0028320000403457214
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,256,128,0.012884444660610624
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,256,2560,0.029814223448435467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,256,64,0.0026791110220882627
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,256,2048,0.02827822168668111
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,256,64,0.012683555483818054
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,256,32,0.0027013334135214486
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,256,32,0.012479110724396177
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,128,65536,0.014811555544535318
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,128,65536,0.15782399972279867
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,128,16384,0.007596444752481248
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,128,16384,0.045112000571356885
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,256,1536,0.02843822373284234
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,128,12288,0.007122666471534305
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,128,12288,0.03743911120626662
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,256,1024,0.026141333911154006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,128,10240,0.007136888802051544
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,128,10240,0.03304266598489549
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,256,768,0.02706577711635166
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,128,8192,0.007019555403126611
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,128,8192,0.02796622117360433
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,256,512,0.026404443714353774
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,128,7168,0.006881777611043718
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,128,7168,0.026333333717452154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,256,256,0.024711110525661047
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,128,6144,0.00666222224632899
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,128,6144,0.024259555670950148
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,256,128,0.023000889354281958
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,128,5120,0.00702133360836241
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,128,5120,0.02204622162712945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,128,65536,0.1831902265548706
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,128,4096,0.006633777585294511
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,128,4096,0.02017333275742001
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,128,16384,0.061122669114006885
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,128,3584,0.007081777685218387
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,128,3584,0.019485332899623446
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,128,12288,0.05197333296140035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,128,3072,0.006771555377377405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,128,3072,0.018345778187115986
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,128,10240,0.04706666535801358
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,128,2560,0.006123555617200003
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,128,2560,0.017308443784713745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,128,8192,0.04089066717359755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,128,2048,0.005422222117582957
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,128,2048,0.01645066671901279
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,128,7168,0.03846399982770284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,128,1536,0.0047404443224271136
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,128,1536,0.015513777732849121
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,128,6144,0.03566222058402167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,128,1024,0.0040702223777771
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,128,1024,0.01463733282354143
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,128,5120,0.03379022081693014
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,128,768,0.003671110918124517
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,128,768,0.014195554786258273
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,128,4096,0.03125955661137899
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,128,512,0.0033902221669753394
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,128,512,0.013758222262064615
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,128,3584,0.03132355544302199
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,128,256,0.002995555599530538
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,128,256,0.013247999880048962
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,128,3072,0.02968177861637539
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,128,128,0.0028328889360030494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,128,128,0.012808000048001608
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,128,2560,0.029686222473780315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,128,64,0.0026826666047175727
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,128,64,0.012724444270133972
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,128,32,0.0026702222724755607
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,128,32,0.012498666842778524
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,64,65536,0.012835555606418185
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,64,65536,0.15667200088500977
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,64,16384,0.006791111081838608
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,64,16384,0.04479022158516777
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,64,12288,0.006700444552633498
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,64,12288,0.035875555541780256
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,64,10240,0.006899555524190267
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,64,10240,0.03202577763133579
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,64,8192,0.006793777975771163
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,64,8192,0.027952889601389568
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,64,7168,0.006697777658700943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,64,7168,0.02607466611597273
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,64,6144,0.006531555619504716
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,64,6144,0.024136000209384497
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,64,5120,0.006875555548402999
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,64,5120,0.022133333815468684
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,64,4096,0.0064560001095136
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,64,4096,0.020249777370029025
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,64,3584,0.0069520001610120135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,64,3584,0.019324445062213473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,64,3072,0.006637333167923822
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,64,3072,0.018272888329293992
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,64,2560,0.006049777898523543
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,64,2560,0.0173342227935791
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,64,2048,0.005381333331267039
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,64,2048,0.016280000408490498
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,64,1536,0.0047360000511010485
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,64,1536,0.015460444821251763
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,64,1024,0.004025777594910728
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,64,1024,0.014641778336630927
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,64,768,0.003670222229427762
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,64,768,0.014312000738249885
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,64,512,0.0033413333197434745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,64,512,0.013777777552604675
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,64,256,0.002970666728085942
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,128,2048,0.027848889430363972
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,64,256,0.013229333692126803
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,64,128,0.002790222151411904
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,64,64,0.0026266665922270883
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,64,128,0.012698666916953193
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,64,32,0.0026604444202449587
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,64,64,0.012656889028019376
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,64,32,0.012439111040698158
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,32,65536,0.01164177805185318
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,32,16384,0.006728000111050076
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,32,65536,0.15732533401913112
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,32,16384,0.04557244314087761
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,32,12288,0.006506666541099548
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,32,10240,0.006758222149478064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,32,12288,0.037237332926856145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,128,1536,0.0284551117155287
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,32,8192,0.006601777755551868
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,32,10240,0.03204088740878635
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,128,1024,0.026075555218590632
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,32,8192,0.02804622385236952
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,32,7168,0.0064995553758409284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,32,6144,0.006425777657164468
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,32,7168,0.02609866691960229
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,32,5120,0.006800000038411882
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,32,6144,0.024043555061022442
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,32,4096,0.0063742221229606206
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,32,5120,0.02219733264711168
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,32,3584,0.0068764446510208984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,32,4096,0.02033333314789666
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,32,3072,0.006513777706358168
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,32,3584,0.019349333312776353
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,32,2560,0.006057777752478917
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,32,3072,0.018274666534529794
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,32,2048,0.005376889059940974
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,32,2560,0.017343999611006845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,32,1536,0.004691555682155821
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,32,2048,0.01628266606065962
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,32,1024,0.00398577791121271
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,32,1536,0.015527110960748462
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,32,768,0.003629333443111844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,32,1024,0.014697778556081982
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,32,512,0.003306666596068276
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,32,768,0.014243554737832812
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,32,256,0.0029279999434947968
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,32,512,0.013803555733627744
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,32,128,0.0027235555979940626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,32,256,0.013167999684810638
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,32,64,0.002593777866827117
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,32,128,0.012740444805887012
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,32,64,0.012456888953844706
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,128,32,32,0.0025964443468385274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,128,32,32,0.012454222473833295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,128,768,0.02733777794573042
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,128,512,0.02628622286849552
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,128,256,0.0240666667620341
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,128,128,128,0.023440000083711412
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,65536,10240,0.20094399982028535
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,65536,16384,0.3134568797217475
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,65536,12288,0.24502134323120117
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,65536,16384,0.320213344362047
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,65536,12288,0.26070578893025714
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,65536,8192,0.16684888468848336
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,65536,10240,0.33260265986124676
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,65536,10240,0.22317510181003145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,65536,16384,0.521942244635688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,65536,12288,0.3970133463541667
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,65536,8192,0.1829813321431478
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,65536,7168,0.1492675542831421
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,65536,8192,0.26652177174886066
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,65536,7168,0.1654568910598755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,65536,5120,0.11096000009112889
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,65536,6144,0.13422311676873103
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,65536,5120,0.13082399633195665
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,65536,7168,0.23377243677775064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,65536,6144,0.1483208868238661
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,65536,6144,0.20231466823154023
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,65536,5120,0.17074044545491537
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,65536,4096,0.0917617744869656
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,65536,4096,0.13815200328826904
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,65536,4096,0.10912799835205078
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,65536,3072,0.07436977492438422
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,65536,3584,0.08455733458201091
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,65536,2560,0.06286133660210504
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,65536,3072,0.09341600206163193
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,65536,3584,0.1210906638039483
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,65536,3584,0.10281155506769817
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,65536,2560,0.08968710899353027
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,65536,2048,0.05372088816430834
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,65536,3072,0.11698222160339355
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,65536,2560,0.0812657806608412
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,65536,1536,0.044300443596310086
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,65536,2048,0.07165244552824232
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,65536,1024,0.03560977843072679
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,65536,1024,0.05317066775427925
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,65536,1536,0.06256710820727877
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,65536,1536,0.05746755335066053
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,65536,768,0.03151911165979173
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,65536,1024,0.040927112102508545
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,65536,256,0.023338667220539514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,65536,768,0.04816533459557427
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,65536,512,0.02719822194841173
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,65536,2048,0.07354311148325603
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,65536,128,0.00925511121749878
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,65536,256,0.013919111755159167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,65536,256,0.03680977887577481
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,65536,512,0.04369422131114536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,65536,512,0.023613333702087402
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,65536,128,0.020990222692489624
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,65536,768,0.03206133180194431
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,65536,64,0.0076657773719893555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,65536,128,0.03130311104986403
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,65536,32,0.007096889118353526
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,65536,64,0.022269333402315777
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,65536,32,0.02270488937695821
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,16384,12288,0.07922666602664523
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,16384,16384,0.12168444527520074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,16384,16384,0.10143644279903835
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,16384,12288,0.09842489163080852
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,16384,12288,0.09651377465989854
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,16384,16384,0.1288693348566691
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,16384,10240,0.06682666805055407
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,16384,10240,0.08402844270070393
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,16384,8192,0.057513780064053006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,16384,10240,0.08332977692286174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,16384,8192,0.07249066564771864
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,16384,8192,0.06798933611975776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,16384,7168,0.05125511023733351
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,16384,6144,0.045794665813446045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,16384,7168,0.06668800115585327
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,16384,7168,0.05921689007017347
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,16384,6144,0.06115021970536974
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,16384,6144,0.05193600058555603
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,16384,65536,0.3469964398278131
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,16384,5120,0.04121066795455085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,16384,5120,0.05637866920895047
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,16384,4096,0.03524533245298598
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,16384,5120,0.044213334719340004
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,16384,4096,0.051227556334601514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,16384,4096,0.037717332442601524
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,16384,3584,0.033124443557527326
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,16384,3072,0.030791110462612573
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,16384,3584,0.04850133260091146
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,16384,3072,0.02875022093454997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,16384,3072,0.046015110280778676
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,16384,3584,0.032496889432271324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,16384,2560,0.024838222397698298
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,16384,2560,0.028516444895002577
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,16384,2560,0.04426755507787069
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,16384,1536,0.02217866645918952
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,16384,2048,0.025793777571784124
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,16384,1536,0.037047111325793795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,16384,1024,0.020111110475328233
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,16384,2048,0.04077777928776211
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,16384,2048,0.020901333954599168
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,16384,1536,0.01670222149954902
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,16384,768,0.009777777724795872
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,16384,1024,0.011941333611806234
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,16384,1024,0.0348639984925588
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,16384,768,0.018758222460746765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,16384,512,0.007719111111429002
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,16384,512,0.017866666118303936
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,16384,256,0.0057884446448749965
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,16384,512,0.031175110075208876
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,16384,768,0.03312888741493225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,16384,256,0.01517688896920946
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,16384,128,0.004200889004601372
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,16384,256,0.025983111725913152
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,16384,128,0.014166222678290473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,16384,64,0.0038497779104444715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,16384,128,0.024140444066789415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,16384,32,0.003966222206751506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,16384,64,0.01459200017982059
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,16384,32,0.014483556151390076
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,16384,65536,0.38938045501708984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,12288,16384,0.10667288965649074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,12288,16384,0.09049155314763387
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,12288,16384,0.11072978046205308
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,12288,12288,0.07100355625152588
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,12288,12288,0.08530044555664062
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,12288,10240,0.06054311328464084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,12288,10240,0.06508800056245592
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,16384,65536,0.583587540520562
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,12288,12288,0.08637777964274089
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,12288,10240,0.07628977961010404
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,12288,8192,0.052432888083987765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,12288,8192,0.053386665052837796
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,12288,8192,0.06454044580459595
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,12288,65536,0.30982311566670734
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,12288,7168,0.0463084446059333
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,12288,7168,0.060953776041666664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,12288,6144,0.04383822282155355
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,12288,7168,0.04624799887339274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,12288,6144,0.05719021956125895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,12288,6144,0.040630221366882324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,12288,5120,0.03711288836267259
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,12288,65536,0.3430817657046848
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,12288,5120,0.05291644401020474
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,12288,4096,0.033145778708987765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,12288,5120,0.034873776965671115
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,12288,4096,0.04810400141610039
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,12288,4096,0.031311998764673867
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,12288,3584,0.03155644403563605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,12288,3584,0.046145776907602944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,12288,3584,0.02590755621592204
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,12288,3072,0.028466665082507666
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,12288,2560,0.026216889421145122
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,12288,3072,0.02292177743381924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,12288,3072,0.043324444029066295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,12288,2560,0.04110577702522278
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,12288,2048,0.024319110645188227
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,12288,2048,0.038246221012539335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,12288,1536,0.021607110897699993
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,12288,1536,0.035707556539111666
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,12288,1024,0.009702222214804756
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,12288,2048,0.01647200021478865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,12288,1024,0.01926488843229082
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,12288,768,0.008280889027648503
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,12288,1536,0.013183111117945777
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,12288,2560,0.019719999697473314
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,12288,1024,0.03421333432197571
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,12288,768,0.018326222896575928
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,12288,512,0.0069297779765393995
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,12288,768,0.0329422222243415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,12288,512,0.017058667209413316
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,12288,256,0.004557333472702238
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,12288,512,0.030784000953038532
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,12288,256,0.015019555886586508
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,12288,128,0.003953777667548922
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,12288,256,0.02583911187118954
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,12288,128,0.01402933398882548
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,12288,64,0.003556444413132138
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,12288,128,0.023779556155204773
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,12288,64,0.014474666780895658
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,12288,32,0.003653333418899112
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,12288,32,0.014317333698272705
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,10240,16384,0.08440177970462376
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,10240,16384,0.09634666972690159
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,10240,12288,0.06740622387992011
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,10240,12288,0.0788373351097107
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,12288,65536,0.41715110672844785
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,10240,12288,0.08087022436989678
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,10240,16384,0.09973066382937962
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,10240,10240,0.057371556758880615
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,10240,8192,0.04845066534148323
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,10240,10240,0.0748951103952196
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,10240,10240,0.06882933113310072
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,10240,8192,0.06183555391099718
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,10240,8192,0.045018666320376925
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,10240,7168,0.044080889887279935
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,10240,7168,0.058509336577521436
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,10240,6144,0.040256887674331665
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,10240,6144,0.03510133425394694
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,10240,5120,0.036531554328070745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,10240,6144,0.05469155642721388
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,10240,7168,0.039926220973332725
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,10240,5120,0.03050044510099623
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,10240,65536,0.2951875527699788
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,10240,5120,0.051046222448349
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,10240,4096,0.03199199835459391
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,10240,4096,0.045419553915659584
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,10240,65536,0.3162933455573188
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,10240,4096,0.025049777494536504
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,10240,3072,0.027514666318893433
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,10240,3584,0.029888000753190782
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,10240,3584,0.02258844508065118
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,10240,3584,0.04403733213742574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,10240,3072,0.04276177618238661
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,10240,2560,0.025255999631351892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,10240,3072,0.019938665959570143
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,10240,2048,0.02241155505180359
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,10240,2560,0.03984355595376756
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,10240,1536,0.01166488892502255
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,10240,2048,0.03700800074471368
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,10240,2560,0.01763911048571269
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,10240,1536,0.020594666401545208
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,10240,1024,0.009106666677527958
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,10240,768,0.007847111258241866
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,10240,2048,0.014682667122946845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,10240,768,0.017796445224020217
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,10240,1536,0.03490666548411051
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,10240,1024,0.019034667147530448
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,10240,1024,0.03358577688535055
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,10240,512,0.006653333289755716
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,10240,768,0.03266577919324239
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,10240,512,0.015951999359660678
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,10240,256,0.0047182221379545005
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,10240,256,0.014992000328169929
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,10240,128,0.0038399998512532976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,10240,512,0.027831110689375136
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,10240,256,0.02544533378548092
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,10240,128,0.014148443937301636
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,10240,64,0.003516444315512975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,10240,128,0.02350133326318529
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,10240,32,0.003535110917356279
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,10240,64,0.014279110564125909
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,10240,32,0.014131555954615274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,8192,16384,0.06533510817421807
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,8192,16384,0.08447466956244574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,8192,12288,0.05836533175574409
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,8192,12288,0.052129778597089976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,8192,16384,0.07374844286176893
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,10240,65536,0.36366844177246094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,8192,10240,0.04457155532307095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,8192,10240,0.0621946652730306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,8192,12288,0.06996799839867486
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,8192,10240,0.045019557078679405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,8192,8192,0.039044443104002215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,8192,8192,0.04050577680269877
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,8192,8192,0.054415110084745616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,8192,7168,0.035587555832333036
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,8192,7168,0.05084266596370273
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,8192,6144,0.032722665203942194
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,8192,6144,0.02856711215443081
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,8192,7168,0.032455109887652926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,8192,6144,0.04825688732994927
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,8192,5120,0.030173334810468886
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,8192,5120,0.0248346659872267
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,8192,5120,0.04523377617200216
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,8192,4096,0.026664889521068994
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,8192,3584,0.02490577763981289
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,8192,4096,0.021152888735135395
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,8192,65536,0.25678666432698566
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,8192,4096,0.04085955686039395
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,8192,65536,0.2121582163704766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,8192,3584,0.03960799839761522
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,8192,3584,0.018567999203999836
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,8192,3072,0.023398222194777593
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,8192,2560,0.02165688905451033
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,8192,3072,0.03791466686460707
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,8192,2560,0.03730310996373495
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,8192,2048,0.0203137778573566
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,8192,3072,0.016576000385814242
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,8192,1536,0.009533333281675974
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,8192,2048,0.035447110732396446
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,8192,2048,0.011582222249772815
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,8192,1536,0.01884533299340142
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,8192,2560,0.01407911048995124
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,8192,1024,0.007824888659848107
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,8192,1024,0.017472000585661996
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,8192,1536,0.033737778663635254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,8192,768,0.006723555425802867
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,8192,1024,0.0328400002585517
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,8192,768,0.015481778317027621
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,8192,512,0.005669333454635408
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,8192,768,0.030850665436850652
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,8192,256,0.004053333153327306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,8192,512,0.01497244503762987
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,8192,512,0.027307555079460144
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,8192,256,0.013968888256284924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,8192,128,0.003644444462325838
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,8192,128,0.013159111142158508
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,8192,256,0.025190222594473097
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,8192,64,0.0032417778339650896
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,8192,128,0.023652444283167522
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,8192,32,0.0035377778112888336
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,8192,64,0.013419555293189155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,8192,32,0.013314666847387949
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,7168,16384,0.060749332110087075
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,7168,12288,0.046148445871141225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,7168,16384,0.08132000101937188
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,7168,12288,0.049341334236992725
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,8192,65536,0.2770462301042345
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,7168,16384,0.06814489099714491
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,7168,10240,0.04216444492340088
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,7168,12288,0.06792800294028388
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,7168,10240,0.0611342191696167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,7168,10240,0.039812445640563965
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,7168,8192,0.03771022293302748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,7168,8192,0.05286222365167406
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,7168,65536,0.19884267118242052
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,7168,7168,0.03386755453215705
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,7168,7168,0.03144711256027222
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,7168,7168,0.05024888780381945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,7168,8192,0.03235377868016561
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,7168,6144,0.0318177772892846
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,7168,6144,0.025234666135576036
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,7168,5120,0.028688000308142766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,7168,4096,0.025370667378107708
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,7168,6144,0.04714666803677877
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,7168,5120,0.04450666573312548
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,7168,5120,0.022075555390781824
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,7168,4096,0.040285332335366145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,7168,3584,0.023951111568344965
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,7168,4096,0.01882844501071506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,7168,3584,0.039784000979529485
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,7168,3072,0.02223555578125848
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,7168,3072,0.03779911001523336
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,7168,3584,0.016460445192125108
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,7168,2560,0.021037333541446265
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,7168,65536,0.24945243199666342
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,7168,3072,0.014508444401952954
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,7168,2048,0.010444444086816577
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,7168,2560,0.0374257763226827
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,7168,2048,0.019719110594855416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,7168,2560,0.012647111382749347
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,7168,1536,0.009487999810112847
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,7168,2048,0.03543377916018168
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,7168,1536,0.018364444375038147
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,7168,1024,0.007102222078376346
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,7168,1536,0.03379822108480666
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,7168,768,0.015396444333924187
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,7168,1024,0.017128000656763714
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,7168,768,0.006246222390068903
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,7168,1024,0.033534222178988986
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,7168,512,0.00490133340160052
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,7168,768,0.03158222304450141
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,7168,512,0.01479022204875946
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,7168,256,0.0038888889054457345
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,7168,512,0.027475555737813313
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,7168,256,0.014117333624098035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,7168,128,0.0035297779573334586
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,7168,128,0.023336889015303716
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,7168,256,0.025199999411900837
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,7168,128,0.013159111142158508
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,7168,64,0.0032177778581778207
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,7168,64,0.013375110924243927
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,7168,32,0.0033395555284288195
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,7168,32,0.013206221991115145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,6144,16384,0.057597332530551486
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,6144,16384,0.08472800254821777
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,7168,65536,0.2569635444217258
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,6144,16384,0.06060622135798136
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,6144,12288,0.04746311240726047
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,6144,12288,0.07383288939793904
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,6144,12288,0.04079466727044847
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,6144,10240,0.04049866729312473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,6144,10240,0.035170667701297335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,6144,8192,0.035535110367669
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,6144,10240,0.06349777513080172
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,6144,8192,0.028809779220157202
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,6144,65536,0.18468088573879668
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,6144,8192,0.054655111498302884
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,6144,7168,0.032850666178597346
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,6144,7168,0.025575111309687298
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,6144,6144,0.030024889442655776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,6144,5120,0.027765333652496338
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,6144,7168,0.051557332277297974
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,6144,6144,0.0484426650736067
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,6144,6144,0.022635555929607813
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,6144,4096,0.02416533397303687
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,6144,5120,0.044884443283081055
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,6144,4096,0.04014044337802463
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,6144,5120,0.019859555694792006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,6144,3584,0.023169777459568445
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,6144,4096,0.016807110773192514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,6144,3584,0.03903822104136149
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,6144,3072,0.02169422143035465
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,6144,65536,0.25838756561279297
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,6144,3584,0.01458044515715705
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,6144,2560,0.011244444383515252
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,6144,3072,0.03676799933115641
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,6144,2560,0.020534222324689228
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,6144,3072,0.012984000146389008
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,6144,2048,0.009638222555319468
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,6144,2560,0.03621422251065572
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,6144,2048,0.019659555620617337
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,6144,1536,0.008104888929261101
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,6144,2048,0.034627556800842285
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,6144,1536,0.01829688913292355
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,6144,1024,0.00666133314371109
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,6144,1024,0.017144888639450073
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,6144,1536,0.0324853327539232
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,6144,512,0.004552888787455029
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,6144,768,0.006001777946949005
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,6144,1024,0.030508445368872747
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,6144,768,0.015432889262835184
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,6144,768,0.03089600139194065
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,6144,512,0.01482488877243466
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,6144,256,0.0038373333712418876
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,6144,512,0.026928888426886663
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,6144,256,0.013937777943081327
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,6144,128,0.003411555455790626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,6144,256,0.024677332904603746
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,6144,128,0.013069333301650153
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,6144,64,0.0031848889258172777
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,6144,128,0.023453333311610754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,6144,32,0.0032755554550223877
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,6144,64,0.013024888932704926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,6144,32,0.013057777451144325
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,5120,16384,0.07493422428766887
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,5120,16384,0.05532000131077237
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,5120,12288,0.042048000627093844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,5120,16384,0.051933334933386914
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,5120,12288,0.04620888829231262
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,5120,12288,0.06211821900473701
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,5120,10240,0.03922399878501892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,5120,10240,0.0366977784368727
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,5120,10240,0.05621688895755344
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,5120,8192,0.034697777695126004
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,5120,8192,0.05002489023738437
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,5120,65536,0.17682933807373047
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,6144,65536,0.2068880001703898
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,5120,8192,0.03126577867401971
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,5120,7168,0.03199111090766059
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,5120,7168,0.028584000137117174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,5120,6144,0.029940442906485662
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,5120,6144,0.04429955614937676
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,5120,7168,0.04725955592261421
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,5120,5120,0.027012444204754297
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,5120,5120,0.04165066613091363
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,5120,6144,0.0203146669599745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,5120,4096,0.02392888896995121
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,5120,4096,0.03725688987308078
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,5120,5120,0.0177208897140291
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,5120,3584,0.022685334086418152
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,5120,3072,0.011674666570292579
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,5120,3584,0.03734488950835334
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,5120,65536,0.23080354266696504
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,5120,3072,0.021355556117163763
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,5120,4096,0.014858666393491956
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,5120,2560,0.010299555957317352
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,5120,2560,0.02016533414522807
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,5120,3072,0.035403556293911405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,5120,3584,0.013089777694808112
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,5120,2048,0.008989333278603023
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,5120,2560,0.03526488939921061
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,5120,2048,0.01944977707333035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,5120,1024,0.006232888748248418
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,5120,1536,0.007671111159854465
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,5120,2048,0.03390489021937052
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,5120,1536,0.01814133259985182
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,5120,1536,0.03230666783120897
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,5120,1024,0.015977778368526034
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,5120,768,0.005437333136796951
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,5120,1024,0.029909332593282063
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,5120,768,0.01538133290078905
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,5120,512,0.004442666553788715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,5120,768,0.029493331909179688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,5120,512,0.014639111028777229
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,5120,256,0.003654222107595868
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,5120,512,0.026797332697444495
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,5120,256,0.013904000322024027
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,5120,128,0.0033679999825027254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,5120,256,0.024837333295080397
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,5120,128,0.013122667041089801
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,5120,64,0.0031066667288541794
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,5120,32,0.003157333367400699
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,5120,128,0.02328266700108846
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,5120,64,0.013030221892727746
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,5120,32,0.0129466669427024
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,5120,65536,0.1832186645931668
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,4096,16384,0.053660445743136935
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,4096,16384,0.07195466756820679
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,4096,16384,0.041725334193971425
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,4096,12288,0.0447982218530443
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,4096,12288,0.034563554657830134
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,4096,12288,0.06012711260053846
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,4096,10240,0.038146668010287814
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,4096,10240,0.05432000093989902
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,4096,10240,0.029289778735902574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,4096,8192,0.03326488865746392
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,4096,8192,0.025661332739724055
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,4096,7168,0.03105777833196852
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,4096,8192,0.047480000389946826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,4096,65536,0.17277067237430146
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,4096,6144,0.028627554575602215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,4096,7168,0.04506311151716444
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,4096,7168,0.023119111855824787
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,4096,6144,0.04265866676966349
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,4096,5120,0.025552888711293537
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,4096,6144,0.020959110723601446
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,4096,5120,0.04060711132155524
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,4096,4096,0.023198222120602924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,4096,5120,0.0188426673412323
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,4096,4096,0.03640000025431315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,4096,3584,0.013082666529549493
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,4096,4096,0.015388444066047668
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,4096,3584,0.022408889399634466
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,4096,3072,0.011752888560295105
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,4096,3584,0.03646311163902283
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,4096,65536,0.22085600429111055
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,4096,3072,0.034544888469907976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,4096,3072,0.021222222182485793
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,4096,2560,0.00889244427283605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,4096,2560,0.020040889581044514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,4096,2048,0.01891911029815674
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,4096,2048,0.007810666329330868
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,4096,2560,0.034485333495669894
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,4096,1536,0.00683111117945777
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,4096,2048,0.03289511137538486
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,4096,1536,0.01686844395266639
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,4096,1024,0.0058666666348775225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,4096,1024,0.015740444262822468
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,4096,1536,0.03135200010405646
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,4096,768,0.0047235555118984645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,4096,1024,0.029667556285858154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,4096,768,0.014977777997652689
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,4096,512,0.003915555361244413
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,4096,768,0.027687110834651526
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,4096,512,0.014682667122946845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,4096,256,0.003555555724435382
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,4096,512,0.02632177703910404
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,4096,256,0.013760888742076026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,4096,128,0.0032524443748924467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,4096,256,0.02442844377623664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,4096,128,0.013212444053755866
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,4096,64,0.003055111194650332
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,4096,128,0.023256889647907678
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,4096,64,0.013129777378506131
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,4096,32,0.003412444351447953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,4096,32,0.012783110969596438
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3584,16384,0.05227555500136482
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,3584,16384,0.06967910793092516
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3584,12288,0.03000266684426202
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3584,16384,0.03872444563441806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3584,12288,0.04292533463901944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,3584,12288,0.05930311150021023
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3584,10240,0.036497778362698026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3584,10240,0.02637600070900387
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,4096,65536,0.14685155285729304
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,3584,10240,0.05282755692799886
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3584,8192,0.031686223215527005
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3584,7168,0.029835555288526747
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,3584,8192,0.04663199848598904
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3584,8192,0.02305511136849721
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,3584,7168,0.04369422131114536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3584,65536,0.16263288921780056
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3584,6144,0.02742044462098016
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,3584,6144,0.041908442974090576
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3584,5120,0.024901333782407973
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3584,7168,0.020784889658292133
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3584,6144,0.018787556224399142
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,3584,5120,0.03937066594759623
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3584,4096,0.012549333274364471
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3584,5120,0.016903110676341586
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3584,4096,0.02314311100376977
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3584,3584,0.011287111375066968
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3584,3584,0.02186311119132572
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,3584,4096,0.03625155488650004
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3584,3072,0.010336000058386061
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3584,3072,0.02083911167250739
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,3584,3584,0.03598399957021078
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3584,2560,0.008378666308191087
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,3584,3072,0.034119112624062434
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3584,2560,0.0196088883611891
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3584,2048,0.007371555599901412
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,3584,2560,0.033879998657438494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3584,1536,0.006657777975002925
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3584,2048,0.018400000201331247
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3584,1536,0.016386666231685214
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,3584,2048,0.03292444348335266
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3584,1024,0.005467555589146084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,3584,1536,0.03122133347723219
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3584,768,0.0047937776479456155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,3584,65536,0.20552533202701143
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3584,1024,0.015623999966515435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3584,768,0.015004444453451367
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,3584,1024,0.02845422095722622
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3584,512,0.003991111285156674
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,3584,768,0.027951998843087092
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3584,512,0.014726221561431885
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3584,256,0.0034346667428811393
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,3584,512,0.02626133296224806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3584,256,0.013779555757840475
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3584,128,0.003169777699642711
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3584,64,0.00300799993177255
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,3584,256,0.024699555502997503
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,3584,128,0.023215111758973863
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3584,128,0.013145777914259167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3584,32,0.0030666666312350165
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3584,64,0.012902222573757172
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3584,32,0.012824888858530255
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3584,65536,0.13188800546858045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3072,16384,0.05093422200944689
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,3072,16384,0.06826044453514947
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3072,12288,0.026963555150561865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3072,16384,0.03585066729121738
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3072,12288,0.0421377784676022
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3072,10240,0.03530755639076233
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,3072,12288,0.0577768882115682
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,3072,10240,0.051912890540228955
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3072,8192,0.03091288937462701
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3072,10240,0.023951111568344965
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,3072,8192,0.04641244477695889
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3072,7168,0.02830044428507487
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3072,8192,0.021016889148288306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3072,6144,0.0259226659933726
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,3072,7168,0.04353955719206068
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3072,7168,0.01880088945229848
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,3072,6144,0.04160088962978787
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3072,5120,0.015190222197108798
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3072,6144,0.016923555069499545
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3072,5120,0.024682665864626568
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3072,4096,0.010710222025712332
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,3072,5120,0.03958844476275974
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3072,4096,0.022613333331214056
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3072,3584,0.009745777481132084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,3072,4096,0.03577955563863119
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3072,3584,0.021570665968788996
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3072,3072,0.00886399961180157
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,3072,3584,0.03582933213975694
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3072,3072,0.020472000042597454
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3072,2560,0.00810222244924969
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,3072,3072,0.03387644555833604
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3072,2560,0.018788443671332467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3072,2048,0.007303111255168915
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3072,65536,0.15926488240559897
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3072,1536,0.0064693333374129395
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3072,2048,0.01742577718363868
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,3072,2560,0.03382488754060533
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3072,1536,0.016410667035314772
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,3072,2048,0.03206488821241591
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3072,1024,0.004987555659479565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,3072,1536,0.029367112451129492
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3072,1024,0.015511999527613321
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3072,768,0.004403555558787452
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3072,768,0.014932443698247274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3072,512,0.0038195554580953387
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,3072,1024,0.028705779049131606
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,3072,65536,0.1988675594329834
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3072,512,0.014503111441930136
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,3072,768,0.02790400054719713
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3072,256,0.0034204444123639
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,3072,512,0.02641422218746609
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3072,256,0.013668444421556262
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3072,128,0.0031617778456873367
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,3072,256,0.024485333098305598
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3072,128,0.013189333180586496
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3072,64,0.003000000077817175
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,3072,128,0.02302933401531643
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3072,32,0.003080000066094928
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3072,64,0.012855999999576144
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,3072,32,0.012682666381200155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,3072,65536,0.11036622524261475
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2560,16384,0.048575109905666776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,2560,16384,0.0675271087222629
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2560,12288,0.02353422178162469
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2560,12288,0.039367109537124634
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2560,10240,0.034435556994544134
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,2560,12288,0.05738933218850029
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2560,16384,0.02850666642189026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,2560,10240,0.050832887490590416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2560,8192,0.029721776644388836
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2560,10240,0.022308443983395893
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2560,8192,0.01774488886197408
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,2560,8192,0.04534222351180183
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2560,7168,0.027850667635599773
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2560,65536,0.1504275533888075
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2560,6144,0.014429334137174817
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,2560,7168,0.042504890097512134
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2560,6144,0.02593066626124912
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2560,7168,0.015976889265908133
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2560,5120,0.011402666568756104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,2560,6144,0.04062044289377
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2560,5120,0.02385422256257799
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2560,4096,0.011615110768212212
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,2560,5120,0.03865244322352939
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2560,4096,0.021776888105604384
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,2560,65536,0.19545333915286592
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2560,3584,0.010287111004193624
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,2560,4096,0.03547822104560004
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2560,3584,0.02095288866096073
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2560,3072,0.008395555946562026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,2560,3584,0.03567377726236979
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2560,3072,0.01997866729895274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2560,2560,0.007979555262459649
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,2560,3072,0.0337137778600057
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2560,2048,0.007225777539942
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2560,2560,0.018424888451894123
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2560,1536,0.006436444405052397
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2560,1536,0.01682400041156345
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2560,1024,0.00479911102188958
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2560,2048,0.01716888944307963
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,2560,2560,0.03388266762097677
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,2560,2048,0.030722667773564655
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,2560,1536,0.028855111863878038
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2560,1024,0.015302222636010913
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2560,768,0.004249777644872665
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,2560,1024,0.02869866622818841
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2560,768,0.014659555421935188
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2560,512,0.0038284444146686127
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,2560,768,0.02751911183198293
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2560,512,0.014300444059901767
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2560,256,0.00342755557762252
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,2560,512,0.026151110728581745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2560,256,0.013632888595263163
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2560,128,0.00315022220214208
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,2560,256,0.02464977734618717
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2560,128,0.013092444174819522
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,2560,128,0.023000889354281958
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2560,64,0.002983999955985281
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2560,65536,0.09320711427264744
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2560,32,0.0029813332690132987
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2560,64,0.012801777985360889
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2560,32,0.0125928885406918
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2048,12288,0.038605332374572754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2048,16384,0.04796444376309713
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,2048,12288,0.056221332814958364
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,2048,16384,0.06583733028835721
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2048,12288,0.020097777247428894
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2048,10240,0.03292799989382426
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2048,16384,0.024545777175161574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,2048,10240,0.049215111467573375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2048,8192,0.02918133470747206
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2048,10240,0.01779911087618934
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2048,8192,0.015098666151364645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2048,7168,0.01406044430202908
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2048,7168,0.027274666561020747
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,2048,8192,0.04427999920315213
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2048,6144,0.012616000241703458
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,2048,7168,0.04150222076310052
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2048,6144,0.025603555970721777
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2048,5120,0.011157333023018308
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,2048,6144,0.04031555520163642
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2048,5120,0.023652444283167522
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2048,4096,0.010016889207892949
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,2048,5120,0.0384533339076572
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2048,4096,0.021693333983421326
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2048,3584,0.009007111191749573
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2048,65536,0.14876533879174128
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,2048,4096,0.03499644332461887
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2048,3584,0.021021333005693223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2048,3072,0.008282666405042013
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,2048,65536,0.18934222062428793
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,2048,3584,0.03489333391189575
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2048,3072,0.019023999571800232
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2048,2560,0.007663999994595845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,2048,3072,0.03321155574586656
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2048,2560,0.01797066628932953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2048,2048,0.00684533350997501
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,2048,2560,0.03257155418395996
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2048,2048,0.0170462214284473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2048,1536,0.005963555640644497
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,2048,2048,0.030193779203626845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2048,1536,0.016026667422718473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2048,1024,0.004826666580306159
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2048,1024,0.015228443675571017
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,2048,1536,0.028877334462271795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2048,65536,0.074182218975491
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2048,768,0.004139555411206352
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,2048,1024,0.028225776222017076
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2048,768,0.014732443624072604
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2048,512,0.0037813331517908308
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,2048,768,0.02772177755832672
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2048,256,0.003359111025929451
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2048,512,0.014377777775128683
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,2048,512,0.026240888569090102
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2048,256,0.013588444226317935
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2048,128,0.0030542222989930045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,2048,256,0.024430221981472437
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2048,128,0.01294133315483729
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2048,64,0.0028577778074476454
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,2048,128,0.022817777262793645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,2048,32,0.0029973333908451926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2048,64,0.0127368892232577
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,2048,32,0.012675555215941535
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1536,16384,0.04682133264011807
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,1536,16384,0.06098577711317274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1536,12288,0.03797333439191183
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1536,16384,0.020803555846214294
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,1536,12288,0.05307644605636597
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1536,12288,0.015984889533784654
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1536,10240,0.014239999983045789
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1536,10240,0.032413333654403687
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1536,8192,0.012311999996503195
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,1536,10240,0.04711911082267761
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1536,8192,0.028756442997190688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1536,7168,0.011523555550310345
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,1536,8192,0.04227911101447212
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1536,7168,0.026885333988401625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1536,6144,0.01073066641887029
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,1536,7168,0.04032355546951294
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1536,6144,0.025303999582926433
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1536,5120,0.011022222538789114
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,1536,6144,0.03881866733233134
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1536,5120,0.02327555583582984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1536,4096,0.009567110902733272
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,1536,5120,0.037067555718951754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1536,65536,0.14432888560824925
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1536,4096,0.021282666259341772
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,1536,65536,0.17267910639444986
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1536,3584,0.008856000171767341
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,1536,4096,0.03412266572316488
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1536,3072,0.008200888832410177
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1536,3584,0.02036177780893114
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1536,3072,0.018417777286635507
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,1536,3584,0.03424177898301019
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1536,2560,0.007228444847795699
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,1536,2560,0.030976888206270006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,1536,3072,0.03153333399030898
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1536,1536,0.005280888742870755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1536,2560,0.01749955614407857
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1536,2048,0.006312888943486744
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1536,2048,0.016567111015319824
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1536,65536,0.06181244717703926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,1536,2048,0.029823111163245306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1536,1536,0.01590666671593984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1536,1024,0.004304888761705822
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,1536,1536,0.028431110911899145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1536,1024,0.0151555554734336
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1536,768,0.003939555750952827
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,1536,1024,0.028110222684012517
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1536,768,0.014713777436150445
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1536,512,0.003638222399685118
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,1536,768,0.027533332506815594
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1536,512,0.014132445057233175
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1536,256,0.0032888888898822996
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,1536,512,0.02583911187118954
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1536,256,0.013377777404255338
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1536,128,0.002966222249799304
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,1536,256,0.024147555232048035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1536,128,0.013028444515334235
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1536,64,0.0028213332924577924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,1536,128,0.0229759994480345
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1536,32,0.0029022221763928733
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1536,64,0.012728888955381183
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1536,32,0.01255022237698237
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1024,16384,0.043876445955700345
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1024,12288,0.012778666284349231
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,1024,16384,0.05884977844026354
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1024,12288,0.036469333701663546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1024,10240,0.011667555405033959
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1024,16384,0.015495111544926962
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,1024,12288,0.053095112244288124
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1024,10240,0.03216000066863166
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1024,8192,0.01260355528857973
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1024,8192,0.028361777464548748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,1024,10240,0.04727466570006477
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1024,7168,0.011368888947698804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,1024,8192,0.04270577761862013
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1024,65536,0.14230488406287298
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1024,7168,0.02660177813635932
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1024,6144,0.010666666759385003
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,1024,7168,0.03993777765168084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1024,6144,0.024282667371961806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1024,5120,0.009923555784755284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,1024,65536,0.17294043964809844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,1024,5120,0.0369795560836792
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,1024,6144,0.039269334740108915
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1024,5120,0.021869333253966436
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1024,4096,0.009097778134875828
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1024,4096,0.019800888167487252
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1024,3584,0.008827555510732863
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1024,65536,0.04690488841798571
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,1024,4096,0.032947556840048894
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1024,3584,0.0195768889453676
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1024,3072,0.008281778130266402
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,1024,3584,0.03201333350605435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1024,3072,0.018072000808186002
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1024,2560,0.006862222320503659
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,1024,3072,0.030842665168974135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1024,2560,0.01742044422361586
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1024,2048,0.00628355559375551
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,1024,2560,0.030886222918828327
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1024,2048,0.016650666793187458
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1024,1536,0.004978666702906291
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,1024,2048,0.029770665698581274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1024,1536,0.015740444262822468
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1024,1024,0.004300444490379757
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,1024,1536,0.028484443823496502
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1024,1024,0.014875556031862894
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1024,768,0.003926222109132343
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,1024,1024,0.028233776489893597
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1024,768,0.014446222119861178
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1024,512,0.003628444340493944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,1024,768,0.027615111735132005
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1024,512,0.014095111025704278
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1024,256,0.003198222153716617
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,1024,512,0.02568000058333079
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1024,256,0.01330844478474723
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1024,128,0.0029395555870400537
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,1024,256,0.024381332927280005
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1024,128,0.012806221842765808
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1024,64,0.0027679999669392905
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,1024,128,0.02348533272743225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1024,64,0.012642666697502136
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,1024,32,0.0028408887899584244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,1024,32,0.01256444470749961
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,768,16384,0.012623111406962076
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,768,12288,0.012196444802814059
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,768,16384,0.043495111995273165
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,768,12288,0.036213331752353244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,768,16384,0.05671199825074938
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,768,10240,0.010847110715177325
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,768,10240,0.03194933467441135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,768,12288,0.05008888906902737
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,768,8192,0.010228444304731157
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,768,8192,0.028120001157124836
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,768,10240,0.0455057786570655
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,768,65536,0.14078756173451742
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,768,7168,0.00981244444847107
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,768,6144,0.00923822240697013
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,768,7168,0.02666666607062022
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,768,8192,0.04044355617629157
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,768,7168,0.03907377852333917
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,768,6144,0.02363288899262746
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,768,5120,0.008783111141787635
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,768,65536,0.16095111105177137
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,768,6144,0.037837333149380155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,768,5120,0.021934222843911912
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,768,4096,0.00811288919713762
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,768,5120,0.03538666831122504
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,768,65536,0.038485334979163274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,768,4096,0.020066666934225295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,768,3584,0.007841777470376756
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,768,3072,0.018028444714016385
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,768,4096,0.03115644388728672
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,768,3584,0.019259555472267997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,768,3072,0.0075911109646161394
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,768,3584,0.03215288784768846
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,768,2560,0.006487999939256244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,768,3072,0.03004266818364461
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,768,2560,0.01744622157679664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,768,2048,0.005650666852792104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,768,2560,0.030360887447992962
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,768,2048,0.01661066710948944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,768,1536,0.004942222187916438
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,768,2048,0.029490666257010564
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,768,1536,0.015807110402319167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,768,1024,0.0041991112132867174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,768,1536,0.028146667612923518
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,768,1024,0.015067555838161044
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,768,768,0.003899555653333664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,768,1024,0.027894222074084814
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,768,768,0.014594667487674289
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,768,512,0.003530666646030214
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,768,768,0.027071111732059058
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,768,512,0.014172444740931192
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,768,256,0.02398577829202016
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,768,256,0.0031999999450312722
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,768,512,0.0258240004380544
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,768,256,0.013384888569513956
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,768,128,0.0029208889851967492
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,768,128,0.01294844432009591
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,768,64,0.0027840000887711844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,768,128,0.02273955610063341
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,768,32,0.0028008888992998335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,768,64,0.01276000009642707
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,768,32,0.012630222572220696
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,512,16384,0.012061333490742577
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,512,16384,0.04267288910018074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,512,12288,0.01051999959680769
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,512,16384,0.05643110805087619
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,512,12288,0.036032888624403216
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,512,10240,0.009659555223253038
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,512,12288,0.04960977700021532
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,512,65536,0.13936089144812688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,512,10240,0.030910223722457886
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,512,8192,0.008985777695973715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,512,10240,0.04547911220126682
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,512,7168,0.008358221915033128
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,512,65536,0.15794755352867976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,512,8192,0.027134221461084154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,512,8192,0.04006133476893107
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,512,7168,0.025654221574465435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,512,65536,0.02940888868437873
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,512,6144,0.007720000214046902
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,512,7168,0.037672887245814
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,512,6144,0.023702222439977858
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,512,5120,0.007859555383523306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,512,6144,0.03608444333076477
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,512,4096,0.019828443725903828
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,512,5120,0.02156266735659705
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,512,4096,0.007017778025733099
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,512,5120,0.03369066781467862
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,512,3584,0.007507555186748505
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,512,4096,0.030984888474146526
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,512,3584,0.018782221608691745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,512,3072,0.006992888947327931
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,512,3584,0.03137066629197862
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,512,3072,0.018032888571421306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,512,2560,0.006535111202134027
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,512,3072,0.02963288956218296
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,512,2560,0.01696711116366916
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,512,2048,0.005584888988071018
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,512,2560,0.029735998974906072
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,512,2048,0.016346666547987197
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,512,1536,0.004885333279768626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,512,2048,0.02905511193805271
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,512,1536,0.015629332926538255
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,512,1024,0.004176889028814104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,512,1536,0.02769777841038174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,512,1024,0.014793778459231058
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,512,768,0.0038177776667806837
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,512,1024,0.027672888504134283
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,512,768,0.014353776971499125
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,512,512,0.0034506666577524612
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,512,512,0.013988445202509562
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,512,768,0.026745777991082933
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,512,256,0.003091555502679613
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,512,512,0.025501334004931982
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,512,256,0.013218666944238873
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,512,128,0.002891555635465516
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,512,256,0.024100444383091394
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,512,128,0.012857777376969656
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,512,64,0.0027102221631341507
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,512,128,0.02334044377009074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,512,32,0.0027911110470692315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,256,16384,0.008960000342792934
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,512,64,0.012648888760142855
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,512,32,0.012486221889654795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,256,16384,0.041928887367248535
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,256,65536,0.13769600126478407
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,256,12288,0.008089777496125963
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,256,16384,0.05593599875768026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,256,65536,0.15724089410569933
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,256,12288,0.03506044546763102
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,256,10240,0.0075786668393347
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,256,65536,0.01940533353222741
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,256,12288,0.0482666658030616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,256,10240,0.03059200114674038
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,256,8192,0.007176000210973952
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,256,10240,0.04356355468432108
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,256,8192,0.02628622286849552
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,256,7168,0.007035555938879649
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,256,8192,0.03811111052831014
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,256,7168,0.024597333537207708
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,256,6144,0.006711999989218182
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,256,7168,0.03673600157101949
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,256,6144,0.023099554909600153
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,256,5120,0.007111111448870764
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,256,6144,0.03526488939921061
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,256,5120,0.021242666575643752
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,256,4096,0.006700444552633498
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,256,5120,0.03327644533581204
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,256,4096,0.019517333971129525
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,256,3584,0.00720088928937912
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,256,4096,0.030643555853101943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,256,3584,0.018575110369258456
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,256,3072,0.006866666591829724
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,256,3584,0.031125333574083116
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,256,3072,0.017662222186724346
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,256,2560,0.006140444427728653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,256,3072,0.029683556821611192
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,256,2560,0.016940444707870483
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,256,2048,0.005480888817045424
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,256,2048,0.01608000033431583
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,256,2560,0.029808001385794744
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,256,1536,0.004761777818202972
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,256,2048,0.02899200055334303
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,256,1536,0.015352000792821249
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,256,1024,0.004127999974621667
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,256,1536,0.027673777606752183
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,256,1024,0.01462311049302419
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,256,768,0.003710222327046924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,256,1024,0.02754844393995073
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,256,768,0.014160000615649752
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,256,512,0.0034097778714365433
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,256,768,0.026847999956872728
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,256,512,0.013857777747843
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,256,256,0.0030382221771611106
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,256,512,0.025426665941874187
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,256,256,0.013162666724787818
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,256,128,0.0028515555378463534
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,256,256,0.02422133253680335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,256,128,0.012612444659074148
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,256,64,0.0026862221873468826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,256,32,0.012536889149083031
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,256,128,0.022541332576009963
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,256,32,0.00267999991774559
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,256,64,0.012609778179062737
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,128,65536,0.014526221487257214
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,128,16384,0.007039111521508958
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,128,65536,0.13562133577134874
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,128,16384,0.0419680012596978
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,128,12288,0.006963555183675554
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,128,65536,0.15741777420043945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,128,12288,0.03406577640109592
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,128,16384,0.05564177698559231
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,128,10240,0.007050666544172499
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,128,12288,0.048298666874567665
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,128,10240,0.02990666694111294
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,128,8192,0.0069777775141927935
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,128,10240,0.04314577910635206
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,128,8192,0.0262435558769438
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,128,7168,0.006779555645253923
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,128,8192,0.03760533200369941
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,128,7168,0.024717332588301763
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,128,6144,0.006640000061856375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,128,7168,0.03648799988958571
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,128,6144,0.023112000690566167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,128,5120,0.006993778049945831
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,128,6144,0.03514844510290358
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,128,4096,0.006599999964237213
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,128,5120,0.02109333376089732
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,128,4096,0.019500444332758587
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,128,5120,0.0331582228342692
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,128,3584,0.007055111229419708
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,128,4096,0.030647112263573542
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,128,3584,0.018498667412334018
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,128,3072,0.006691555596060223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,128,3584,0.030943999687830608
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,128,3072,0.017607111069891188
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,128,2560,0.00609688874748018
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,128,3072,0.029873778422673542
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,128,2560,0.016902221573723685
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,128,2048,0.005402666827042897
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,128,2560,0.029567112525304157
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,128,2048,0.01603555513752831
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,128,1536,0.0047457776963710785
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,128,2048,0.028731554746627808
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,128,1536,0.01536799967288971
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,128,1024,0.0040960001448790235
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,128,1536,0.027394667267799377
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,128,1024,0.014548444085650973
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,128,768,0.003699555579158995
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,128,1024,0.027387556102540758
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,128,768,0.014104000396198697
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,128,512,0.003354666754603386
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,128,768,0.026624000734753076
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,128,512,0.013752000199423896
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,128,256,0.003003555453485913
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,128,512,0.02516266703605652
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,128,256,0.013163555827405719
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,128,128,0.002838222309947014
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,128,256,0.02386755579047733
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,128,128,0.012703110774358114
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,128,64,0.0026862221873468826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,96,128,128,0.022699554761250813
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,128,64,0.012511110968059964
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,128,32,0.0027146666414207886
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,128,32,0.012503110700183444
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,64,65536,0.012744888663291931
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,64,16384,0.006671999891599019
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,64,65536,0.13598488436804876
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,64,16384,0.04148711098564996
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,64,12288,0.006586666736337874
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,64,12288,0.033599111768934466
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,64,10240,0.006812444577614467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,64,10240,0.029669334491093952
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,64,8192,0.006636444479227066
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,64,7168,0.006745777610275481
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,64,6144,0.022993778189023335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,64,8192,0.026457778281635706
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,64,6144,0.006495999793211619
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,64,7168,0.02464888824356927
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,64,5120,0.006903111106819577
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,64,4096,0.006467555546098285
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,64,5120,0.021055110626750525
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,64,4096,0.01945333348380195
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,64,3584,0.006948444578382704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,64,3072,0.006616000086069107
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,64,3584,0.01852888862291972
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,64,2560,0.006171555568774541
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,64,3072,0.017685333887736004
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,64,2048,0.005405333307054307
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,64,2560,0.016857778032620747
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,64,2048,0.016139555308553908
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,64,1536,0.0047271110945277745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,64,1536,0.015267555912335714
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,64,1024,0.004025777594910728
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,64,768,0.0036720000207424164
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,64,1024,0.014502222339312235
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,64,768,0.014133334159851074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,64,512,0.003334222154484855
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,64,256,0.002982222164670626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,64,512,0.01366311146153344
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,64,256,0.013032889200581444
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,64,128,0.002809777855873108
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,64,64,0.002630222174856398
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,64,128,0.01258488910065757
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,64,32,0.002624888800912433
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,64,64,0.012433778080675336
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,64,32,0.012292444705963135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,32,65536,0.01128000020980835
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,32,16384,0.006695999867386288
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,32,65536,0.13515467113918728
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,32,16384,0.04023911224471198
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,32,12288,0.006521777974234686
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,32,10240,0.006711111300521427
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,32,12288,0.03382133444150289
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,32,8192,0.006581333362393909
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,32,10240,0.0299217767185635
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,32,7168,0.006529777828190062
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,32,8192,0.026421333352724712
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,32,7168,0.024639111426141527
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,32,6144,0.0063635553750726916
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,32,6144,0.02275111112329695
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,32,5120,0.006808888994985157
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,32,4096,0.00638755535085996
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,32,5120,0.02109777761830224
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,32,3584,0.006877333339717653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,32,4096,0.019541333119074505
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,32,3584,0.018544000056054857
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,32,3072,0.006495111104514863
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,32,3072,0.017621333400408428
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,32,2560,0.006054222169849608
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,32,2560,0.016905777984195285
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,32,2048,0.0053591111467944244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,32,1536,0.0047217777205838096
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,32,512,0.0032835555159383347
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,32,768,0.014194667339324951
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,32,2048,0.01609511176745097
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,32,1024,0.004009777886999978
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,32,1536,0.015256888336605497
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,32,768,0.0036106668412685394
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,32,1024,0.014488000008794995
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,32,256,0.0029475556479560006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,32,512,0.013674666484196981
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,32,256,0.012993777791659037
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,32,128,0.002761777697337998
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,32,128,0.012681778106424542
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,32,64,0.002601777720782492
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,96,32,32,0.002597333242495855
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,32,64,0.012558221817016602
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,96,32,32,0.012255111502276527
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,65536,10240,0.2118142180972629
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,65536,12288,0.25852621926201713
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,65536,16384,0.3306106726328532
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,65536,12288,0.25630754894680446
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,65536,10240,0.35834309789869523
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,65536,10240,0.21981422106424967
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,65536,16384,0.32483379046122235
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,65536,8192,0.1859297752380371
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,65536,12288,0.42877689997355145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,65536,8192,0.2857511043548584
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,65536,7168,0.1642248895433214
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,65536,8192,0.18057421843210855
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,65536,16384,0.5676870875888401
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,65536,6144,0.14328800307379827
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,65536,7168,0.2547644509209527
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,65536,7168,0.16090577178531223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,65536,6144,0.14726489120059544
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,65536,6144,0.2189697821935018
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,65536,5120,0.12146933873494466
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,65536,5120,0.1284266710281372
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,65536,4096,0.10094844632678562
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,65536,5120,0.18471111191643608
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,65536,4096,0.1483751138051351
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,65536,3584,0.08990933497746785
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,65536,4096,0.11084889041052924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,65536,3072,0.0789511137538486
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,65536,3584,0.10055555237664117
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,65536,3584,0.1298417780134413
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,65536,2560,0.06515466504626803
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,65536,3072,0.11484532886081272
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,65536,3072,0.08997599946127997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,65536,2048,0.05543555484877693
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,65536,1536,0.04455733299255371
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,65536,2560,0.09511199924680923
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,65536,1536,0.062223109934065074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,65536,2560,0.08026488622029622
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,65536,2048,0.07368622223536174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,65536,2048,0.07732622490988837
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,65536,1024,0.037695109844207764
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,65536,1536,0.06070578098297119
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,65536,1024,0.05305777655707466
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,65536,256,0.014289778139856128
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,65536,768,0.03114933437771267
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,65536,512,0.027186666925748188
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,65536,768,0.04751822352409363
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,65536,256,0.023078221413824294
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,65536,512,0.04194044404559665
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,65536,128,0.00942844483587477
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,65536,768,0.03363022208213806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,65536,512,0.025482666161325242
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,65536,256,0.036248889234330915
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,65536,64,0.022322666313913133
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,65536,128,0.021031111478805542
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,65536,64,0.007931555310885111
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,65536,128,0.03112799922625224
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,65536,32,0.006741333338949416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,65536,1024,0.04338666796684265
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,65536,32,0.023339556323157415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,16384,12288,0.0811671084827847
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,16384,16384,0.10601155625449286
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,16384,16384,0.12610133488972983
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,16384,12288,0.09821689128875732
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,16384,12288,0.09896000226338704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,16384,16384,0.12840710745917425
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,16384,10240,0.0694942209455702
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,16384,10240,0.08333244588640001
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,16384,8192,0.060699555608961314
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,16384,10240,0.0870293312602573
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,16384,8192,0.07250844770007663
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,16384,7168,0.05428711242145962
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,16384,7168,0.06774044699139066
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,16384,8192,0.06731644603941175
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,16384,7168,0.05917688873079088
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,16384,6144,0.04830933279461331
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,16384,6144,0.05154044429461161
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,16384,5120,0.04267200165324741
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,16384,6144,0.06299377812279595
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,16384,5120,0.05575022432539198
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,16384,65536,0.3587324354383681
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,16384,4096,0.0369742214679718
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,16384,4096,0.05099022057321337
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,16384,5120,0.04406755500369602
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,16384,4096,0.03654399845335219
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,16384,3584,0.0351164440313975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,16384,3584,0.04847822255558438
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,16384,3584,0.032163557079103254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,16384,2560,0.02848177817132738
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,16384,3072,0.032016889916525945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,16384,3072,0.047863112555609814
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,16384,3072,0.028287112712860107
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,16384,2560,0.024459555745124817
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,16384,2560,0.04425422350565592
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,16384,1536,0.02253155575858222
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,16384,2048,0.02563200063175625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,16384,2048,0.04039644532733493
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,16384,2048,0.020269332660569083
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,16384,1536,0.03801244497299194
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,16384,1024,0.020024889045291476
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,16384,1536,0.01792266633775499
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,16384,1024,0.03449866506788466
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,16384,768,0.009642666412724389
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,16384,1024,0.011838222543398539
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,16384,768,0.018948444061809115
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,16384,512,0.007647111184067196
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,16384,768,0.03304800060060289
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,16384,512,0.017816889617178176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,16384,256,0.005616000129116907
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,16384,512,0.03083911206987169
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,16384,256,0.015059555570284525
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,16384,128,0.004071111066473855
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,16384,256,0.025934222671720717
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,16384,128,0.014134221606784396
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,16384,64,0.0038168889780839286
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,16384,128,0.024059555596775476
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,16384,32,0.0038373333712418876
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,16384,64,0.014533332652515836
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,16384,32,0.01497155593501197
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,16384,65536,0.3945431179470486
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,12288,16384,0.09103466404808892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,12288,16384,0.10432622167799209
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,12288,12288,0.06936355431874593
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,12288,12288,0.08531910843319362
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,12288,16384,0.10937955644395615
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,12288,12288,0.084297776222229
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,12288,10240,0.06102044714821709
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,12288,10240,0.0679360032081604
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,12288,10240,0.07367200321621366
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,12288,8192,0.05240089032385084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,12288,8192,0.05503733290566338
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,12288,8192,0.06424799892637464
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,12288,65536,0.31721067428588867
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,12288,7168,0.0473911099963718
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,12288,7168,0.06042933464050293
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,12288,6144,0.04321066538492838
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,12288,6144,0.0565297802289327
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,12288,7168,0.04756888747215271
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,12288,6144,0.04147288865513272
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,12288,65536,0.329893324110243
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,12288,5120,0.03866844375928243
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,12288,5120,0.035972442891862653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,12288,4096,0.033985777033699885
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,12288,5120,0.05123555660247803
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,12288,4096,0.047726223866144814
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,12288,3584,0.02610933283964793
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,12288,3584,0.03154311246342129
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,12288,4096,0.02956355611483256
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,12288,3584,0.04500444398985969
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,12288,3072,0.029343999094433252
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,12288,3072,0.04383377896414863
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,12288,2560,0.026573333475324843
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,12288,3072,0.02314311100376977
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,12288,2048,0.023610666394233704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,12288,2560,0.040404445595211454
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,12288,2048,0.03762844536039565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,12288,1536,0.02125155594613817
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,12288,2560,0.020311110549502902
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,12288,1024,0.010148444109492833
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,12288,1536,0.03660088777542114
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,12288,1536,0.013433777623706393
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,12288,768,0.008562666674455007
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,12288,2048,0.016945777667893302
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,12288,1024,0.01941333380010393
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,12288,1024,0.03362755642996894
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,12288,512,0.006953777538405524
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,12288,768,0.032532443602879844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,12288,768,0.018552000323931377
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,12288,512,0.016364443633291457
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,12288,256,0.015052444405025907
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,12288,256,0.0047315553658538395
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,12288,512,0.029992000924216375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,12288,128,0.003918222255176968
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,12288,256,0.02568977740075853
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,12288,128,0.014038221703635322
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,12288,64,0.0035955554081334006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,12288,128,0.023971555961502924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,12288,64,0.01442488862408532
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,12288,32,0.0036275556517971885
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,12288,32,0.014858666393491956
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,16384,65536,0.570741335550944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,10240,16384,0.08612889051437378
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,10240,16384,0.09426133500205146
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,10240,12288,0.06611022022035387
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,10240,16384,0.09775377644432916
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,10240,12288,0.07688799831602308
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,10240,12288,0.0779048866695828
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,12288,65536,0.41240978240966797
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,10240,10240,0.05686577823426989
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,10240,8192,0.050141334533691406
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,10240,10240,0.06694310903549194
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,10240,10240,0.07259911298751831
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,10240,8192,0.04474577638838026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,10240,8192,0.0606364475356208
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,10240,7168,0.044256889157825045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,10240,6144,0.0409822232193417
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,10240,65536,0.29328711827596027
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,10240,7168,0.039114667309655085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,10240,7168,0.057542224725087486
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,10240,6144,0.03462666604253981
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,10240,6144,0.053951111104753285
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,10240,5120,0.036665777365366616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,10240,5120,0.049119111564424306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,10240,4096,0.032403555181291364
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,10240,5120,0.029768887493345473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,10240,4096,0.04602311054865519
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,10240,3584,0.03009422289000617
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,10240,4096,0.024626667300860088
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,10240,3584,0.02186133298608992
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,10240,3072,0.027871999475691054
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,10240,3584,0.043043557140562266
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,10240,2560,0.024871110916137695
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,10240,3072,0.041687998506757945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,10240,2560,0.03885777791341146
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,10240,3072,0.019478221734364826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,10240,2048,0.022408889399634466
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,10240,1536,0.01148977792925305
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,10240,1536,0.020672000116772123
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,10240,2048,0.03641511003176371
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,10240,2048,0.014724445011880664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,10240,1024,0.008962666822804345
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,10240,1536,0.03549244337611728
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,10240,1024,0.018823999497625563
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,10240,2560,0.017430222696728177
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,10240,768,0.00775822252035141
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,10240,1024,0.0333280000421736
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,10240,768,0.01798488861984677
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,10240,512,0.006613333192136552
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,10240,768,0.03174577818976508
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,10240,512,0.015957333975368075
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,10240,256,0.004451555510361989
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,10240,512,0.02793866727087233
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,10240,256,0.014977777997652689
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,10240,128,0.003920000046491623
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,10240,256,0.026455110973782007
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,10240,128,0.014175999495718213
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,10240,64,0.0034933334423436057
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,10240,128,0.02374222212367588
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,10240,32,0.003506666670242945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,10240,64,0.01426666643884447
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,10240,32,0.014771555860837301
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,10240,65536,0.3046915531158447
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,8192,16384,0.06549244456821017
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,8192,12288,0.05096977618005541
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,8192,16384,0.07400621970494588
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,8192,16384,0.0862319999270969
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,8192,12288,0.05824355284372965
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,10240,65536,0.3634675608740912
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,8192,12288,0.06863911284340753
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,8192,10240,0.044482668240865074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,8192,10240,0.04401510953903198
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,8192,10240,0.061271111170450844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,8192,8192,0.036136888795428805
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,8192,8192,0.03975733452373081
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,8192,7168,0.03613244493802389
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,8192,8192,0.053386665052837796
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,8192,7168,0.05202844407823351
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,8192,65536,0.21193688445621064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,8192,7168,0.03492177857293023
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,8192,6144,0.033366223176320396
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,8192,6144,0.02797244323624505
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,8192,6144,0.048941334088643394
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,8192,5120,0.03054844339688619
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,8192,5120,0.04519377814398872
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,8192,4096,0.027360000544124182
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,8192,5120,0.024192000428835552
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,8192,4096,0.020768000019921195
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,8192,3584,0.025216889050271776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,8192,3584,0.039672000540627375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,8192,4096,0.041780445310804576
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,8192,3072,0.02358666724628872
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,8192,3072,0.03916977842648824
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,8192,65536,0.25488800472683376
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,8192,3584,0.018385777870814007
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,8192,2560,0.02164622147878011
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,8192,3072,0.01610844499535031
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,8192,2560,0.03687289026048448
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,8192,2048,0.020429333051045735
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,8192,1536,0.00958933350112703
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,8192,2048,0.03460888730155097
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,8192,2560,0.014101333088344999
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,8192,1536,0.018936889039145578
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,8192,1024,0.007540444533030192
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,8192,2048,0.011533333195580376
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,8192,1024,0.017668444249365065
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,8192,768,0.00675022229552269
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,8192,1536,0.034632888105180525
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,8192,768,0.015604444675975375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,8192,1024,0.032327112224366926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,8192,512,0.005559111220969095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,8192,768,0.030736890104081895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,8192,512,0.015027556154463025
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,8192,256,0.003937777959638172
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,8192,512,0.027803556786643133
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,8192,128,0.003527111063400904
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,8192,256,0.014002665877342224
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,8192,256,0.024961777859263953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,8192,128,0.013355555633703867
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,8192,64,0.0032835555159383347
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,8192,128,0.023577777875794306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,8192,64,0.01332622187005149
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,8192,32,0.0034684443639384378
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,8192,32,0.01369866645998425
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,7168,16384,0.06172000037299263
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,7168,16384,0.08285155561235216
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,7168,12288,0.045372443066702954
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,7168,12288,0.048420445786582104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,7168,12288,0.06679022312164307
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,7168,16384,0.06745244397057427
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,7168,10240,0.04249688982963562
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,7168,10240,0.039255999856524997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,7168,10240,0.060157332155439586
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,7168,8192,0.03741777936617533
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,7168,8192,0.05250044332610237
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,7168,8192,0.032495998673968844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,7168,7168,0.03415288858943515
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,7168,65536,0.20063643985324434
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,7168,7168,0.02832711074087355
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,7168,7168,0.050795555114746094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,8192,65536,0.2744862238566081
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,7168,6144,0.03211911188231574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,7168,6144,0.04767910970581902
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,7168,5120,0.029147555430730183
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,7168,6144,0.024899555577172175
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,7168,4096,0.026200888885392085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,7168,5120,0.021689777572949726
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,7168,65536,0.259680880440606
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,7168,5120,0.04410755634307861
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,7168,4096,0.0184879998366038
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,7168,3584,0.02402488887310028
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,7168,4096,0.040878223048316106
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,7168,3072,0.02237777743074629
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,7168,3584,0.03899822301334805
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,7168,3072,0.03846844368510776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,7168,2560,0.020979555116759405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,7168,3584,0.01623288955953386
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,7168,2560,0.03696711195839776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,7168,2048,0.010470222267839642
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,7168,2048,0.01997244523631202
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,7168,3072,0.014336888988812765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,7168,1536,0.008611555728647444
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,7168,1536,0.018365333477656048
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,7168,2560,0.012619555824332766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,7168,2048,0.03479199939303928
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,7168,1024,0.007065777977307637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,7168,1024,0.03296800123320685
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,7168,1536,0.034816887643602155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,7168,1024,0.01700088878472646
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,7168,768,0.006272000157170826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,7168,768,0.015468445089128284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,7168,512,0.01499644418557485
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,7168,512,0.004972444640265571
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,7168,768,0.02968888812594944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,7168,256,0.0037368887828456033
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,7168,512,0.02716444432735443
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,7168,256,0.01389244364367591
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,7168,128,0.003439999909864532
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,7168,256,0.025591111845440332
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,7168,128,0.01331822243001726
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,7168,64,0.0032382222513357797
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,7168,128,0.023583110835817125
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,7168,64,0.013228444589508904
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,7168,32,0.0033191111352708605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,7168,32,0.013722666435771518
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,6144,16384,0.05795555644565158
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,6144,16384,0.08426400025685628
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,6144,12288,0.04006133476893107
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,7168,65536,0.26183822419908315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,6144,16384,0.059991108046637646
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,6144,12288,0.04571733209821913
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,6144,12288,0.06869955857594807
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,6144,10240,0.04114044374889798
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,6144,10240,0.06165866719351875
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,6144,10240,0.03467288944456313
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,6144,8192,0.03594311078389486
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,6144,8192,0.028463999430338543
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,6144,8192,0.05389599998792013
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,6144,7168,0.03323910964859857
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,6144,7168,0.050999999046325684
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,6144,7168,0.025116443634033203
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,6144,65536,0.18438133928510878
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,6144,6144,0.03030844529469808
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,6144,5120,0.028117332193586562
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,6144,6144,0.048082666264639966
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,6144,6144,0.022241777843899194
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,6144,5120,0.043783999151653714
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,6144,4096,0.025450666745503742
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,6144,5120,0.01938311093383365
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,6144,4096,0.04111288984616598
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,6144,3584,0.014454222387737699
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,6144,3584,0.023335999912685815
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,6144,4096,0.016378665963808697
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,6144,3584,0.03819110989570618
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,6144,3072,0.021888888544506494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,6144,2560,0.011108444796668159
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,6144,2560,0.020502222908867728
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,6144,3072,0.038021332687801786
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,6144,2048,0.009528000321653154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,6144,3072,0.012645333177513547
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,6144,2560,0.03611022233963013
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,6144,2048,0.019818666908476088
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,6144,1536,0.01819022165404426
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,6144,1536,0.00796977761718962
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,6144,2048,0.033957332372665405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,6144,1024,0.0064195555945237475
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,6144,1024,0.01605599953068627
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,6144,1536,0.03393066591686673
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,6144,768,0.005992888990375731
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,6144,512,0.004289777742491828
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,6144,1024,0.03159911102718777
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,6144,768,0.015311999453438653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,6144,512,0.014723555909262763
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,6144,768,0.028406222661336262
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,6144,256,0.0037706668178240457
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,6144,256,0.013950222068362765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,6144,512,0.026403556267420452
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,6144,128,0.003419555516706573
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,6144,256,0.024596444434589807
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,6144,128,0.013090666797426013
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,6144,64,0.0031848889258172777
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,6144,128,0.023440000083711412
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,6144,32,0.0032524443748924467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,6144,64,0.013248888982666863
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,6144,32,0.013463999662134381
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,6144,65536,0.2557004557715522
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,5120,16384,0.05386666787995232
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,5120,16384,0.08296533425649007
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,5120,16384,0.051775111092461475
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,5120,12288,0.04332000017166138
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,5120,12288,0.04110044572088454
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,5120,12288,0.06350311305787829
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,5120,10240,0.036130666732788086
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,5120,10240,0.03846933444341024
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,5120,10240,0.0584684411684672
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,5120,65536,0.1714222166273329
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,5120,8192,0.03428355521625943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,5120,8192,0.030647112263573542
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,6144,65536,0.2062657806608412
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,5120,7168,0.03145066565937466
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,5120,8192,0.05067822337150574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,5120,7168,0.022258667482270136
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,5120,6144,0.02963288956218296
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,5120,6144,0.04522844486766391
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,5120,65536,0.23913423220316568
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,5120,5120,0.025647999511824712
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,5120,7168,0.04832355512513054
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,5120,5120,0.041526221566730075
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,5120,4096,0.023002665903833177
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,5120,6144,0.01958755486541324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,5120,3584,0.02257866660753886
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,5120,4096,0.03901066713862949
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,5120,5120,0.017196445001496207
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,5120,4096,0.014375999569892883
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,5120,3072,0.01128088848458396
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,5120,3584,0.03734222385618422
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,5120,3072,0.02142222225666046
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,5120,3584,0.012820444173283048
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,5120,2560,0.02012977831893497
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,5120,2560,0.010065777434243096
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,5120,3072,0.0366924438211653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,5120,2048,0.008551999926567078
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,5120,2560,0.035255998373031616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,5120,2048,0.01940000057220459
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,5120,1536,0.0074648890230390764
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,5120,2048,0.03348711133003235
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,5120,1536,0.018079110317760043
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,5120,1024,0.0061048890153567
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,5120,1536,0.033203555477990046
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,5120,1024,0.01592088904645708
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,5120,768,0.005398222141795688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,5120,1024,0.03124622172779507
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,5120,768,0.015276443627145557
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,5120,512,0.0041271112859249115
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,5120,768,0.028423110644022625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,5120,512,0.014711111783981323
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,5120,256,0.003640000190999773
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,5120,512,0.026307556364271376
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,5120,256,0.013754666679435305
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,5120,128,0.003370666669474708
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,5120,256,0.024696000748210486
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,5120,128,0.013179555535316467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,5120,64,0.0031395554542541504
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,5120,128,0.02330133318901062
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,5120,64,0.013179555535316467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,5120,32,0.003167111012670729
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,5120,32,0.013471110827393003
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,5120,65536,0.1831893391079373
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,4096,16384,0.052914665804968945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,4096,12288,0.042083554797702365
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,4096,16384,0.07665333482954237
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,4096,12288,0.03388533327314589
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,4096,16384,0.040949334700902305
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,4096,12288,0.061153776115841336
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,4096,10240,0.03714933329158359
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,4096,8192,0.03277422322167291
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,4096,10240,0.05449600021044413
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,4096,10240,0.030677331818474665
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,4096,7168,0.029471112622155085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,4096,8192,0.024762666887707178
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,4096,7168,0.04561333192719353
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,4096,6144,0.027287110686302185
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,4096,8192,0.04857244425349765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,4096,6144,0.04295555419392056
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,4096,7168,0.02254666719171736
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,4096,5120,0.02501511077086131
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,4096,6144,0.020395555429988436
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,4096,4096,0.012849777936935425
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,4096,5120,0.04014755619896783
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,4096,4096,0.022895110978020564
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,4096,5120,0.018030222919252183
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,4096,4096,0.037956443097856306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,4096,3584,0.011699555648697747
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,4096,3584,0.022026666336589392
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,4096,3072,0.010122666756312052
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,4096,3584,0.03601866629388597
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,4096,3072,0.020647111866209242
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,4096,2560,0.008848889006508721
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,4096,2560,0.01979111135005951
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,4096,3072,0.035712000396516584
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,4096,2048,0.008160889148712158
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,4096,2560,0.03461511267556085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,4096,2048,0.018943111101786297
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,4096,1536,0.006730666591061487
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,4096,65536,0.1648791101243761
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,4096,1536,0.01627288924323188
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,4096,65536,0.21778578228420684
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,4096,2048,0.032905777295430504
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,4096,1024,0.005903111149867375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,4096,1536,0.03244977858331468
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,4096,1024,0.015671999918089975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,4096,768,0.004527111020353106
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,4096,512,0.0039022221333450745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,4096,1024,0.028179556131362915
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,4096,768,0.01517599986659156
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,4096,768,0.027477333943049114
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,4096,512,0.01462044484085507
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,4096,256,0.0034942221310403612
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,4096,512,0.026269333230124578
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,4096,256,0.01351466692156262
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,4096,128,0.003175111073586676
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,4096,256,0.0244186669588089
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,4096,128,0.013153778182135688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,4096,64,0.003071111109521654
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,4096,128,0.02316355539692773
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,4096,64,0.012819555898507437
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,4096,32,0.0032231110251612137
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,4096,32,0.013287111288971372
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,4096,65536,0.14351377222273085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3584,16384,0.05028177632225884
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,3584,16384,0.07154044177797106
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3584,12288,0.040264887942208186
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3584,16384,0.03713599840799967
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,3584,12288,0.05983466572231717
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3584,12288,0.029172443681293066
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3584,10240,0.0351217786471049
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,3584,10240,0.05531911055246989
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3584,8192,0.032489776611328125
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3584,10240,0.025987555583318073
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,3584,8192,0.046880000167422824
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3584,7168,0.02881244487232632
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,3584,7168,0.045778665277693004
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3584,8192,0.022582223018010456
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3584,6144,0.02680444386270311
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3584,65536,0.15515466531117758
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3584,7168,0.020319999919997323
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3584,6144,0.018223999275101554
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3584,5120,0.024163555767801072
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,3584,6144,0.04219733344184028
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3584,4096,0.011356444822417365
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,3584,5120,0.04033155573738946
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3584,4096,0.022011554903454248
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3584,3584,0.010064000056849586
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3584,5120,0.016417778200573392
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,3584,4096,0.03752533263630337
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3584,3584,0.021405332618289526
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3584,3072,0.009335999687512716
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,3584,3584,0.03566933340496487
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,3584,3072,0.035417778624428645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3584,3072,0.020631111330456205
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3584,2560,0.008359111017651027
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3584,2560,0.019542222221692402
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3584,2048,0.007353777686754863
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,3584,2560,0.03416266706254747
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3584,2048,0.01831999917825063
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3584,1536,0.006605333338181178
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,3584,2048,0.032567110326555036
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3584,1536,0.01629688839117686
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,3584,1536,0.03153955605294969
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3584,1024,0.005746666755941179
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3584,1024,0.01552177800072564
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3584,768,0.004402666870090696
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,3584,65536,0.20287822352515328
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,3584,1024,0.028386665715111628
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3584,768,0.015036443869272867
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3584,512,0.0038240001433425476
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,3584,768,0.027494221925735474
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3584,512,0.014583999911944071
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,3584,512,0.02602311140961117
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3584,256,0.0033902221669753394
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3584,256,0.013613333304723104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3584,128,0.003126222226354811
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,3584,256,0.024170666933059692
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3584,128,0.01315822203954061
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3584,64,0.002986666642957263
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,3584,128,0.023207111491097346
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3584,32,0.0030088888274298776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3584,64,0.012958221965365939
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3584,32,0.013236444857385425
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3072,16384,0.04913155568970574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3072,12288,0.02571822206179301
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,3072,16384,0.07312089204788208
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3072,16384,0.03189511100451151
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3072,12288,0.039022223816977605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3072,10240,0.034663110971450806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3584,65536,0.13225689199235705
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,3072,12288,0.060324443711174854
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,3072,10240,0.05305155449443393
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3072,8192,0.030846221579445735
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,3072,8192,0.04629066586494446
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3072,7168,0.02851111027929518
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3072,10240,0.0229066660006841
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3072,8192,0.020119110743204754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,3072,7168,0.04478489028082954
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3072,6144,0.026032888227038916
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3072,7168,0.017830222845077515
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3072,65536,0.1530088848537869
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3072,5120,0.01423911088042789
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,3072,6144,0.04163288738992479
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3072,5120,0.023945776952637568
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3072,4096,0.012551111479600271
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,3072,5120,0.03917955689960056
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3072,6144,0.016121778223249648
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3072,4096,0.021680000755521987
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3072,3584,0.011764444410800934
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,3072,4096,0.03734222385618422
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3072,3584,0.021394666698243883
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3072,3072,0.011129777464601727
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,3072,3584,0.035615109735065036
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3072,3072,0.02030933399995168
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,3072,65536,0.20418311489952934
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3072,2560,0.008004444340864817
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,3072,3072,0.035439110464519925
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3072,2560,0.018791110979186166
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3072,2048,0.007262222468852997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,3072,2560,0.03475644522243076
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3072,2048,0.017163554827372234
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3072,1536,0.0064177778032090925
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,3072,2048,0.031522668070263334
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3072,1536,0.01603911154799991
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3072,1024,0.00562755556570159
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,3072,1536,0.03025777803526984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3072,1024,0.015405333704418607
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3072,768,0.004567999806669024
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,3072,1024,0.028471999698215063
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3072,768,0.014814222852389017
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3072,512,0.0038053331275780997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,3072,768,0.027356444133652583
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3072,512,0.014461333552996317
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3072,256,0.0034026667061779234
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,3072,512,0.026280000805854797
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3072,256,0.013560000393125745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3072,128,0.0031066667288541794
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,3072,256,0.02426311042573717
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3072,65536,0.10789599683549668
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3072,128,0.013070222404268054
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3072,64,0.0029813332690132987
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,3072,128,0.023203555080625746
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3072,64,0.012812444733248817
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,3072,32,0.0030622221529483795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,3072,32,0.01315822203954061
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2560,12288,0.03731733229425218
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2560,16384,0.04722933305634392
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,2560,12288,0.05877333217196994
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,2560,16384,0.0724088880750868
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2560,16384,0.028205331828859117
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2560,12288,0.02264444364441766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2560,10240,0.033055110110176936
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,2560,10240,0.05111555589569939
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2560,10240,0.020117332537968952
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2560,8192,0.029134223858515423
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,2560,8192,0.04698400033844841
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2560,7168,0.02719288898838891
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2560,8192,0.017120000388887193
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,2560,7168,0.043812443812688194
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2560,6144,0.014109333356221518
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2560,6144,0.02481511069668664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2560,5120,0.012852444416946836
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2560,7168,0.015799999237060547
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,2560,6144,0.041514668199751116
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2560,5120,0.023228444986873206
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2560,4096,0.011699555648697747
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2560,65536,0.14376800590091282
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,2560,5120,0.03920355439186096
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2560,4096,0.02163022259871165
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2560,3584,0.009496889180607265
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,2560,4096,0.037083556254704796
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2560,3584,0.020670221911536325
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2560,3072,0.008822222550710043
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,2560,3584,0.03529422150717841
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2560,3072,0.0196115556690428
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,2560,65536,0.1925599972407023
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2560,2560,0.007813333637184566
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,2560,3072,0.03501510951254103
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2560,2560,0.017851556340853374
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2560,2048,0.00721866637468338
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,2560,2560,0.03385600116517808
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2560,2048,0.017315554949972365
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2560,1536,0.00629688882165485
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,2560,2048,0.030458665556377832
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2560,1536,0.01604711181587643
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2560,1024,0.004853333450025982
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,2560,1536,0.02995733420054118
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2560,1024,0.015213333898120456
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2560,512,0.0037671112351947357
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2560,768,0.004570666700601578
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,2560,1024,0.02800977892345852
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2560,768,0.014710222681363424
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,2560,768,0.02716444432735443
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2560,512,0.014282666974597506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2560,256,0.0033804445217053094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,2560,512,0.02573777735233307
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2560,65536,0.09163288937674628
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2560,256,0.013409777647919126
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2560,128,0.003119999956753519
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,2560,256,0.024505777491463557
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2560,128,0.013013333082199097
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2560,64,0.0029288888391521242
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,2560,128,0.023009777069091797
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2560,64,0.012808889150619507
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2560,32,0.0029884444342719186
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2560,32,0.013334222137928009
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2048,16384,0.04554577668507894
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,2048,16384,0.06884266932805379
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2048,12288,0.036045332749684654
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2048,12288,0.019332443674405415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2048,16384,0.023465777436892193
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2048,10240,0.0325546662012736
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,2048,12288,0.05515199899673462
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,2048,10240,0.050498667690489024
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2048,10240,0.016941333810488384
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2048,8192,0.028455999162462022
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2048,7168,0.013482666677898832
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,2048,8192,0.04472800095876058
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2048,8192,0.015096889601813423
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2048,7168,0.027058665951093037
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2048,65536,0.14024355676439074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2048,6144,0.012327999704413943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2048,6144,0.025407999753952026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,2048,7168,0.04284888837072584
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2048,5120,0.011528888510333167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,2048,6144,0.04064177804523044
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2048,5120,0.023081777824295893
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2048,4096,0.010435555544164447
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,2048,65536,0.1918755504820082
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,2048,5120,0.03808177842034234
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2048,4096,0.02124711043304867
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2048,3584,0.009214222431182861
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2048,3584,0.020394666327370536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,2048,4096,0.03674933314323425
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2048,3072,0.008472888833946651
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,2048,3584,0.03477599885728624
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2048,3072,0.018591110905011494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2048,2560,0.007640000018808577
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2048,2560,0.017380444539917838
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2048,2048,0.0069573331210348345
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,2048,3072,0.03482044405407376
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,2048,2560,0.03333244389957852
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2048,2048,0.017027555240525138
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2048,1536,0.005696889013051987
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,2048,2048,0.029840889904234145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2048,1536,0.015844444433848064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2048,1024,0.00481599983241823
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,2048,1536,0.03013155526585049
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2048,1024,0.015217777755525378
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2048,768,0.004217777815130022
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,2048,1024,0.027903109788894653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2048,512,0.014263111684057446
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2048,768,0.014726221561431885
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2048,512,0.0037786666717794207
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,2048,768,0.0270115551021364
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2048,256,0.003343111111058129
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,2048,512,0.026014222039116755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2048,256,0.013507555756303998
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2048,128,0.0030533334033356774
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,2048,256,0.024803555674023096
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2048,128,0.012959111067983838
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2048,64,0.0029271110478374693
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2048,65536,0.07441955804824829
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,2048,128,0.022801778382725183
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,2048,32,0.0029537777105967202
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2048,64,0.012837332983811697
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,2048,32,0.01310488912794325
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1536,16384,0.04408266809251574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1536,12288,0.034689777427249484
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,1536,16384,0.06319377819697063
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,1536,12288,0.05429688758320279
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1536,10240,0.01387111097574234
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1536,10240,0.03123555580774943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1536,12288,0.015801777442296345
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1536,8192,0.012074666718641916
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1536,16384,0.019928889142142404
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,1536,10240,0.04779466655519274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1536,8192,0.028055999014112685
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1536,7168,0.011197333534558615
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,1536,8192,0.04330133398373922
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1536,7168,0.026310222016440496
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1536,6144,0.010495999621020423
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,1536,7168,0.0406257775094774
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1536,6144,0.024357333779335022
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1536,5120,0.010207110808955299
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1536,65536,0.13595822122361925
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,1536,5120,0.03702488872740004
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,1536,6144,0.03898933198716905
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1536,5120,0.02254044512907664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1536,4096,0.009541333549552495
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1536,4096,0.01995022263791826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1536,3584,0.009851555857393477
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,1536,4096,0.03487199876043532
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1536,3584,0.018935999936527677
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1536,3072,0.00904888908068339
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,1536,3584,0.033230221933788724
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1536,3072,0.018029333816634286
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1536,2560,0.007198221981525421
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,1536,65536,0.16989244355095756
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,1536,3072,0.033222221665912204
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1536,2560,0.01719288859102461
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1536,2048,0.0064195555945237475
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,1536,2560,0.030774222479926214
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1536,2048,0.01663733356528812
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1536,1536,0.005264889034960005
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1536,65536,0.0598506662580702
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,1536,2048,0.029510223203235205
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1536,1536,0.01567911108334859
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1536,1024,0.004318222403526306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,1536,1536,0.029719110992219713
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1536,1024,0.01519733336236742
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1536,768,0.004018666843573253
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,1536,1024,0.027477333943049114
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1536,768,0.01461955573823717
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1536,512,0.003620444486538569
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,1536,768,0.026975999275843304
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1536,512,0.01418133411142561
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1536,256,0.0032711111836963226
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,1536,512,0.02568888829814063
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1536,256,0.013281777501106262
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1536,128,0.0029315555261241067
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,1536,256,0.024117334021462336
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1536,128,0.012912888493802814
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1536,64,0.002830222249031067
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,1536,128,0.022868444522221882
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1536,64,0.012784889174832238
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1536,32,0.0029075555503368378
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1536,32,0.01309511148267322
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1024,16384,0.015003555350833468
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1024,16384,0.041831112570232816
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1024,12288,0.012487110992272695
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1024,12288,0.034430222378836736
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,1024,16384,0.06529866986804538
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1024,10240,0.011459555890825061
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,1024,12288,0.05448355608516269
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1024,10240,0.03110399842262268
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1024,8192,0.010495999621020423
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,1024,10240,0.04920977685186598
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1024,8192,0.02773777809407976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1024,7168,0.01017155581050449
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1024,7168,0.026114667455355327
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,1024,8192,0.043402665191226535
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1024,65536,0.13325244850582546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1024,6144,0.009626666704813639
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,1024,7168,0.04118400149875217
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1024,6144,0.023893333143658105
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,1024,65536,0.17132977644602457
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1024,5120,0.009126221968068017
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,1024,6144,0.039429333474900984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1024,5120,0.021078222327762183
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1024,4096,0.008656000097592672
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1024,3584,0.008830221990744272
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,1024,5120,0.035551998350355364
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1024,4096,0.019485332899623446
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,1024,4096,0.03520977828237746
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1024,3072,0.00795199970404307
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1024,3584,0.018699554933442008
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1024,3072,0.018018666240904067
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,1024,3584,0.032103111346562706
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1024,2560,0.00703111125363244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,1024,3072,0.03194933467441135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1024,2560,0.017172444197866652
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1024,65536,0.04673244555791219
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1024,2048,0.005776888794369168
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,1024,2560,0.03132355544302199
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1024,2048,0.016559999850061204
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1024,1536,0.005271111097600725
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,1024,2048,0.02944177720281813
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1024,1536,0.01586844523747762
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1024,1024,0.00433688900536961
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1024,768,0.014631110760900708
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,1024,1536,0.029567999972237483
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1024,1024,0.01495911180973053
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1024,768,0.003975999852021535
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,1024,1024,0.02751022246148851
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1024,512,0.003593777616818746
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,1024,768,0.027024888330035742
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1024,256,0.0031795555518733132
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1024,512,0.014101333088344999
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1024,256,0.01330133361948861
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,1024,512,0.025990222891171772
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1024,128,0.0029271110478374693
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,1024,256,0.02409066590997908
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1024,128,0.012824888858530255
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1024,64,0.002768888862596618
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,1024,128,0.023065777288542852
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1024,64,0.012795555922720166
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,1024,32,0.0029235554652081597
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,1024,32,0.01316177762216992
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,768,16384,0.012325333224402534
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,768,16384,0.04163555635346307
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,768,12288,0.01054133309258355
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,768,16384,0.06087999873691135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,768,12288,0.034141331911087036
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,768,10240,0.010250666903124915
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,768,12288,0.05013422171274821
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,768,10240,0.03049955434269375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,768,8192,0.011587555209795633
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,768,10240,0.046247111426459424
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,768,8192,0.026837334036827087
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,768,65536,0.1327244440714518
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,768,7168,0.010026666853162978
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,768,8192,0.04188444548183017
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,768,65536,0.15931910938686794
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,768,7168,0.025498666697078284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,768,6144,0.00923733330435223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,768,7168,0.038719111018710665
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,768,6144,0.02362666692998674
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,768,65536,0.038021332687801786
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,768,5120,0.009296889106432596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,768,6144,0.03830044468243917
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,768,5120,0.02146666745344798
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,768,4096,0.00792266676823298
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,768,5120,0.033929778469933405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,768,4096,0.01957066688272688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,768,3584,0.008184000021881526
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,768,4096,0.03312888741493225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,768,3584,0.01906755566596985
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,768,2560,0.0063564446237352155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,768,3072,0.007616889145639207
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,768,3584,0.030927111705144245
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,768,3072,0.018090666996108163
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,768,3072,0.03174044357405768
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,768,2560,0.0170160002178616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,768,2048,0.005699555493063397
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,768,2560,0.030420445733600195
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,768,2048,0.016501333978441026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,768,1536,0.004927111168702443
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,768,2048,0.029218667083316382
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,768,1536,0.015805333852767944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,768,1024,0.004214222232500712
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,768,1536,0.02905333373281691
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,768,1024,0.015015999476114908
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,768,768,0.003855111284388436
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,768,1024,0.027560888065232172
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,768,512,0.013887999786270989
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,768,768,0.014575999644067554
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,768,512,0.025619554850790236
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,768,512,0.003573333223660787
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,768,768,0.027060444156328838
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,768,256,0.0032133333798911837
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,768,256,0.01330311099688212
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,768,128,0.0029075555503368378
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,768,256,0.023961777488390606
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,768,128,0.012915555801656512
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,768,64,0.0027457777824666765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,768,128,0.022833777798546687
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,768,32,0.0028657778683635923
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,768,64,0.012780444489585029
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,768,32,0.012963555753231049
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,512,16384,0.010919111470381418
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,512,16384,0.041535112592909075
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,512,12288,0.009911110831631554
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,512,16384,0.05496622125307719
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,512,12288,0.033540444241629705
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,512,10240,0.00944088896115621
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,512,12288,0.0476800004641215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,512,10240,0.02983022067281935
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,512,65536,0.12987644142574734
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,512,8192,0.008894222478071848
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,512,8192,0.02621511121590932
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,512,10240,0.043649779425727
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,512,65536,0.145541336801317
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,512,7168,0.008574222524960836
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,512,65536,0.02887644370396932
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,512,8192,0.03744711147414313
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,512,7168,0.02440888848569658
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,512,6144,0.00778311077091429
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,512,7168,0.03656889001528422
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,512,6144,0.023237332701683044
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,512,5120,0.008290666672918532
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,512,6144,0.036883556180530123
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,512,5120,0.020836444364653692
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,512,4096,0.008015111088752747
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,512,5120,0.03222844335767958
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,512,4096,0.01926399932967292
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,512,3584,0.007710222568776872
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,512,4096,0.03169155451986525
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,512,3584,0.018610666195551556
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,512,3072,0.007106666763623555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,512,3584,0.030271109607484605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,512,3072,0.01769333415561252
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,512,2560,0.006234666539563074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,512,3072,0.030433777305814955
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,512,2560,0.01682044400109185
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,512,2048,0.005600888695981767
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,512,2560,0.029714667134814795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,512,2048,0.015968888998031616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,512,1536,0.004847110973464118
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,512,2048,0.028415110376146104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,512,1536,0.01537333263291253
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,512,1024,0.0041475556790828705
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,512,1536,0.028564443190892536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,512,1024,0.014914666612943014
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,512,768,0.003823111040724648
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,512,1024,0.02721244427892897
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,512,768,0.014398222168286642
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,512,512,0.025144888295067683
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,512,512,0.0034560000316964257
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,512,768,0.02656800051530202
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,512,512,0.013983110586802164
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,512,256,0.0031102223114834893
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,512,256,0.01314933349688848
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,512,128,0.002871111035346985
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,512,256,0.02387733260790507
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,512,128,0.01275288893116845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,512,64,0.0027395555128653846
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,512,128,0.02271466619438595
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,512,32,0.0027662221756246355
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,512,64,0.012695999609099494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,512,32,0.012891555825869242
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,256,65536,0.019263111882739596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,256,16384,0.010607110957304636
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,256,65536,0.128328005472819
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,256,16384,0.04108622339036729
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,256,12288,0.009268444445398118
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,256,65536,0.14465778403811985
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,256,16384,0.0547288888030582
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,256,12288,0.03295288814438714
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,256,10240,0.008294222255547842
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,256,12288,0.047467556264665395
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,256,10240,0.029174221886528864
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,256,8192,0.0075128889746136135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,256,10240,0.04176622298028734
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,256,8192,0.02588622272014618
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,256,7168,0.00721155520942476
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,256,8192,0.03708622190687392
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,256,7168,0.02458133300145467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,256,6144,0.006809777683681912
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,256,7168,0.03628355595800612
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,256,6144,0.022407111194398668
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,256,5120,0.00720888872941335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,256,6144,0.03446400165557861
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,256,5120,0.02071644365787506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,256,4096,0.006855999843941794
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,256,5120,0.03215377860599094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,256,4096,0.019137778215938143
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,256,3584,0.007119110888904995
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,256,4096,0.03144444359673394
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,256,3584,0.018378666705555387
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,256,3072,0.006871999965773688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,256,3584,0.03010400136311849
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,256,3072,0.01752000053723653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,256,2560,0.006115555349323485
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,256,3072,0.030284444491068523
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,256,2560,0.01664355562792884
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,256,2048,0.005389333185222413
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,256,2560,0.029453333881166246
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,256,2048,0.016000888413853116
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,256,1536,0.0047395556337303585
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,256,2048,0.028062221076753404
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,256,1024,0.004055110944641961
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,256,1536,0.015280889140235053
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,256,1024,0.014524444937705994
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,256,768,0.0036968886852264404
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,256,1536,0.028379556205537584
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,256,512,0.013760888742076026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,256,1024,0.02713866697417365
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,256,768,0.01423111061255137
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,256,512,0.003359111025929451
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,256,768,0.02641155487961239
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,256,256,0.003010666618744532
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,256,64,0.0026844443960322272
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,256,512,0.025705777936511572
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,256,128,0.02252799934811062
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,256,256,0.01316088851955202
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,256,128,0.0028728888266616394
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,256,256,0.02363555630048116
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,256,128,0.012718222207493253
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,256,32,0.0026906666656335196
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,256,64,0.01278133359220293
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,256,32,0.01295733369059033
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,128,65536,0.014516444669829475
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,128,65536,0.12627555264367
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,128,16384,0.007903111477692923
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,128,16384,0.03916710946294997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,128,65536,0.14510933558146158
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,128,12288,0.0069253332912921906
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,128,12288,0.03196177879969279
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,128,16384,0.052707556221220225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,128,10240,0.007075555622577667
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,128,12288,0.044997334480285645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,128,10240,0.029490666257010564
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,128,8192,0.007052444749408298
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,128,10240,0.041160000695122614
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,128,8192,0.025628444221284654
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,128,7168,0.006861333217885759
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,128,8192,0.03678933448261685
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,128,7168,0.023874666955735948
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,128,6144,0.006589333216349284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,128,7168,0.0358862214618259
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,128,6144,0.022260444031821355
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,128,5120,0.00703111125363244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,128,6144,0.03428711162673102
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,128,5120,0.02074666652414534
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,128,4096,0.006540444576077991
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,128,5120,0.03202844328350491
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,128,4096,0.018980445133315194
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,128,3584,0.006996444529957241
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,128,4096,0.03128711051411099
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,128,3584,0.018238221605618794
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,128,3072,0.006699555450015598
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,128,3584,0.02985422147644891
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,128,3072,0.01735911104414198
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,128,2560,0.006114666660626729
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,128,3072,0.030234667989942763
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,128,2048,0.015826667348543804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,128,2560,0.016604445046848722
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,128,2048,0.005433777968088786
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,128,2560,0.029431111282772485
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,128,1536,0.0047128887640105355
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,128,2048,0.028083556228213843
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,128,1024,0.004037333445416557
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,128,1536,0.015260444747077094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,128,1536,0.028372443384594385
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,128,1024,0.01464000013139513
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,128,768,0.003718222181002299
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,128,1024,0.026904889278941687
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,128,768,0.014087999860445658
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,128,512,0.003383111208677292
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,128,512,0.02493600050608317
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,128,768,0.02591466738118066
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,128,512,0.01370666672786077
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,128,256,0.002973333415057924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,128,256,0.013051555388503604
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,128,128,0.002838222309947014
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,128,256,0.023727110690540735
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,128,128,0.012771555946932899
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,128,64,0.0026551110463009942
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,80,128,128,0.022444445225927565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,128,32,0.0026613333159022858
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,128,64,0.012712889247470431
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,128,32,0.01258133351802826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,64,65536,0.012875555290116204
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,64,16384,0.006834666762087081
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,64,65536,0.12676800621880427
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,64,12288,0.006710222197903528
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,64,16384,0.039036442836125694
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,64,12288,0.03197155396143595
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,64,10240,0.0069422221018208405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,64,8192,0.006768888897365994
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,64,10240,0.028656889994939167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,64,7168,0.006599999964237213
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,64,8192,0.02570933269129859
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,64,7168,0.024085332949956257
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,64,6144,0.006452444526884291
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,64,5120,0.0069751110341813825
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,64,6144,0.022308443983395893
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,64,5120,0.020639111598332722
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,64,2560,0.006046222315894232
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,64,3072,0.01735377808411916
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,64,4096,0.0064702220261096954
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,64,4096,0.018996444013383653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,64,3584,0.0070497774415545994
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,64,3584,0.018246221873495314
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,64,3072,0.00665155549844106
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,64,2560,0.016558221644825406
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,64,2048,0.005380444642570283
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,64,1536,0.0046986668474144405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,64,2048,0.015818667080667283
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,64,1536,0.015242666006088257
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,64,1024,0.004041777716742622
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,64,1024,0.014529777897728814
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,64,768,0.003644444462325838
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,64,768,0.014125333891974555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,64,512,0.003335111050142182
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,64,512,0.013658666776286231
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,64,256,0.002971555623743269
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,64,128,0.0027840000887711844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,64,256,0.013020444247457715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,64,128,0.012610666453838348
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,64,64,0.002609777781698439
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,64,32,0.0026684444811609057
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,64,64,0.01257955531279246
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,32,65536,0.011503111157152386
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,64,32,0.012652444342772165
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,32,16384,0.007093333535724216
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,32,65536,0.1267226669523451
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,32,12288,0.006642666541867786
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,32,16384,0.038979556825425886
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,32,10240,0.00685066646999783
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,32,12288,0.032023111979166664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,32,8192,0.006647999915811751
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,32,10240,0.028868443436092798
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,32,7168,0.006528888725572162
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,32,8192,0.025431111454963684
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,32,6144,0.006325333482689328
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,32,7168,0.024086222052574158
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,32,6144,0.022152000003390845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,32,5120,0.00674222202764617
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,32,4096,0.0063866666621632045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,32,5120,0.020627554919984605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,32,3584,0.006888000087605582
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,32,4096,0.019123555885420907
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,32,3584,0.01824711097611321
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,32,3072,0.00658044425977601
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,32,2560,0.006039111150635614
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,32,3072,0.01738133364253574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,32,2048,0.005306666509972678
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,32,2560,0.016639111770523917
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,32,2048,0.01581333412064446
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,32,1536,0.004679999831649992
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,32,1024,0.003983111017280155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,32,1536,0.015271110667122735
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,32,768,0.00360977773865064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,32,1024,0.014530667000346713
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,32,512,0.0032968889507982465
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,32,768,0.0140337778462304
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,32,512,0.013718222578366598
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,32,256,0.002944888960984018
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,32,256,0.013081777426931592
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,32,128,0.002712888850106133
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,32,128,0.012616888516479068
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,32,64,0.00258577780591117
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,32,64,0.01238666640387641
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,80,32,32,0.0025840000145965153
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,80,32,32,0.01257066677014033
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,65536,12288,0.21526400248209634
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,65536,10240,0.18337955739763048
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,65536,16384,0.28125421206156415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,65536,10240,0.3574568960401747
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,65536,8192,0.15199466546376547
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,65536,8192,0.28465067015753853
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,65536,12288,0.42777156829833984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,65536,16384,0.5683306588066949
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,65536,7168,0.14294755458831787
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,65536,7168,0.2536124388376872
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,65536,6144,0.11848266919453938
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,65536,6144,0.24463200569152832
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,65536,5120,0.10145866870880127
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,65536,5120,0.1838915612962511
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,65536,12288,0.24684890111287436
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,65536,4096,0.14782488346099854
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,65536,16384,0.3036239941914876
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,65536,4096,0.08572177754508124
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,65536,10240,0.21123733785417345
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,65536,3584,0.07969866858588324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,65536,3584,0.1295564439561632
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,65536,8192,0.1748373375998603
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,65536,3072,0.073216888639662
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,65536,3072,0.11285777886708577
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,65536,2560,0.09392799933751424
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,65536,2560,0.06260710954666138
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,65536,7168,0.15814577208624944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,65536,2048,0.05555644300248888
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,65536,2048,0.07649688588248359
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,65536,6144,0.13982310560014513
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,65536,1536,0.062393777900271945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,65536,1536,0.04715822140375773
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,65536,1024,0.041931556330786816
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,65536,1024,0.04019733270009359
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,65536,5120,0.12368888325161403
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,65536,768,0.032553778754340276
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,65536,768,0.037951111793518066
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,65536,4096,0.10464266935984294
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,65536,512,0.02609511050913069
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,65536,512,0.03407288922203912
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,65536,3584,0.09294222460852729
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,65536,256,0.013735111388895245
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,65536,256,0.02926666537920634
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,65536,3072,0.0841111143430074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,65536,128,0.008747555315494537
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,65536,128,0.028129776318868
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,65536,2560,0.07907733652326795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,65536,64,0.007110222346252865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,65536,64,0.02602844436963399
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,65536,32,0.005903999838564131
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,65536,32,0.02593155536386702
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,65536,2048,0.06621422370274861
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,65536,1536,0.059007108211517334
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,65536,256,0.03334133492575751
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,65536,512,0.04024444354905023
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,65536,768,0.04488355583614773
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,65536,128,0.02927910950448778
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,65536,1024,0.04993066522810194
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,16384,16384,0.13639377223120794
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,16384,16384,0.08816088570488824
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,16384,12288,0.06911822160085042
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,16384,12288,0.0970639983812968
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,16384,10240,0.06059910853703817
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,16384,10240,0.08226755592558119
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,16384,8192,0.05160977774196201
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,16384,8192,0.06636533472273085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,16384,65536,0.31419555346171063
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,16384,7168,0.0475413335694207
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,16384,7168,0.058493336041768394
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,16384,65536,0.5760293536716038
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,16384,6144,0.045662220981385976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,16384,6144,0.055693335003323026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,16384,5120,0.043544001049465604
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,16384,5120,0.037965334124035306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,16384,16384,0.1086862219704522
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,16384,12288,0.08894666698243882
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,16384,4096,0.03599288728502061
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,16384,4096,0.03462666604253981
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,16384,3584,0.03172533379660712
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,16384,10240,0.076910224225786
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,16384,3584,0.03234400020705329
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,16384,8192,0.06680177979999118
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,16384,3072,0.027882665395736694
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,16384,65536,0.3424515459272597
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,16384,3072,0.030142221185896132
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,16384,2560,0.02757155564096239
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,16384,7168,0.06229155593448215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,16384,2560,0.02398666739463806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,16384,2048,0.02000266644689772
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,16384,2048,0.025192888246642217
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,16384,6144,0.05697511302100288
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,16384,1536,0.022095110681321885
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,16384,1536,0.015973332855436537
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,16384,1024,0.01148444414138794
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,16384,1024,0.020201777418454487
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,16384,5120,0.05350133445527819
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,16384,768,0.00944622192117903
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,16384,768,0.019288889235920377
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,16384,4096,0.04892711175812615
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,16384,512,0.0074240002367231585
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,16384,512,0.018058665924602084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,16384,3584,0.0477439992957645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,16384,256,0.0053191110491752625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,16384,256,0.014512000812424554
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,16384,3072,0.04501422246297201
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,16384,128,0.004256000121434529
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,16384,128,0.014484445254007975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,16384,2560,0.04173333446184794
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,16384,64,0.003759110967318217
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,16384,64,0.015277332729763456
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,16384,32,0.003802666647566689
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,16384,32,0.015418666932317944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,16384,2048,0.03808622227774726
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,16384,1536,0.03638666537072923
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,16384,1024,0.03317777646912469
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,16384,768,0.03183555603027344
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,16384,512,0.029708445072174072
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,12288,16384,0.07431999842325847
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,12288,16384,0.10730489095052083
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,16384,256,0.024283554818895128
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,12288,12288,0.05907466676500109
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,12288,12288,0.07740622096591525
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,16384,128,0.023936000135209825
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,12288,65536,0.26042132907443577
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,12288,10240,0.05182133449448479
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,12288,8192,0.0447324448161655
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,12288,10240,0.06632266441980998
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,12288,8192,0.053455998500188194
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,12288,7168,0.04296355446179708
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,12288,7168,0.04662222332424588
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,12288,65536,0.405240880118476
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,12288,6144,0.04063466522428725
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,12288,6144,0.03914311197068956
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,12288,5120,0.03495466709136963
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,12288,5120,0.03344355689154731
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,12288,16384,0.0912346641222636
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,12288,12288,0.07418844434950086
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,12288,4096,0.03109688891304864
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,12288,4096,0.029045333464940388
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,12288,3584,0.025566221939192876
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,12288,3584,0.028866665230857
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,12288,65536,0.300617774327596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,12288,10240,0.06670222017500135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,12288,3072,0.026958222190539043
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,12288,3072,0.024471999870406255
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,12288,2560,0.019724445210562814
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,12288,2560,0.024400000770886738
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,12288,8192,0.061174220508999295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,12288,2048,0.01632000009218852
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,12288,2048,0.0225591113169988
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,12288,7168,0.05615466833114624
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,12288,1536,0.013074666261672974
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,12288,1536,0.021016000045670405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,12288,6144,0.05312622255749173
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,12288,1024,0.009840889109505547
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,12288,1024,0.019661333825853135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,12288,5120,0.048933333820766874
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,12288,768,0.008206222620275285
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,12288,768,0.01872977779971229
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,12288,4096,0.04466222061051262
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,12288,512,0.006738666858938005
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,12288,512,0.01645066671901279
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,12288,3584,0.0438435541258918
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,12288,256,0.00433955548538102
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,12288,256,0.01424711114830441
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,12288,3072,0.04271822174390157
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,12288,128,0.0037831109431054858
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,12288,2560,0.03834133346875509
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,12288,128,0.014205333259370593
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,12288,64,0.00342399999499321
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,12288,64,0.014853333433469137
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,12288,32,0.0034764446318149567
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,12288,32,0.014998222390810648
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,12288,2048,0.035283555587132774
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,12288,1536,0.034938666555616595
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,12288,1024,0.03211466802491082
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,12288,768,0.030766222212049697
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,10240,16384,0.06830222076839872
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,10240,65536,0.23487112257215711
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,10240,16384,0.08058400286568536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,10240,12288,0.06669066349665324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,12288,256,0.023947555157873366
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,12288,512,0.02847466535038418
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,12288,128,0.023315555519527856
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,10240,12288,0.055303110016716846
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,10240,10240,0.049230221245023936
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,10240,10240,0.05299733413590325
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,10240,65536,0.3398844401041667
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,10240,8192,0.042597333590189614
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,10240,8192,0.04252800014283922
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,10240,7168,0.038903110557132296
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,10240,7168,0.03786399960517883
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,10240,6144,0.03757066528002421
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,10240,6144,0.03315644462903341
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,10240,5120,0.028578665521409776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,10240,5120,0.03236888845761617
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,10240,16384,0.08218489090601604
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,10240,4096,0.023573334018389385
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,10240,4096,0.029523554775449965
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,10240,65536,0.25698312123616535
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,10240,12288,0.06737155384487577
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,10240,3584,0.02108000053299798
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,10240,10240,0.06122399701012505
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,10240,3584,0.028059555424584284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,10240,3072,0.025545777546034917
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,10240,3072,0.019519999623298645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,10240,8192,0.055051555236180626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,10240,2560,0.017321777012613084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,10240,2560,0.02372088862790002
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,10240,7168,0.05275911092758179
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,10240,2048,0.013935999737845527
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,10240,2048,0.0221742226017846
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,10240,1536,0.011205332974592844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,10240,1536,0.020640000700950623
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,10240,6144,0.04807200034459432
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,10240,1024,0.008930666579140557
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,10240,1024,0.01902577777703603
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,10240,5120,0.04569422205289205
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,10240,768,0.007448888487286038
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,10240,768,0.017521777086787753
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,10240,4096,0.04149333304829068
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,10240,512,0.006360000206364526
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,10240,512,0.015173332558737861
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,10240,3584,0.0414648883872562
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,10240,256,0.0038177776667806837
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,10240,256,0.014177777700954013
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,10240,3072,0.038633777035607234
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,10240,128,0.0034426665968365143
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,10240,128,0.014001778430408902
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,10240,2560,0.036559998989105225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,10240,64,0.003141333245568805
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,10240,64,0.014327110515700446
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,10240,32,0.003244444520937072
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,10240,32,0.014479999740918478
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,10240,1536,0.033815999825795494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,10240,2048,0.03359555535846286
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,10240,1024,0.03224710954560174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,10240,768,0.03048355711830987
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,8192,16384,0.07201421923107572
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,8192,16384,0.054930667082468666
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,8192,65536,0.179439107577006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,8192,12288,0.044693334235085376
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,10240,512,0.027828445037206013
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,8192,12288,0.056220445368025035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,8192,10240,0.039613333013322614
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,8192,10240,0.04336622357368469
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,10240,256,0.024122666981485155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,8192,8192,0.03741600116093954
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,10240,128,0.023574221465322707
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,8192,65536,0.26780356301201713
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,8192,8192,0.034697777695126004
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,8192,7168,0.03197600113021003
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,8192,7168,0.03275288807021247
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,8192,6144,0.03125422199567159
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,8192,6144,0.029636445972654555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,8192,5120,0.024874667326609295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,8192,5120,0.027063111464182537
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,8192,16384,0.07349777883953519
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,8192,4096,0.021147555775112573
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,8192,65536,0.22333510716756186
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,8192,4096,0.023980443676312763
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,8192,12288,0.06136888927883572
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,8192,3584,0.018431110514534842
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,8192,3584,0.021947556071811255
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,8192,3072,0.01590755581855774
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,8192,3072,0.020418667131000094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,8192,10240,0.05568710962931315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,8192,2560,0.014347554908858405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,8192,2560,0.01943111088540819
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,8192,8192,0.05059911145104302
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,8192,7168,0.04843022094832527
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,8192,2048,0.012134222520722283
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,8192,2048,0.018231110440360177
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,8192,1536,0.00960622231165568
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,8192,1536,0.01705777810679542
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,8192,6144,0.044897778166664966
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,8192,1024,0.007763555480374231
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,8192,1024,0.015841777125994366
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,8192,5120,0.043040888177024
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,8192,768,0.006628444625271692
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,8192,768,0.01370666672786077
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,8192,4096,0.03918222255176968
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,8192,512,0.005088888936572605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,8192,512,0.013391111459996967
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,8192,3584,0.03878488805558946
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,8192,256,0.003935111065705617
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,8192,256,0.01260444439119763
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,8192,3072,0.03648533423741659
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,8192,128,0.003539555602603488
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,8192,128,0.012527999778588613
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,8192,2560,0.03488444288571676
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,8192,64,0.0033173333439562055
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,8192,64,0.012904889053768583
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,8192,32,0.0033448889023727844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,8192,32,0.012921777864297232
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,8192,2048,0.03253777821858724
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,8192,1536,0.03326933251486884
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,8192,1024,0.03137866655985514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,7168,16384,0.07246933380762736
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,7168,16384,0.05164444446563721
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,7168,65536,0.1669306622611152
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,8192,768,0.028611554039849177
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,7168,12288,0.041519999504089355
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,7168,12288,0.044659554958343506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,8192,512,0.0255022214518653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,8192,256,0.023758222659428913
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,7168,65536,0.24487378862169054
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,7168,10240,0.03707022137112088
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,7168,10240,0.03941511114438375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,8192,128,0.022832888695928786
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,7168,8192,0.03286755416128371
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,7168,8192,0.031493332650926374
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,7168,7168,0.030209776428010728
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,7168,7168,0.02772444486618042
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,7168,6144,0.028064890040291682
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,7168,6144,0.024673778149816725
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,7168,5120,0.021362667282422382
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,7168,5120,0.02516088883082072
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,7168,16384,0.07079911231994629
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,7168,4096,0.018223999275101554
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,7168,4096,0.022428444690174524
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,7168,65536,0.21162488725450304
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,7168,12288,0.05838221973843045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,7168,3584,0.02077511118517982
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,7168,3584,0.017214222086800467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,7168,10240,0.05392533209588793
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,7168,3072,0.01995733380317688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,7168,3072,0.01421688828203413
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,7168,8192,0.04868888854980469
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,7168,2560,0.012676444318559436
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,7168,2560,0.01886311173439026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,7168,2048,0.010758221977286868
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,7168,2048,0.017834666702482436
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,7168,7168,0.047855999734666615
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,7168,1536,0.008952889177534316
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,7168,1536,0.016891555653678048
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,7168,6144,0.04377333323160807
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,7168,1024,0.007135111424658034
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,7168,1024,0.015067555838161044
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,7168,5120,0.04196711050139534
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,7168,768,0.006265777680608962
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,7168,768,0.013597333596812354
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,7168,4096,0.037864890363481306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,7168,512,0.0046640001237392426
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,7168,512,0.013199110825856527
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,7168,3584,0.03715022073851691
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,7168,256,0.0038648889296584656
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,7168,256,0.012527999778588613
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,7168,3072,0.035615109735065036
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,7168,128,0.0034764446318149567
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,7168,128,0.0123848890264829
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,7168,2560,0.033933334880405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,7168,64,0.003264888914095031
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,7168,64,0.01272266689274046
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,7168,32,0.0033200000309281876
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,7168,32,0.012723555167516073
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,7168,2048,0.032052444087134466
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,7168,1536,0.03270755542649163
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,7168,1024,0.030942221482594807
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,7168,768,0.029056890143288508
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,6144,16384,0.04786755641301473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,6144,16384,0.0601466695467631
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,6144,65536,0.1560266680187649
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,6144,12288,0.04024799995952182
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,6144,12288,0.03885866536034478
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,6144,10240,0.03501777847607931
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,6144,10240,0.034816887643602155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,7168,512,0.02564711040920681
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,7168,256,0.024146666129430134
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,6144,65536,0.20158310731252035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,7168,128,0.02341777748531765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,6144,8192,0.03134222163094415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,6144,8192,0.028562668297025893
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,6144,7168,0.02862311071819729
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,6144,7168,0.024744888146718342
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,6144,6144,0.025669333007600572
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,6144,6144,0.0221742226017846
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,6144,5120,0.019304000669055514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,6144,5120,0.02346399923165639
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,6144,16384,0.06600266695022583
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,6144,12288,0.055227554506725736
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,6144,65536,0.19668977790408662
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,6144,4096,0.022149332695537146
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,6144,4096,0.016601777738995023
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,6144,3584,0.014490667316648694
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,6144,3584,0.019927110936906602
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,6144,3072,0.012845333251688214
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,6144,3072,0.019121777680185106
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,6144,10240,0.050923556089401245
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,6144,2560,0.011170667078759937
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,6144,2560,0.018065777089860704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,6144,8192,0.04786400000254313
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,6144,2048,0.009715555442704095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,6144,2048,0.017249777913093567
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,6144,7168,0.04476088947719998
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,6144,1536,0.008186666501892937
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,6144,1536,0.016529776983790927
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,6144,6144,0.041136887338426374
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,6144,1024,0.006685333533419504
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,6144,1024,0.013889777991506787
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,6144,5120,0.03968977928161621
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,6144,768,0.0058240000572469496
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,6144,768,0.013531555732091268
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,6144,4096,0.035754667388068304
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,6144,512,0.004563555535342959
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,6144,512,0.013155555559529198
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,6144,3584,0.036299556493759155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,6144,256,0.0035946667194366455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,6144,256,0.012367111113336353
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,6144,3072,0.035029331843058266
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,6144,128,0.0033440000067154565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,6144,128,0.012395555774370829
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,6144,2560,0.034027556578318276
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,6144,64,0.00309955556359556
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,6144,64,0.012483555409643384
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,6144,32,0.003156444471743372
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,6144,32,0.012625777886973487
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,6144,2048,0.032050665881898664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,6144,1536,0.032590223683251276
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,6144,1024,0.030377778742048476
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,6144,768,0.026878222823143005
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,5120,16384,0.05041066805521647
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,5120,16384,0.0433573325475057
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,5120,65536,0.13684088653988308
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,5120,65536,0.17612000306447348
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,5120,12288,0.03994399971432156
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,5120,12288,0.03605955508020189
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,5120,10240,0.032590223683251276
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,5120,10240,0.034070223569869995
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,5120,8192,0.02427644365363651
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,6144,512,0.025272889269722834
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,6144,256,0.023579556081030104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,5120,8192,0.029158221350775823
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,5120,7168,0.021538666552967493
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,5120,7168,0.026320889592170715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,6144,128,0.022816000713242426
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,5120,6144,0.024455111887719896
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,5120,6144,0.019168888529141743
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,5120,5120,0.016711999972661335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,5120,5120,0.022016889519161646
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,5120,16384,0.061759114265441895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,5120,4096,0.014384888940387301
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,5120,4096,0.019987555013762582
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,5120,65536,0.1729511155022515
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,5120,12288,0.05383111039797465
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,5120,3584,0.019288000133302476
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,5120,3584,0.012358222570684222
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,5120,3072,0.011051555474599203
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,5120,3072,0.018283555905024212
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,5120,10240,0.0483528905444675
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,5120,2560,0.009699555734793345
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,5120,2560,0.01759999990463257
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,5120,8192,0.044079999128977455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,5120,2048,0.008450667063395182
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,5120,2048,0.016855110724767048
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,5120,7168,0.04247555467817518
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,5120,1536,0.007299555672539606
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,5120,6144,0.03913066784540812
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,5120,1536,0.015847111741701763
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,5120,1024,0.006041777630647023
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,5120,1024,0.01369244439734353
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,5120,5120,0.03801777627733018
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,5120,768,0.005255110975768831
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,5120,768,0.013440888788965015
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,5120,4096,0.03479555580351088
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,5120,512,0.004232888834344016
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,5120,512,0.013186666700575085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,5120,3584,0.03536266750759549
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,5120,256,0.0034471110751231513
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,5120,256,0.01237422227859497
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,5120,3072,0.03441866570048862
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,5120,128,0.0031831111345026228
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,5120,128,0.012229333321253458
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,5120,2560,0.033542222446865506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,5120,64,0.0030542222989930045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,5120,64,0.012576888832781049
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,5120,32,0.003116444374124209
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,5120,32,0.012481778032249875
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,5120,2048,0.03128622306717767
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,5120,1536,0.03218933277659946
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,5120,1024,0.028001778655582007
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,5120,768,0.02664799988269806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,4096,65536,0.11947111288706462
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,4096,16384,0.040222220950656466
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,4096,16384,0.040264887942208186
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,4096,12288,0.03384888834423489
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,4096,12288,0.03222666515244378
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,4096,65536,0.14125066333346897
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,4096,10240,0.028335111008750066
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,4096,10240,0.03149688906139798
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,4096,8192,0.0243848893377516
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,4096,8192,0.026643556025293138
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,5120,512,0.025804445147514343
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,4096,7168,0.018272888329293992
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,5120,256,0.023354666100607976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,4096,7168,0.02437777817249298
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,5120,128,0.022799111074871484
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,4096,6144,0.022615111536449854
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,4096,6144,0.016163556112183463
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,4096,5120,0.014127110441525778
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,4096,5120,0.02072266737620036
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,4096,65536,0.15701422426435682
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,4096,16384,0.05746489100986057
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,4096,4096,0.01934400035275353
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,4096,4096,0.011924444801277585
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,4096,3584,0.010696888797812991
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,4096,3584,0.01867999964290195
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,4096,12288,0.04910222358173794
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,4096,3072,0.009696889254781935
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,4096,3072,0.017901332841979135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,4096,10240,0.04535911149448819
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,4096,2560,0.008555555509196388
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,4096,2560,0.01739644507567088
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,4096,8192,0.04188977678616842
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,4096,2048,0.007631111476156447
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,4096,2048,0.016238222519556682
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,4096,7168,0.04001955522431268
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,4096,1536,0.006601777755551868
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,4096,1536,0.014161777165200977
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,4096,6144,0.03714399867587619
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,4096,1024,0.005613333235184352
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,4096,1024,0.013392888837390475
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,4096,5120,0.03671111000908746
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,4096,768,0.004606222112973531
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,4096,768,0.013304000099500021
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,4096,4096,0.03386844529045953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,4096,512,0.003748444633351432
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,4096,512,0.012982221941153208
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,4096,3584,0.03479822145568
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,4096,256,0.0033617777129014335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,4096,256,0.012186666329701742
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,4096,3072,0.033655110332700945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,4096,128,0.0030666666312350165
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,4096,128,0.012334222594896952
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,4096,2560,0.03218844532966614
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,4096,64,0.002946666752298673
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,4096,64,0.012291555603345236
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,4096,32,0.0029608888758553397
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,4096,32,0.012159110771285163
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,4096,2048,0.031173331869973078
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,4096,1536,0.03128800127241347
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,4096,1024,0.028904888365003798
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,4096,768,0.026364444030655756
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3584,16384,0.03735911183887058
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3584,16384,0.03901422354910109
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3584,65536,0.11829422579871283
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3584,12288,0.02866311205757989
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3584,12288,0.0326275560590956
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3584,65536,0.12337156136830647
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3584,10240,0.030745777818891738
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3584,10240,0.02520888878239526
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3584,8192,0.022078222698635522
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3584,8192,0.025681777132882014
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,4096,512,0.025137777129809063
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,4096,256,0.023337778117921617
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3584,7168,0.023823110593689814
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3584,7168,0.019694222344292533
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,4096,128,0.022664000590642292
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3584,6144,0.01770755483044518
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3584,6144,0.021898667017618816
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3584,5120,0.01572800013754103
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3584,5120,0.020530665914217632
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,3584,16384,0.056295997566647
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3584,4096,0.011219555305110084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3584,4096,0.018949333164427016
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,3584,65536,0.14764711591932508
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3584,3584,0.010623111493057676
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3584,3584,0.018221333622932434
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,3584,12288,0.04782755507363213
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3584,3072,0.0092604441775216
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3584,3072,0.01777777738041348
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,3584,10240,0.044290665123197764
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3584,2560,0.008273777862389883
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3584,2560,0.017113778326246474
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,3584,8192,0.04048088855213589
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3584,2048,0.007319110963079665
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3584,2048,0.015917332635985482
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,3584,7168,0.03959822323587205
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3584,1536,0.006523555351628198
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3584,1536,0.014223110344674854
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,3584,6144,0.037652442852656044
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,3584,5120,0.03666311171319749
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3584,1024,0.005477333234416113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3584,1024,0.013657777673668332
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3584,768,0.004256000121434529
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3584,768,0.013254221942689685
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,3584,4096,0.033913777934180364
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3584,512,0.00371111101574368
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3584,512,0.012820444173283048
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,3584,3584,0.03465155429310269
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3584,256,0.0033191111352708605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3584,256,0.012177777787049612
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,3584,3072,0.033371554480658636
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3584,128,0.003104000041882197
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3584,128,0.012178666889667511
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,3584,2560,0.03216622273127238
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3584,64,0.0029608888758553397
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3584,64,0.01222222215599484
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3584,32,0.002969777832428614
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3584,32,0.012097777591811286
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,3584,2048,0.03087110983000861
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,3584,1536,0.029491557015313044
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3072,65536,0.11636355188157822
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,3584,1024,0.02736977736155192
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3072,16384,0.031918221049838595
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3072,16384,0.0382008883688185
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3072,65536,0.10696178012424046
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,3584,768,0.026239110363854304
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3072,12288,0.03189599845144484
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3072,12288,0.025684444440735713
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3072,10240,0.02849688794877794
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3072,10240,0.022441777918073866
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3072,8192,0.01976444489426083
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3072,8192,0.024643555283546448
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,3584,512,0.02500088844034407
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,3584,256,0.02387199964788225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,3584,128,0.023225777679019507
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3072,7168,0.02399377855989668
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3072,7168,0.017638223038779367
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3072,6144,0.02163822286658817
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3072,6144,0.015779554843902588
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3072,5120,0.013956444131003486
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3072,5120,0.02010844482315911
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,3072,16384,0.055128888951407544
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,3072,65536,0.14454132980770534
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3072,4096,0.010594666832023196
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3072,4096,0.018784888916545443
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,3072,12288,0.047453333934148155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3072,3584,0.009376000199053023
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3072,3584,0.018059555027219985
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3072,3072,0.008743111458089616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3072,3072,0.01754488878779941
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,3072,10240,0.04388977752791511
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3072,2560,0.007921777665615082
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3072,2560,0.01663377715481652
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,3072,8192,0.04000000158945719
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3072,2048,0.007184888753626082
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3072,2048,0.015212444795502557
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,3072,7168,0.0395760006374783
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3072,1536,0.0063182223174307085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3072,1536,0.0140684445699056
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,3072,6144,0.03658755620320638
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3072,1024,0.005159111072619756
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3072,1024,0.01332711097266939
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,3072,5120,0.03580710954136319
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3072,768,0.004219555606444676
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3072,768,0.013269333375824822
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,3072,4096,0.03358755509058634
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3072,512,0.003714666598372989
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,3072,3584,0.03414311011632284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3072,512,0.012975999878512489
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3072,256,0.0033440000067154565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3072,256,0.012297777665985955
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,3072,3072,0.03323288758595785
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3072,128,0.003010666618744532
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3072,128,0.012127111355463663
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,3072,2560,0.03188088867399428
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3072,64,0.0028808888875775864
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3072,64,0.01222222215599484
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,3072,32,0.003003555453485913
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,3072,32,0.012053333222866058
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,3072,2048,0.030373334884643555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,3072,1536,0.030949334303538006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2560,65536,0.11443289120992024
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,3072,1024,0.02769600020514594
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,3072,768,0.02607911162906223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2560,65536,0.09989777538511489
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2560,16384,0.03743022349145677
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2560,16384,0.028039111031426325
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2560,12288,0.031088001198238794
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2560,12288,0.022646221849653456
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2560,10240,0.028039111031426325
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2560,10240,0.01979466610484653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2560,8192,0.017072889539930556
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2560,8192,0.024359999431504145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,3072,512,0.02510844502184126
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,3072,256,0.02353688908947839
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,3072,128,0.02292622294690874
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2560,7168,0.023198222120602924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2560,6144,0.013972444666756524
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2560,7168,0.015580443872345818
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2560,6144,0.021394666698243883
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2560,5120,0.012786666552225748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2560,5120,0.020115555988417733
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,2560,16384,0.05463644531038073
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,2560,12288,0.04659733176231384
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2560,4096,0.009904888768990835
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2560,4096,0.018744889232847426
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2560,3584,0.009186666872766282
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,2560,10240,0.0432897773053911
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2560,3584,0.01835555500454373
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,2560,65536,0.14094489150577122
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2560,3072,0.00849422232972251
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2560,3072,0.017128888103697035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2560,2560,0.007749333149856991
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2560,2560,0.016381333271662395
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,2560,8192,0.03915466533766852
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2560,2048,0.0069360000391801195
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2560,2048,0.014477334088749357
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,2560,7168,0.038776887787712946
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2560,1536,0.006143110907740063
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2560,1536,0.01402311192618476
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,2560,6144,0.03668355610635545
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2560,1024,0.0048071112897661
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2560,1024,0.013363555901580386
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,2560,5120,0.03570222192340427
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2560,768,0.0040702223777771
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2560,768,0.01316088851955202
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,2560,4096,0.0332640012105306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2560,512,0.0036640001667870414
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2560,512,0.0129457778400845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,2560,3584,0.03402400016784668
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2560,256,0.0032924444725116095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,2560,3072,0.033231112692091204
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2560,256,0.012235555383894177
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2560,128,0.0029768887907266617
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2560,128,0.012216000093354119
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,2560,2560,0.032186667124430336
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2560,64,0.002833777831660377
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2560,64,0.01219555570019616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2560,32,0.002921777880854077
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2560,32,0.011995555626021491
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,2560,2048,0.02871733241611057
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2048,65536,0.10986133416493733
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2048,65536,0.07240711318122016
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2048,16384,0.02334311107794444
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2048,16384,0.0362115568584866
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,2560,1536,0.02819377846188015
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,2560,1024,0.02738488879468706
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2048,12288,0.019129777948061626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2048,12288,0.030247999562157526
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2048,10240,0.016729777057965595
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2048,10240,0.027154665854242113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,2560,768,0.026390223039521113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2048,8192,0.014731556177139282
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2048,8192,0.023965333898862202
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,2560,512,0.024824889169798955
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,2560,256,0.023192889160580103
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2048,7168,0.013769778112570444
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2048,7168,0.02256533337963952
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,2560,128,0.022470222579108343
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2048,6144,0.012489777472284106
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2048,6144,0.02110933264096578
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2048,5120,0.01125955581665039
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2048,5120,0.019685332973798115
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,2048,65536,0.13694222768147787
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2048,4096,0.010223111344708337
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2048,4096,0.018384888768196106
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,2048,16384,0.052946666876475014
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2048,3584,0.00960000024901496
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2048,3584,0.017436444759368896
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,2048,12288,0.045555555158191256
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2048,3072,0.009167999857001835
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2048,3072,0.015644444359673392
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,2048,10240,0.04236266679233975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2048,2560,0.007418666448858049
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2048,2560,0.014697778556081982
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,2048,8192,0.03844177722930908
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2048,2048,0.006671111202902264
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2048,2048,0.014193778236707052
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,2048,7168,0.03795911206139459
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,2048,6144,0.03568355573548211
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2048,1536,0.005471111171775394
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2048,1536,0.01391377713945177
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2048,1024,0.004359999878538979
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2048,1024,0.013497778111033969
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,2048,5120,0.035169776942994856
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2048,768,0.003945777813593547
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2048,768,0.01309333327743742
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,2048,4096,0.0331928895579444
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2048,512,0.003573333223660787
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2048,512,0.012805333567990197
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,2048,3584,0.033344000577926636
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2048,256,0.0032906666811969546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2048,256,0.012297777665985955
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,2048,3072,0.03051911128891839
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2048,128,0.003006222140457895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2048,128,0.012086222569147745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,2048,2560,0.029630220598644678
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2048,64,0.0028417776856157514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,2048,2048,0.02793955471780565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2048,64,0.01201244443655014
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,2048,32,0.0029155556112527847
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,2048,32,0.011869333684444427
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,2048,1536,0.02863022353914049
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,2048,1024,0.02736088964674208
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,2048,768,0.026149332523345947
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1536,16384,0.03511111272705926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1536,16384,0.019595555133289762
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1536,65536,0.05958577659395006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1536,65536,0.10607733329137166
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1536,12288,0.028971556160185073
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1536,10240,0.013882666826248169
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1536,12288,0.015726221932305228
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1536,10240,0.026022222306993272
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1536,8192,0.012225777738624148
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1536,8192,0.023394667439990576
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,2048,512,0.025544888443417017
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1536,7168,0.01148977792925305
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,2048,256,0.023781332704755995
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1536,7168,0.0221377776728736
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1536,6144,0.010515555739402771
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1536,6144,0.020693333612547982
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,2048,128,0.022950222094853718
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1536,5120,0.010497777826256223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1536,5120,0.01908177799648709
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,1536,12288,0.045066667927636035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,1536,16384,0.05205066667662727
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1536,4096,0.009326222042242685
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1536,4096,0.017455110947291057
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1536,3584,0.009184000392754873
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,1536,10240,0.041561779048707746
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1536,3584,0.01577333278126187
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,1536,65536,0.13418400287628174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1536,3072,0.008636444807052612
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1536,3072,0.015321777926550971
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1536,2560,0.0069120000633928515
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1536,2560,0.014704888065656027
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,1536,8192,0.03793955511516995
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1536,2048,0.006137777947717243
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1536,2048,0.014189332723617554
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,1536,7168,0.03790311018625895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1536,1536,0.0052800000541739995
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1536,1536,0.013834666874673633
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,1536,6144,0.035460445615980364
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1536,1024,0.00424177779091729
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1536,1024,0.013184888495339287
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,1536,5120,0.03513688842455546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1536,768,0.0038817777401871155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1536,768,0.012999999854299756
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,1536,4096,0.032431999842325844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1536,512,0.0035253332720862497
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,1536,3584,0.03330844309594896
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1536,512,0.01277599980433782
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1536,256,0.003165333428316646
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1536,256,0.01204177737236023
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,1536,3072,0.03033511175049676
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1536,128,0.00291377781993813
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1536,128,0.012038222617573209
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,1536,2560,0.02884622083769904
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1536,64,0.0027475555737813315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1536,64,0.012011555333932241
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1536,32,0.002798222212327851
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,1536,2048,0.027694221999910142
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1536,32,0.011774222056070963
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1024,65536,0.07789333661397298
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1024,65536,0.03993955585691664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1024,16384,0.014911111858155994
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1024,16384,0.027401778433057997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,1536,1536,0.02819733487235175
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1024,12288,0.01279022213485506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1024,12288,0.024313777685165405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,1536,1024,0.027013333307372198
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1024,10240,0.011557333171367645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1024,10240,0.02250755495495266
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,1536,768,0.025920000341203477
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1024,8192,0.01053600013256073
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1024,8192,0.020798222886191476
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,1536,512,0.02481599979930454
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1024,7168,0.010056888891590966
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1024,7168,0.019840889506869845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,1536,256,0.023158222436904907
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1024,6144,0.009472000102202097
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,1536,128,0.022431999444961548
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1024,6144,0.017794667018784415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1024,5120,0.008955555657545725
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1024,5120,0.016338666280110676
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,1024,65536,0.13368088669247097
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1024,4096,0.008377778033415476
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1024,4096,0.01577333278126187
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,1024,16384,0.05125511023733351
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1024,3584,0.008009778128729926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1024,3584,0.015226667126019796
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,1024,12288,0.0440977778699663
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1024,3072,0.007540444533030192
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1024,3072,0.014735110931926303
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,1024,10240,0.0411119990878635
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1024,2560,0.006654222392373615
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1024,2560,0.014305777019924588
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,1024,8192,0.03771200113826328
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1024,2048,0.005995555470387141
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1024,2048,0.014087999860445658
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,1024,7168,0.037392000357309975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1024,1536,0.005325333525737126
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1024,1536,0.013694222602579327
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,1024,6144,0.035142223040262856
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1024,1024,0.004271111140648524
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1024,1024,0.013259555730554791
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,1024,5120,0.03319022059440613
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1024,768,0.003912000192536248
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1024,768,0.0129439996348487
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,1024,4096,0.031231999397277832
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1024,512,0.003584889074166616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1024,512,0.01275466630856196
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,1024,3584,0.03123289015558031
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1024,256,0.003232888877391815
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1024,256,0.012079999678664737
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,1024,3072,0.029853334029515583
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1024,128,0.0029644444584846497
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1024,128,0.011963555382357704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,1024,2560,0.028953777419196233
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1024,64,0.0027599999060233435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1024,64,0.011752000285519494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,1024,32,0.002893333426780171
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,1024,32,0.011439111497667102
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,768,65536,0.03591022226545546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,768,65536,0.07588711049821642
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,1024,2048,0.02709688908523983
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,768,16384,0.012416889270146688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,768,16384,0.027486221657858953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,1024,1536,0.028382221857706707
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,768,12288,0.011104889214038849
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,768,12288,0.023974221613672044
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,1024,768,0.0258240004380544
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,1024,1024,0.026872888207435608
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,768,10240,0.010108444425794814
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,768,10240,0.022301332818137273
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,768,8192,0.009296000003814697
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,768,8192,0.020230222079488967
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,1024,512,0.024600888291994732
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,768,7168,0.008880889250172509
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,768,7168,0.0185253338681327
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,1024,256,0.023205333285861548
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,768,6144,0.00833777752187517
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,1024,128,0.02308977809217241
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,768,6144,0.017695999807781644
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,768,5120,0.00850666645500395
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,768,5120,0.016395555602179635
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,768,65536,0.13334222634633383
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,768,4096,0.007361777954631382
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,768,4096,0.015579554769727917
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,768,16384,0.05081955591837565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,768,3584,0.007663999994595845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,768,3584,0.015092444088723926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,768,12288,0.044066667556762695
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,768,3072,0.007254222200976477
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,768,3072,0.014771555860837301
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,768,10240,0.040878223048316106
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,768,2560,0.0064204442832205035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,768,2560,0.014167999227841696
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,768,8192,0.03763733307520548
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,768,2048,0.005559111220969095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,768,2048,0.013939556148317126
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,768,7168,0.037018666664759316
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,768,6144,0.03494044476085239
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,768,1536,0.004852444347408083
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,768,1536,0.013497778111033969
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,768,1024,0.0042168887125121225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,768,1024,0.013159111142158508
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,768,5120,0.03250222073660956
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,768,768,0.0038195554580953387
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,768,768,0.012816888590653738
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,768,4096,0.0303111109468672
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,768,512,0.0035182221068276297
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,768,512,0.01257511145538754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,768,3584,0.03068355719248454
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,768,256,0.003104000041882197
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,768,256,0.012000888586044312
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,768,3072,0.030032889710532293
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,768,128,0.002879111096262932
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,768,128,0.011934222446547614
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,768,2560,0.02846222122510274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,768,64,0.0027013334135214486
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,768,64,0.011701333026091257
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,768,2048,0.027183110515276592
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,768,32,0.0027395555128653846
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,768,32,0.011483555866612328
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,512,65536,0.07410044140285917
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,512,65536,0.02720977697107527
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,512,16384,0.010589333044158088
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,512,16384,0.02711733347839779
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,768,1536,0.027997334798177082
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,512,12288,0.00962755580743154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,768,1024,0.02666666607062022
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,512,12288,0.023504000571038988
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,512,10240,0.009100444614887238
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,768,768,0.025906667113304138
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,512,10240,0.020869332883093093
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,512,8192,0.00868444475862715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,512,8192,0.018922666708628338
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,768,512,0.02463111115826501
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,512,7168,0.00832444429397583
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,512,7168,0.017839110559887357
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,768,256,0.022895110978020564
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,512,6144,0.007606222397751278
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,512,6144,0.0169840008020401
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,768,128,0.022524444593323603
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,512,5120,0.00775822252035141
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,512,5120,0.016176889340082806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,512,65536,0.13373511367373997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,512,4096,0.006797333144479328
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,512,4096,0.015226667126019796
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,512,16384,0.05067733261320326
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,512,3584,0.007327111230956183
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,512,3584,0.014871999621391296
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,512,12288,0.04420444369316101
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,512,3072,0.007043555378913879
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,512,3072,0.01441511180665758
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,512,10240,0.04009600149260627
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,512,2560,0.006134222365087933
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,512,2560,0.013968000809351603
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,512,8192,0.03616266780429416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,512,2048,0.005477333234416113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,512,2048,0.013716444373130798
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,512,7168,0.03532888823085361
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,512,1536,0.004809777769777509
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,512,1536,0.013316444224781461
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,512,6144,0.033542222446865506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,512,1024,0.004093333250946469
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,512,1024,0.012970666918489667
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,512,5120,0.03257600135273404
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,512,768,0.0037840000457233856
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,512,768,0.012898666991127862
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,512,4096,0.02996088729964362
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,512,512,0.0033866665843460294
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,512,512,0.012523555921183692
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,512,3584,0.030823998981051978
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,512,256,0.003031111011902491
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,512,256,0.011973333027627734
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,512,3072,0.029620445436901514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,512,128,0.0028693332440323303
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,512,128,0.011855999628702799
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,512,2560,0.028456889920764502
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,512,64,0.0027066665805048416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,512,64,0.011570666399266986
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,512,32,0.0027111110587914786
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,512,32,0.011405333048767514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,256,65536,0.0203448881705602
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,256,65536,0.07190222210354276
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,512,2048,0.027102222045262653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,256,16384,0.00870400004916721
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,256,16384,0.026308443811204698
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,512,1536,0.028041776683595445
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,256,12288,0.007425777614116669
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,256,12288,0.022320000661744013
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,512,1024,0.026742221580611333
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,256,10240,0.007698666718271043
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,256,10240,0.020495999190542433
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,512,768,0.025622222158643935
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,256,8192,0.007449777589903937
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,256,8192,0.01871288816134135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,512,512,0.024525332782003615
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,256,7168,0.007112888826264276
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,512,256,0.023001778456899855
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,256,7168,0.017885333961910672
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,256,6144,0.00675999994079272
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,512,128,0.023080888721677992
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,256,6144,0.016905777984195285
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,256,5120,0.007095110913117726
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,256,5120,0.016024889217482675
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,256,65536,0.133659561475118
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,256,4096,0.006617777877383762
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,256,4096,0.015322667029168872
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,256,16384,0.050103111399544605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,256,3584,0.007090667055712805
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,256,12288,0.042605333858066134
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,256,3584,0.014997333288192749
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,256,3072,0.006748444504208035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,256,3072,0.014454222387737699
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,256,10240,0.03919466667705112
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,256,2560,0.006124444305896759
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,256,2560,0.013960889644092984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,256,8192,0.03511022196875678
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,256,2048,0.005443555613358815
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,256,2048,0.013751111096805997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,256,7168,0.03460622164938185
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,256,1536,0.0047475554876857335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,256,6144,0.03250222073660956
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,256,1536,0.013283555706342062
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,256,1024,0.0040693332751592
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,256,1024,0.012850667039553324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,256,5120,0.0321084459622701
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,256,768,0.0036684444381131064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,256,768,0.012712000144852532
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,256,4096,0.030157334274715845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,256,512,0.003391999958289994
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,256,512,0.012523555921183692
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,256,3584,0.030731555488374498
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,256,256,0.00299466670387321
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,256,256,0.012034667034943899
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,256,3072,0.029751999510659113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,256,2560,0.028705779049131606
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,256,128,0.002838222309947014
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,256,128,0.011686222420798408
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,256,64,0.0027031112048361036
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,256,64,0.011521778172916837
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,256,32,0.002663111107216941
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,256,32,0.011310222248236338
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,128,65536,0.013458666702111563
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,128,65536,0.07321777608659533
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,256,2048,0.027004443936877783
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,128,16384,0.0070088886552386815
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,128,16384,0.025568889247046575
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,256,1536,0.027829332484139338
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,128,12288,0.006682666639486949
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,256,1024,0.026711111267407734
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,128,12288,0.02188266648186578
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,128,10240,0.00685777763525645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,128,10240,0.020219556159443326
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,256,768,0.025544888443417017
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,128,8192,0.006739555547634761
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,128,8192,0.018616888258192275
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,256,512,0.024522667129834492
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,128,7168,0.006583111153708563
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,128,7168,0.017672888106769986
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,256,256,0.022898667388492163
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,128,6144,0.0064737776087390045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,128,6144,0.01701155636045668
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,256,128,0.022801778382725183
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,128,5120,0.006815111057625876
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,128,5120,0.01607288916905721
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,128,65536,0.13373777601453993
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,128,4096,0.006497777584526274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,128,4096,0.01499111122555203
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,128,16384,0.049031111929151744
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,128,3584,0.006918222126033571
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,128,3584,0.01497066683239407
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,128,10240,0.038584887981414795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,128,12288,0.0414266652531094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,128,3072,0.006584888945023219
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,128,3072,0.014316444595654806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,128,2560,0.006040888941950268
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,128,2560,0.013832888669437833
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,128,8192,0.03498311175240411
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,128,2048,0.005395555661784277
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,128,2048,0.01368622233470281
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,128,7168,0.03449511196878221
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,128,1536,0.0047173334492577445
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,128,1536,0.013183111117945777
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,128,6144,0.03243555625279745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,128,1024,0.004031110968854693
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,128,1024,0.01294311136007309
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,128,5120,0.03223022156291538
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,128,768,0.0036968886852264404
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,128,768,0.012720888687504662
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,128,4096,0.029983109898037378
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,128,512,0.0033679999825027254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,128,512,0.01240355521440506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,128,3584,0.03054844339688619
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,128,256,0.00300444434914324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,128,256,0.011976888610257043
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,128,3072,0.029579556650585596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,128,128,0.0027911110470692315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,128,128,0.011766222616036734
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,128,2560,0.028295109669367473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,128,64,0.0026462222966882917
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,128,64,0.011570666399266986
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,128,32,0.002642666714058982
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,128,32,0.011349333657158745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,128,2048,0.02711822258101569
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,64,65536,0.01311466677321328
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,64,16384,0.006631999793979857
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,64,65536,0.07266844643486871
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,64,16384,0.025060445070266724
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,64,12288,0.006541333264774746
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,64,10240,0.006700444552633498
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,64,12288,0.02167644434505039
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,64,8192,0.006618666566080517
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,64,10240,0.020187555087937247
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,64,8192,0.018352889352374606
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,64,7168,0.00646577775478363
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,64,6144,0.0063182223174307085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,64,7168,0.01759288873937395
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,64,5120,0.006601777755551868
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,64,6144,0.01684266659948561
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,64,4096,0.006247111078765657
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,64,5120,0.015760888655980427
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,64,3584,0.006700444552633498
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,64,4096,0.015106666419241162
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,128,1536,0.02806666824552748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,64,3584,0.014846222268210517
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,64,3072,0.0064284445510970215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,64,2560,0.0060159998635451
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,64,3072,0.014159111513031853
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,64,2048,0.0053680001033677
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,64,2560,0.013808888693650564
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,64,2048,0.013415111435784234
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,64,1536,0.004680888934267892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,64,1024,0.003995555556482739
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,64,1536,0.013208000196350945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,64,768,0.003640000190999773
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,64,1024,0.012720000412729053
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,64,512,0.0033173333439562055
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,64,768,0.012437333663304647
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,64,256,0.002967111145456632
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,64,512,0.012347555822796293
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,64,128,0.0027742222365405825
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,64,256,0.011780444118711682
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,64,64,0.0026640000028742682
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,64,128,0.011667555405033959
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,64,32,0.0026337777574857077
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,64,64,0.011416888899273343
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,32,65536,0.011714666253990598
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,64,32,0.01124622258875105
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,32,16384,0.006515555497672822
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,32,65536,0.07188266515731812
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,32,16384,0.02511466708448198
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,32,12288,0.006474666711356904
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,32,12288,0.0217902229891883
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,32,10240,0.006680888848172293
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,32,8192,0.006551110910044775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,32,10240,0.020074667202101815
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,32,8192,0.018405333161354065
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,32,7168,0.006355555521117316
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,32,6144,0.006200888918505774
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,32,7168,0.017609778377744887
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,32,5120,0.006594666590293248
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,32,6144,0.01680177781316969
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,32,4096,0.006171555568774541
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,32,5120,0.015993777248594496
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,32,3584,0.006632888896597757
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,32,4096,0.015056888262430826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,32,3072,0.00629688882165485
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,32,3584,0.014743111199802823
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,32,2560,0.0059573331640826324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,32,3072,0.014284445179833306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,32,2048,0.005250666704442766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,32,2560,0.013820444544156393
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,32,2048,0.013463999662134381
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,32,1536,0.004600888739029567
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,32,1024,0.003943110919660992
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,32,1536,0.013100444442696042
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,32,768,0.003592888928121991
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,32,1024,0.012847999731699625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,32,768,0.012568000290128918
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,32,512,0.0032400000426504347
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,128,1024,0.026532444689008925
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,32,256,0.0029093333416514923
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,32,512,0.01239822225438224
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,32,128,0.0026942222482628296
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,32,256,0.011791999969217511
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,32,64,0.0025475554996066624
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,32,128,0.011640888949235281
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,32,64,0.01146755533085929
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,64,32,32,0.002587555597225825
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,64,32,32,0.011160000330872007
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,128,768,0.02537422213289473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,128,512,0.024303111765119765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,128,256,0.022872888379626807
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,64,128,128,0.022878222995334204
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,65536,12288,0.2157226668463813
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,65536,16384,0.27932267718844944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,65536,10240,0.1836684412426419
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,65536,12288,0.2446071041954888
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,65536,16384,0.29936620924207896
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,65536,12288,0.396417776743571
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,65536,8192,0.15260622236463758
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,65536,10240,0.2108088864220513
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,65536,8192,0.1727377838558621
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,65536,10240,0.33144889937506783
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,65536,7168,0.1348364485634698
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,65536,16384,0.5282506412929958
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,65536,8192,0.266336891386244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,65536,7168,0.1564719941880968
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,65536,6144,0.11860356065962051
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,65536,5120,0.10247111320495605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,65536,5120,0.12255911032358806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,65536,6144,0.13892622788747153
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,65536,4096,0.08618755473030938
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,65536,5120,0.16948444313473174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,65536,6144,0.20155110624101427
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,65536,4096,0.10436711046430801
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,65536,7168,0.23336977428860137
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,65536,4096,0.1367759969499376
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,65536,3584,0.07939555247624715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,65536,3584,0.09569244252310859
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,65536,3584,0.11890578269958496
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,65536,3072,0.07494488689634535
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,65536,3072,0.08698488606346978
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,65536,2560,0.06342844168345134
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,65536,2048,0.056890666484832764
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,65536,3072,0.10418755478329128
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,65536,1536,0.04653333293067085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,65536,2048,0.0669599970181783
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,65536,2560,0.07532444265153673
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,65536,2048,0.07063200076421101
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,65536,1536,0.05472266674041748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,65536,1024,0.04002844293912252
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,65536,1024,0.04909244510862562
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,65536,1536,0.058486223220825195
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,65536,1024,0.03886399997605218
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,65536,2560,0.09831466939714219
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,65536,768,0.03052799900372823
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,65536,768,0.03716266817516751
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,65536,512,0.03352977832158407
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,65536,512,0.02207022243075901
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,65536,512,0.039671109782324895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,65536,768,0.044360889328850635
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,65536,128,0.0083155557513237
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,65536,256,0.028883556524912517
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,65536,128,0.027062222361564636
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,65536,64,0.006603555546866522
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,65536,256,0.03289688958062066
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,65536,64,0.026777777406904433
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,65536,32,0.005435555759403441
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,65536,128,0.029368887344996136
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,65536,32,0.025923555095990498
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,65536,256,0.012482666307025485
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,16384,16384,0.08803377548853557
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,16384,16384,0.1071351104312473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,16384,12288,0.0681191086769104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,16384,12288,0.09559822082519531
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,16384,10240,0.05983289082845052
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,16384,12288,0.08571110831366645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,16384,16384,0.12537777423858643
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,16384,8192,0.0509128901693556
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,16384,10240,0.08103466696209378
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,16384,10240,0.07923199733098348
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,16384,8192,0.0654675563176473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,16384,7168,0.04637155599064297
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,16384,8192,0.06509688827726576
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,16384,7168,0.05784889062245687
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,16384,6144,0.04452622267935011
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,16384,7168,0.06142399708429972
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,16384,6144,0.05033422178692288
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,16384,6144,0.05678400066163805
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,16384,5120,0.03802666730350918
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,16384,65536,0.3056960105895996
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,16384,5120,0.04303822252485487
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,16384,5120,0.05413866705364651
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,16384,4096,0.034033778640959
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,16384,4096,0.04873244298828972
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,16384,4096,0.035367111365000405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,16384,3584,0.03180444571706984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,16384,3072,0.02968177861637539
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,16384,3584,0.03128000100453695
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,16384,3584,0.04610844453175863
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,16384,3072,0.043587555487950645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,16384,2560,0.027521777484152053
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,16384,2560,0.023723555935753718
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,16384,3072,0.027414222558339436
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,16384,2048,0.024687111377716064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,16384,2560,0.04270844327078926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,16384,1536,0.022209778428077698
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,16384,1024,0.02014666630162133
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,16384,1536,0.03611466619703505
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,16384,2048,0.03884977764553494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,16384,2048,0.01984355515903897
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,16384,1536,0.01587288909488254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,16384,1024,0.01129333343770769
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,16384,768,0.009309333231714036
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,16384,65536,0.3318844371371799
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,16384,1024,0.03249955508444045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,16384,768,0.019126221537590027
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,16384,512,0.007374222079912822
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,16384,768,0.031473779016070895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,16384,512,0.017287110288937885
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,16384,512,0.029214223225911457
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,16384,256,0.014555555250909595
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,16384,128,0.0038613333470291565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,16384,256,0.024540444215138752
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,16384,256,0.005604444278611078
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,16384,128,0.01441333360142178
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,16384,64,0.0034799998005231223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,16384,128,0.023364444573720295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,16384,32,0.003629333443111844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,16384,64,0.015240000353919135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,16384,32,0.016100444727473788
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,12288,16384,0.07533955574035645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,12288,16384,0.09094666772418553
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,12288,12288,0.07857955826653375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,12288,12288,0.058331555790371366
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,12288,16384,0.10378044181399876
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,12288,10240,0.054175999429490834
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,12288,12288,0.07385866509543525
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,12288,10240,0.06989955239825778
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,12288,10240,0.06730933321846856
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,12288,8192,0.04483644498719109
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,12288,8192,0.05827644136216906
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,16384,65536,0.5410355461968316
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,12288,8192,0.05433422327041626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,12288,7168,0.0472942226462894
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,12288,7168,0.039890666802724205
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,12288,7168,0.05533688929345873
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,12288,6144,0.03661422265900506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,12288,6144,0.041470223002963595
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,12288,6144,0.05374222331576877
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,12288,5120,0.03423733181423611
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,12288,5120,0.04931555522812737
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,12288,4096,0.030457778109444514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,12288,4096,0.044701334502961904
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,12288,5120,0.03523200088077121
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,12288,4096,0.029050668080647785
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,12288,65536,0.2894035445319282
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,12288,3584,0.02567111121283637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,12288,65536,0.2552773422665066
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,12288,3584,0.02857511242230733
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,12288,3584,0.04291022155019972
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,12288,2560,0.024303111765119765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,12288,3072,0.022664000590642292
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,12288,3072,0.02682311170630985
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,12288,2560,0.039303110705481634
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,12288,2048,0.0232924438185162
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,12288,3072,0.04015200005637275
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,12288,2048,0.03637422124544779
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,12288,2048,0.01641155613793267
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,12288,1024,0.00980177770058314
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,12288,1536,0.020823111136754353
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,12288,2560,0.01963200006220076
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,12288,1536,0.034493333763546415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,12288,1024,0.019380443625979953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,12288,768,0.008260444634490544
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,12288,1024,0.031720889939202204
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,12288,1536,0.01293688929743237
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,12288,512,0.006643555644485686
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,12288,768,0.0181688881582684
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,12288,768,0.03070488903257582
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,12288,512,0.015678221980730694
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,12288,256,0.004278222305907143
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,12288,512,0.025943999489148457
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,12288,256,0.014119111829333834
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,12288,128,0.003679111186001036
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,12288,256,0.0240631103515625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,12288,128,0.013996443814701505
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,12288,64,0.0033617777129014335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,12288,128,0.02314044369591607
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,12288,32,0.003375111147761345
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,12288,64,0.014575110541449653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,12288,32,0.01531377765867445
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,10240,16384,0.06625955634646945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,10240,16384,0.08018310864766438
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,10240,12288,0.053915556934144765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,10240,12288,0.07264800204171075
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,10240,16384,0.09322133329179551
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,10240,12288,0.06977333625157674
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,12288,65536,0.4009697702195909
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,10240,10240,0.047613332668940224
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,10240,10240,0.06113511323928833
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,10240,10240,0.05394488904211256
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,10240,8192,0.04398666818936666
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,10240,8192,0.04340799980693393
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,10240,7168,0.03842844565709432
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,10240,8192,0.05405511127577888
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,10240,7168,0.05264266663127475
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,10240,7168,0.03831911087036133
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,10240,65536,0.22594843970404732
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,10240,6144,0.033575110965304904
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,10240,6144,0.035263111193974815
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,10240,5120,0.032073779238594904
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,10240,6144,0.047839999198913574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,10240,4096,0.02865511178970337
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,10240,5120,0.028387556473414104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,10240,5120,0.046112888389163546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,10240,4096,0.023965333898862202
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,10240,4096,0.042360888587103955
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,10240,3072,0.02511022157139248
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,10240,3584,0.027024888330035742
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,10240,2560,0.02353688908947839
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,10240,3072,0.03795377744568719
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,10240,3584,0.040656887822681
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,10240,3584,0.021353777911927965
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,10240,3072,0.01889600025282966
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,10240,2048,0.021910222040282354
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,10240,2560,0.03743911120626662
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,10240,2560,0.016393777396943834
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,10240,1536,0.010880889164076911
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,10240,2048,0.03562044435077243
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,10240,1536,0.020469332734743755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,10240,1024,0.00850666645500395
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,10240,2048,0.013783111340469785
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,10240,1536,0.033725334538353816
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,10240,1024,0.019000000423855253
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,10240,768,0.007136888802051544
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,10240,768,0.017808889349301655
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,10240,512,0.006217777729034424
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,10240,1024,0.03168088859981961
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,10240,768,0.030375109778510198
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,10240,256,0.0038577777643998465
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,10240,512,0.01514933341079288
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,10240,512,0.025976889663272437
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,10240,128,0.003555555724435382
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,10240,256,0.014129777749379476
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,10240,256,0.02401599950260586
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,10240,128,0.02295200030008952
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,10240,32,0.0032826666202810076
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,10240,64,0.0144533332851198
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,10240,128,0.013887999786270989
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,10240,64,0.003256888853179084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,10240,32,0.015056888262430826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,10240,65536,0.24644888771904838
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,8192,16384,0.053928888506359525
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,8192,16384,0.06995555427339342
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,8192,16384,0.070433775583903
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,8192,12288,0.055211553970972695
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,8192,12288,0.04314488834804959
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,10240,65536,0.35434577200147843
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,8192,12288,0.06009778049257067
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,8192,10240,0.038718223571777344
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,8192,8192,0.03382577829890781
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,8192,10240,0.05571555429034763
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,8192,8192,0.03507644600338406
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,8192,10240,0.04294844468434652
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,8192,65536,0.1708444489373101
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,8192,8192,0.0500204430686103
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,8192,7168,0.03129955463939243
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,8192,6144,0.028888887829250757
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,8192,7168,0.031303998496797346
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,8192,7168,0.049271110031339854
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,8192,6144,0.044703112708197705
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,8192,5120,0.026622222529517278
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,8192,6144,0.027447111076778833
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,8192,4096,0.023541332946883306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,8192,5120,0.04296977652443779
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,8192,65536,0.21286843882666695
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,8192,4096,0.02001066671477424
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,8192,4096,0.03992799917856852
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,8192,3584,0.021778666310840186
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,8192,3072,0.020219556159443326
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,8192,3584,0.03755733370780945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,8192,5120,0.02375288969940609
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,8192,3072,0.036045332749684654
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,8192,2560,0.01923022170861562
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,8192,2560,0.03568622138765123
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,8192,2048,0.018198221921920776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,8192,3584,0.01775911119249132
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,8192,2560,0.01370933320787218
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,8192,3072,0.0170462214284473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,8192,2048,0.03405155407057868
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,8192,1536,0.009431111315886179
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,8192,1536,0.017147555947303772
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,8192,1024,0.0074142225914531285
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,8192,2048,0.011504000259770287
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,8192,1536,0.032277332411872015
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,8192,1024,0.015871110889646742
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,8192,768,0.006450666735569636
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,8192,1024,0.030256887276967365
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,8192,768,0.013812444276279874
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,8192,512,0.005229333208666907
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,8192,768,0.029334220621320937
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,8192,512,0.013178666432698568
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,8192,256,0.003751111113362842
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,8192,512,0.025599110457632277
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,8192,256,0.012479999827014076
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,8192,128,0.003581333491537306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,8192,256,0.023471110396915015
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,8192,128,0.012451555165979596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,8192,64,0.003185777821474605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,8192,128,0.023437332775857713
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,8192,32,0.0033937777496046494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,8192,64,0.012866666747464074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,8192,32,0.013491555220550962
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,7168,16384,0.04932266804907057
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,7168,16384,0.06697866651746961
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,8192,65536,0.2669484350416395
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,7168,12288,0.040413333310021296
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,7168,12288,0.04590310984187656
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,7168,12288,0.05734755595525106
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,7168,16384,0.0595031115743849
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,7168,10240,0.03600177831119961
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,7168,10240,0.039055999782350324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,7168,10240,0.05314933260281881
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,7168,8192,0.03224444389343262
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,7168,8192,0.047581334908803306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,7168,8192,0.03216622273127238
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,7168,7168,0.029521776570214167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,7168,65536,0.15873955355750188
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,7168,6144,0.027309333284695942
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,7168,7168,0.0462337765428755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,7168,7168,0.02847822176085578
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,7168,6144,0.024824000067181055
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,7168,5120,0.025229333175553217
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,7168,4096,0.02165777815712823
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,7168,6144,0.04342755675315857
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,7168,5120,0.04176533222198486
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,7168,5120,0.021581333544519212
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,7168,3584,0.02064533366097344
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,7168,4096,0.01964000033007728
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,7168,4096,0.03814222084151374
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,7168,3584,0.03651288814014859
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,7168,3072,0.019508444600635104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,7168,65536,0.19768444697062174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,7168,3584,0.01627466744846768
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,7168,2560,0.01873333255449931
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,7168,3072,0.03509511219130622
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,7168,2048,0.010524444282054901
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,7168,2560,0.03512977891498142
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,7168,3072,0.014304889572991265
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,7168,2048,0.017681777477264404
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,7168,1536,0.008772444393899705
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,7168,2560,0.012468444804350534
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,7168,1536,0.016585777203241985
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,7168,1024,0.014942222171359591
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,7168,2048,0.03338399860594008
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,7168,1024,0.0070222218831380205
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,7168,1536,0.03288088904486762
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,7168,768,0.006137777947717243
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,7168,512,0.00464088883664873
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,7168,1024,0.030244443151685927
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,7168,768,0.013602666556835175
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,7168,768,0.02755022214518653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,7168,512,0.013057777451144325
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,7168,256,0.0035617777870761026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,7168,512,0.025125333004527624
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,7168,256,0.012460444536474017
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,7168,128,0.0033297776761982176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,7168,64,0.0030942221896515954
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,7168,256,0.023575110567940608
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,7168,128,0.012325333224402534
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,7168,128,0.023222222924232483
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,7168,64,0.012605333493815528
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,7168,32,0.0032311110860771606
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,7168,32,0.013283555706342062
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,6144,16384,0.06286133660210504
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,6144,16384,0.04687111245261299
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,7168,65536,0.2507590982649061
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,6144,16384,0.0531217787000868
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,6144,12288,0.037795556916130915
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,6144,12288,0.04107288850678338
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,6144,12288,0.054788443777296275
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,6144,10240,0.03384622269206577
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,6144,10240,0.05076533224847582
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,6144,10240,0.03535200158754984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,6144,8192,0.03051111102104187
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,6144,8192,0.04595111144913567
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,6144,8192,0.028928889168633357
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,6144,7168,0.029079112741682265
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,6144,7168,0.025438222620222304
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,6144,7168,0.04436444573932224
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,6144,6144,0.02552177839808994
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,6144,6144,0.04187555445565117
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,6144,6144,0.022760000493791368
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,6144,5120,0.0229066660006841
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,6144,5120,0.039776888158586286
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,6144,4096,0.021031111478805542
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,6144,5120,0.019703111714786954
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,6144,4096,0.03674844569630093
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,6144,3584,0.02014400064945221
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,6144,4096,0.01680888897842831
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,6144,3584,0.03566666775279575
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,6144,3072,0.01885777711868286
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,6144,65536,0.18028710948096383
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,6144,3584,0.014884443746672736
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,6144,65536,0.14864888456132677
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,6144,2560,0.017895999881956313
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,6144,2560,0.011415111521879831
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,6144,2048,0.01017066670788659
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,6144,3072,0.03415911065207587
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,6144,3072,0.013162666724787818
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,6144,2560,0.03458666801452637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,6144,1536,0.008448888858159384
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,6144,2048,0.017315554949972365
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,6144,1536,0.016531555189026725
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,6144,1024,0.006620444357395172
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,6144,2048,0.033294220765431724
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,6144,1536,0.0321048895517985
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,6144,1024,0.013999111122555204
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,6144,768,0.005762666463851929
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,6144,1024,0.02975911233160231
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,6144,768,0.013617777989970313
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,6144,512,0.004558222161398994
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,6144,768,0.02665422194533878
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,6144,512,0.012959111067983838
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,6144,256,0.0035413333939181436
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,6144,512,0.02519911030928294
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,6144,256,0.012220444778601328
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,6144,128,0.0032764443506797156
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,6144,256,0.024096000525686476
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,6144,128,0.012145777543385824
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,6144,64,0.003097777772280905
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,6144,128,0.023359111613697473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,6144,32,0.003093333293994268
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,6144,64,0.0125608891248703
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,6144,32,0.012998222476906247
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,5120,16384,0.042781333128611244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,5120,16384,0.05871466795603434
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,6144,65536,0.19660354985131157
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,5120,12288,0.03849777910444472
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,5120,12288,0.034845332304636635
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,5120,16384,0.04820800158712599
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,5120,12288,0.05156355433993869
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,5120,10240,0.03128622306717767
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,5120,10240,0.048602667119767934
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,5120,10240,0.035791112316979304
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,5120,8192,0.023962666591008503
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,5120,8192,0.028275556034511987
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,5120,8192,0.04399822155634562
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,5120,6144,0.023631110787391663
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,5120,7168,0.025755556093321905
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,5120,7168,0.021568889419237774
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,5120,7168,0.042747557163238525
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,5120,6144,0.03982133335537381
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,5120,6144,0.018750222192870248
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,5120,5120,0.021680000755521987
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,5120,4096,0.01979466610484653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,5120,5120,0.03848355677392747
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,5120,5120,0.016695999436908297
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,5120,3584,0.01897333396805657
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,5120,4096,0.03644888930850559
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,5120,4096,0.01408088869518704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,5120,3072,0.01105955574247572
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,5120,3584,0.034741332133611046
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,5120,3584,0.01237777786122428
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,5120,65536,0.13166577286190456
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,5120,3072,0.018208000395033095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,5120,2560,0.009749333063761393
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,5120,3072,0.03380444314744737
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,5120,2560,0.01754666699303521
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,5120,2048,0.008479110896587372
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,5120,2560,0.03384088807635837
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,5120,2048,0.01682577696111467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,5120,1536,0.007320000065697565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,5120,2048,0.03324266605907016
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,5120,1536,0.015671999918089975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,5120,1024,0.006024000131421619
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,5120,1536,0.031594667169782854
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,5120,1024,0.01369955556260215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,5120,768,0.0052844443255000645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,5120,1024,0.027735999888843958
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,5120,768,0.013314666847387949
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,5120,512,0.0038622220357259116
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,5120,768,0.026538666751649644
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,5120,512,0.012867555850081973
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,5120,512,0.025031111306614343
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,5120,256,0.0035173334181308746
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,5120,65536,0.1699288951026069
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,5120,256,0.012223111258612739
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,5120,128,0.0032213332338465583
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,5120,256,0.023728888895776536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,5120,128,0.012118221984969245
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,5120,64,0.0030213333666324615
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,5120,128,0.022665777140193518
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,5120,32,0.0030657777355776895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,5120,64,0.012344000240166983
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,5120,32,0.012705778082211813
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,4096,16384,0.03832622369130453
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,5120,65536,0.19024977419111463
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,4096,16384,0.05614666806326973
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,4096,16384,0.03917333483695984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,4096,12288,0.03136533498764038
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,4096,12288,0.03240711159176297
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,4096,12288,0.04984088738759359
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,4096,10240,0.02938133478164673
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,4096,10240,0.026941332552168105
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,4096,10240,0.04667822188801236
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,4096,8192,0.02570933269129859
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,4096,8192,0.02387911081314087
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,4096,7168,0.02370133333735996
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,4096,8192,0.041858666472964816
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,4096,7168,0.04066488809055752
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,4096,6144,0.021943999661339655
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,4096,7168,0.02083555526203579
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,4096,6144,0.0376782218615214
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,4096,5120,0.020447999238967896
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,4096,6144,0.018803555104467604
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,4096,5120,0.037026666932635836
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,4096,4096,0.01908533275127411
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,4096,5120,0.016720000240537856
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,4096,3584,0.01332266628742218
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,4096,3584,0.01830844415558709
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,4096,4096,0.03556266758177016
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,4096,4096,0.014583110809326172
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,4096,3072,0.012127999630239276
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,4096,3584,0.034493333763546415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,4096,3072,0.01770755483044518
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,4096,65536,0.15458844767676458
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,4096,2560,0.008752889103359645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,4096,65536,0.11366755432552761
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,4096,3072,0.03254400028122796
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,4096,2560,0.01716444392999013
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,4096,2048,0.007717333734035492
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,4096,2560,0.033562666840023465
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,4096,2048,0.016615110966894362
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,4096,1536,0.006706666615274217
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,4096,2048,0.03222844335767958
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,4096,1024,0.013636444177892474
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,4096,1536,0.014464888307783337
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,4096,1024,0.005597333527273602
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,4096,1536,0.02985244327121311
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,4096,768,0.004266666869322459
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,4096,1024,0.02713955607679155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,4096,768,0.013312888642152151
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,4096,512,0.0038782221575578055
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,4096,768,0.02626488937271966
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,4096,512,0.012778666284349231
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,4096,256,0.003383111208677292
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,4096,512,0.02500444485081567
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,4096,256,0.012117333710193634
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,4096,128,0.0031022222505675424
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,4096,256,0.023716444770495098
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,4096,128,0.012183999849690331
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,4096,64,0.0029297777348094513
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,4096,128,0.02349333299530877
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,4096,64,0.01205244412024816
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,4096,32,0.003079111170437601
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,4096,32,0.012688000169065265
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,4096,65536,0.13845511277516684
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3584,16384,0.037081778049468994
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,3584,16384,0.053766220808029175
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3584,16384,0.03537333342764113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3584,12288,0.031121777163611516
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3584,12288,0.028408888313505385
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,3584,12288,0.048410667313469775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3584,10240,0.028183112541834515
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3584,8192,0.025776000486479864
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,3584,10240,0.044813334941864014
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3584,10240,0.025028443998760645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,3584,8192,0.040163556734720864
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3584,7168,0.022693332698610093
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3584,8192,0.021336888273557026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,3584,7168,0.03958488835228814
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3584,6144,0.0214008887608846
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3584,7168,0.019289778338538278
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,3584,6144,0.038104888465669416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3584,5120,0.020153777466879953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3584,6144,0.01740533279048072
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3584,4096,0.013455111119482251
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3584,4096,0.018625777628686693
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3584,5120,0.01536977787812551
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,3584,5120,0.036786665519078575
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3584,65536,0.11076177491082086
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3584,3584,0.012499555945396423
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,3584,4096,0.03470044334729513
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3584,3584,0.018072888255119324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3584,3072,0.0173653331067827
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,3584,3584,0.033598221010631986
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3584,3072,0.011557333171367645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3584,2560,0.008314666648705801
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,3584,3072,0.032369779215918645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3584,2560,0.01681866745154063
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3584,2048,0.0075493330756823225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,3584,2560,0.03306666678852505
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3584,2048,0.015461333923869662
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3584,1536,0.006477333191368315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,3584,65536,0.14503555827670628
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,3584,2048,0.032264000839657254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3584,1536,0.01404888927936554
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3584,1024,0.005388444496525659
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,3584,1536,0.029184887806574505
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3584,1024,0.01367022262679206
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,3584,1024,0.026749332745869953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3584,768,0.004047111090686587
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3584,768,0.01312711089849472
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3584,512,0.003656889001528422
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,3584,768,0.026378666361172993
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3584,512,0.012806221842765808
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3584,256,0.0033119999700122406
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,3584,512,0.02516977820131514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3584,256,0.012260444462299347
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3584,128,0.003002666764789157
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,3584,256,0.02328444520632426
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3584,128,0.011992000043392181
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3584,64,0.0028364445186323593
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,3584,128,0.022608889473809138
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3584,32,0.0029582221888833572
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3584,64,0.012159999873903064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3584,32,0.012628444366984896
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3584,65536,0.12438666820526123
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3072,12288,0.030116445488399927
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3072,16384,0.03666755557060242
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,3072,16384,0.05221688747406006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3072,16384,0.03195200032658047
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3072,10240,0.027706666125191584
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3072,12288,0.025669333007600572
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,3072,12288,0.04665333363744947
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,3072,10240,0.043571554952197604
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3072,8192,0.024088000257809956
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,3072,8192,0.03972355524698893
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3072,7168,0.02243822150760227
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3072,10240,0.022671999202834234
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3072,8192,0.0196142229768965
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,3072,7168,0.03922755519549052
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3072,6144,0.021542222963439092
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3072,7168,0.017664889494578045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3072,5120,0.014269333746698169
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,3072,6144,0.036601778533723615
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3072,5120,0.019827556278970506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3072,65536,0.10926844676335652
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3072,4096,0.012684444586435953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,3072,5120,0.03603644503487481
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3072,6144,0.015867556134859722
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3072,4096,0.018610666195551556
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3072,3584,0.011790222591824003
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3072,3584,0.017774222625626456
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3072,3072,0.010951110886202918
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,3072,4096,0.034471111165152654
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,3072,3584,0.03327644533581204
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3072,3072,0.017475555340449016
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3072,2560,0.007800888684060838
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,3072,3072,0.03272711237271627
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3072,2560,0.016328000360065036
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,3072,2560,0.03307288885116577
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3072,2048,0.007038222418891058
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,3072,65536,0.13850311438242593
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3072,2048,0.014736889137162102
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3072,1536,0.0063386667105886675
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,3072,2048,0.03135022189882066
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3072,1536,0.01423288881778717
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3072,1024,0.0047866664826869965
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,3072,1536,0.028817776176664565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3072,1024,0.013553777502642738
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3072,768,0.013287111288971372
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3072,768,0.004045333299371931
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,3072,1024,0.02708622150950962
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3072,512,0.003633777714437909
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,3072,768,0.026192888617515564
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3072,256,0.0032799999333090256
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3072,512,0.012672888735930124
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,3072,512,0.025136000580257837
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3072,256,0.012155555188655853
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3072,128,0.00299466670387321
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,3072,256,0.02312444481584761
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3072,128,0.011901333100265928
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3072,64,0.0028604444944196274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3072,32,0.002893333426780171
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,3072,128,0.023215111758973863
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3072,64,0.011988444460762871
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,3072,32,0.012416889270146688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,3072,65536,0.10282844305038452
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2560,16384,0.03583555420239767
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,2560,16384,0.051282667451434664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2560,12288,0.03009777930047777
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2560,12288,0.02219377789232466
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2560,16384,0.027727110518349543
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,2560,12288,0.04607199960284763
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2560,10240,0.026963555150561865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,2560,10240,0.04310399956173367
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2560,8192,0.023660444551044043
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,2560,8192,0.03867644402715895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2560,10240,0.019499555230140686
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2560,8192,0.016576889488432143
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2560,7168,0.022314666046036616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2560,6144,0.013721778160995908
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,2560,7168,0.0381279985109965
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2560,6144,0.020991999242040846
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2560,7168,0.015264888604482016
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2560,5120,0.012448888685968188
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,2560,6144,0.036087112294303045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2560,5120,0.019575999842749704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2560,4096,0.011020444333553314
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,2560,5120,0.03604711095492045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2560,4096,0.018434666925006442
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2560,3584,0.01038044442733129
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2560,3584,0.017814222309324477
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,2560,4096,0.03411466545528836
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2560,3072,0.009844444692134857
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2560,65536,0.10809866587320964
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,2560,3584,0.03319022059440613
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2560,3072,0.01683911184469859
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,2560,65536,0.13080533345540366
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2560,2560,0.007667555577225155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2560,2560,0.015217777755525378
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,2560,2560,0.032595554987589516
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,2560,3072,0.03195200032658047
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2560,2048,0.006931555353932911
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2560,2048,0.014653333359294467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2560,1536,0.0058453331391016645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,2560,2048,0.029496000872717962
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2560,1024,0.013529777526855469
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2560,1536,0.01385244478782018
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2560,1024,0.004804444395833545
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,2560,1536,0.028406222661336262
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2560,768,0.004006222304370669
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,2560,1024,0.02643111182583703
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2560,512,0.0036435553597079385
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2560,768,0.01311822235584259
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,2560,768,0.02603911028967963
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2560,512,0.012759110993809171
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2560,256,0.0032746667663256326
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,2560,512,0.02504355543189579
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2560,256,0.012048000262843238
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2560,128,0.0029519999192820657
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,2560,256,0.023228444986873206
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2560,128,0.012054222325483957
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2560,65536,0.09787644280327691
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2560,64,0.0028453332682450614
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,2560,128,0.02307555576165517
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2560,64,0.01202044470442666
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2560,32,0.0028942223224374983
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2560,32,0.012336889074908363
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2048,12288,0.029278222057554457
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2048,16384,0.03481066558096144
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,2048,12288,0.04619644416703118
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,2048,16384,0.05147644546296862
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2048,12288,0.01828444500764211
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2048,10240,0.025983999172846477
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2048,16384,0.022836445106400385
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,2048,10240,0.04270133376121521
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2048,10240,0.016492444607946608
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2048,8192,0.02329866588115692
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2048,7168,0.01348533315791024
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,2048,8192,0.038418667183982
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2048,8192,0.014852444330851236
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2048,7168,0.021783111823929682
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,2048,7168,0.03814399904674954
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2048,65536,0.10282933049731785
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2048,6144,0.012551999754375882
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2048,6144,0.02049511174360911
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2048,5120,0.011525332927703857
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,2048,6144,0.03550488750139872
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2048,5120,0.01942933268017239
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2048,4096,0.010266666611035665
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,2048,5120,0.03551822238498264
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2048,3584,0.008849778109126622
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2048,4096,0.01813155578242408
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,2048,65536,0.13744355572594538
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2048,3072,0.00814844419558843
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2048,3584,0.01718577742576599
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,2048,4096,0.033972445461485125
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2048,3072,0.01570755574438307
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,2048,3584,0.03315644462903341
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2048,2560,0.007476444045702617
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,2048,3072,0.032015111711290144
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2048,2560,0.01484266585773892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2048,2048,0.006741333338949416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,2048,2560,0.03200177682770623
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2048,2048,0.014563555518786112
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2048,1536,0.0059057776298787855
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,2048,2048,0.029320889049106177
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2048,1536,0.013758222262064615
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2048,1024,0.004815111143721474
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2048,1024,0.01332711097266939
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,2048,1536,0.028397334946526423
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2048,768,0.004064889003833135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,2048,1024,0.02646488944689433
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2048,768,0.013030221892727746
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2048,512,0.0036720000207424164
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2048,65536,0.07047288947635226
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,2048,768,0.02609155575434367
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2048,512,0.012607110871209038
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2048,256,0.0032791110376516977
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,2048,512,0.02464888824356927
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2048,256,0.012034667034943899
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2048,128,0.002970666728085942
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,2048,256,0.02385955552260081
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2048,128,0.01182577759027481
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2048,64,0.0028035555862718155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,2048,128,0.022648000054889258
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,2048,32,0.0028862222615215513
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2048,64,0.01183999992079205
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,2048,32,0.012320888539155325
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1536,16384,0.03304800060060289
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1536,12288,0.027965333726671007
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,1536,16384,0.04888622297181023
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,1536,12288,0.044116444057888456
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1536,10240,0.014000889327791003
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1536,12288,0.015683554940753512
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1536,16384,0.019440000255902607
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1536,10240,0.02531555626127455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1536,8192,0.01216533366176817
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,1536,10240,0.04105510976579454
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1536,8192,0.022590221630202398
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1536,7168,0.011504888534545898
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,1536,8192,0.037547555234697126
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1536,7168,0.021739555729760066
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1536,6144,0.010588444769382477
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1536,6144,0.020192889703644645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1536,65536,0.09903644190894233
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,1536,7168,0.03708088729116652
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1536,5120,0.009840889109505547
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,1536,65536,0.12337333626217312
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,1536,6144,0.03544266687499152
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1536,5120,0.01884355478816562
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1536,4096,0.009201777478059133
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1536,4096,0.01716977854569753
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1536,3584,0.0162782222032547
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1536,3584,0.009009777671760982
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,1536,5120,0.03520800007714166
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1536,3072,0.01499466598033905
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,1536,4096,0.03348088926739163
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1536,3072,0.007981333467695449
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,1536,3584,0.031360887818866305
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1536,2560,0.007102222078376346
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,1536,3072,0.030125333203209773
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1536,2560,0.014705777168273926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1536,2048,0.006141333530346553
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,1536,2560,0.029863112502627905
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1536,2048,0.014340443743599786
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1536,1536,0.004909333255555895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,1536,1536,0.028176890479193792
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,1536,2048,0.029417776399188574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1536,1536,0.013749333719412485
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1536,1024,0.0042462220622433555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1536,1024,0.013191999660597907
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1536,768,0.0038817777401871155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,1536,1024,0.02637422250376807
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1536,768,0.01296533313062456
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1536,512,0.003504000190231535
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,1536,768,0.026226666238572862
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1536,512,0.01260088880856832
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1536,256,0.003167111012670729
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,1536,512,0.02460177739461263
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1536,65536,0.05796977546479967
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1536,256,0.01223288890388277
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1536,128,0.0029351111087534162
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,1536,256,0.022970666488011677
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1536,128,0.011889778077602386
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1536,64,0.0027457777824666765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,1536,128,0.023053333163261414
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1536,32,0.002809777855873108
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1536,64,0.011853333148691388
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1536,32,0.012223111258612739
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1024,16384,0.026258667310078938
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1024,12288,0.012847111456924014
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,1024,16384,0.04821155468622843
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1024,12288,0.023239110906918842
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1024,10240,0.01165688865714603
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1024,16384,0.015065777632925244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,1024,12288,0.04361866580115425
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1024,10240,0.021242666575643752
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1024,8192,0.010647111468844943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1024,8192,0.019693333241674636
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,1024,10240,0.04074044360054864
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1024,65536,0.07007110781139798
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1024,7168,0.009647111097971598
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,1024,8192,0.03709244396951463
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1024,7168,0.018906666172875296
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1024,6144,0.010427555276287926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,1024,7168,0.037089778317345515
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,1024,65536,0.12242755624983047
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1024,6144,0.01814133259985182
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1024,5120,0.009757333331637913
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,1024,6144,0.034966223769717746
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1024,5120,0.01627111103799608
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1024,4096,0.008981333010726506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,1024,5120,0.033082667324278094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1024,4096,0.015467555986510383
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1024,65536,0.04140355520778232
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1024,3584,0.008640000389681922
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,1024,4096,0.03121244576242235
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1024,3584,0.015112888481881885
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1024,3072,0.008160889148712158
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,1024,3584,0.030392888519499037
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1024,3072,0.01479466590616438
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,1024,2560,0.02951288885540432
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1024,2560,0.006937777830494775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,1024,3072,0.029206222958034937
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1024,2560,0.014308444327778287
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1024,2048,0.00620888877246115
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1024,2048,0.014002665877342224
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1024,1536,0.004919111314747068
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,1024,2048,0.02845777736769782
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1024,1536,0.013603555659453074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1024,1024,0.004224888980388641
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,1024,1536,0.02786311176088121
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1024,1024,0.01292266696691513
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1024,768,0.0038853333228164246
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,1024,1024,0.026112889250119526
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1024,768,0.01297511077589459
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1024,512,0.003540444291300244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,1024,768,0.025880000657505457
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1024,512,0.012512889173295764
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,1024,512,0.024654222859276667
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1024,256,0.0031893334041039148
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,1024,128,0.023141332798533972
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1024,256,0.01184266640080346
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1024,128,0.002903999967707528
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,1024,256,0.02294844388961792
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1024,128,0.01183733344078064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1024,64,0.0027253333893087176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,1024,32,0.0028640000770489373
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1024,64,0.011699555648697747
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,1024,32,0.01200355589389801
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,768,16384,0.01238666640387641
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,768,16384,0.025800888737042744
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,768,12288,0.01073155552148819
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,768,12288,0.022822222775883142
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,768,16384,0.047800001170900136
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,768,10240,0.009951999617947472
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,768,12288,0.04359555575582716
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,768,10240,0.02108000053299798
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,768,8192,0.009278222090668148
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,768,10240,0.040757334894604154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,768,8192,0.01906222270594703
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,768,7168,0.00869688888390859
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,768,8192,0.037019554111692644
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,768,65536,0.12182577451070149
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,768,65536,0.06852266523573133
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,768,7168,0.01810933318403032
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,768,6144,0.009128000173303816
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,768,7168,0.0367253323396047
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,768,6144,0.016827555166350473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,768,5120,0.008558221989207799
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,768,65536,0.03498488995763991
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,768,6144,0.034352001216676496
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,768,5120,0.01612444387541877
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,768,4096,0.008216000265545314
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,768,5120,0.03280622098180983
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,768,4096,0.015455999308162265
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,768,3584,0.0075786668393347
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,768,4096,0.030773331721623737
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,768,3584,0.01499288943078783
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,768,3072,0.007017778025733099
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,768,3584,0.02995377779006958
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,768,3072,0.014581332604090372
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,768,2560,0.006616000086069107
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,768,3072,0.028941333293914795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,768,2560,0.014160888062583076
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,768,2048,0.005572444448868434
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,768,2560,0.029529776838090684
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,768,2048,0.013783111340469785
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,768,1536,0.00489066665371259
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,768,2048,0.02814133299721612
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,768,1536,0.013376000026861826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,768,1024,0.0041777777175108595
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,768,1536,0.02792888879776001
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,768,1024,0.013007111019558378
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,768,768,0.0038355555799272326
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,768,1024,0.0259244441986084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,768,768,0.012799999780125089
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,768,512,0.003463111196955045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,768,768,0.025839999318122864
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,768,512,0.01242844429281023
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,768,256,0.0030942221896515954
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,768,512,0.02443555494149526
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,768,128,0.01163644426398807
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,768,256,0.012048000262843238
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,768,128,0.0028773333049482773
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,768,256,0.023192000057962205
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,768,64,0.0027502222607533136
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,768,128,0.022311111291249592
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,768,32,0.0028088889602157804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,768,64,0.011471110913488599
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,768,32,0.011872000164455838
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,512,16384,0.010447111394670276
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,512,16384,0.02530577778816223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,512,12288,0.010272888673676385
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,512,12288,0.02161066730817159
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,512,16384,0.04766755633884006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,512,10240,0.009557333257463243
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,512,65536,0.06672800249523587
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,512,12288,0.043453332450654775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,512,10240,0.019838222199016146
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,512,8192,0.008756444685988957
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,512,10240,0.04059911105367873
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,512,7168,0.008427555362383524
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,512,8192,0.017993777990341187
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,512,65536,0.12200266785091823
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,512,8192,0.035162667433420815
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,512,7168,0.017148445049921673
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,512,6144,0.007753777835104201
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,512,65536,0.025980444418059453
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,512,6144,0.03264266583654616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,512,7168,0.03639111253950331
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,512,6144,0.016376889414257474
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,512,5120,0.007536888950400882
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,512,5120,0.015593777100245157
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,512,4096,0.0069013333155049225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,512,5120,0.03226222263442145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,512,4096,0.014876445134480795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,512,3584,0.007229333122571309
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,512,3584,0.014704000618722705
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,512,4096,0.030735111898846094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,512,3072,0.006844444407357111
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,512,3584,0.030014223522610132
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,512,3072,0.014173333843549093
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,512,2560,0.006164444403515921
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,512,3072,0.029109332296583388
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,512,2560,0.013818666338920593
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,512,2048,0.00545155546731419
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,512,2560,0.029306666718588933
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,512,2048,0.013630222115251752
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,512,1536,0.004776000148720211
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,512,2048,0.02829866607983907
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,512,1536,0.013398222625255585
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,512,1024,0.004074666649103165
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,512,1536,0.027683556079864502
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,512,1024,0.012902222573757172
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,512,1024,0.02581866747803158
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,512,768,0.00370755543311437
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,512,768,0.012648888760142855
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,512,512,0.003403555601835251
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,512,768,0.02570488883389367
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,512,512,0.012356444365448423
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,512,256,0.00305600009030766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,512,512,0.024159111910396155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,512,256,0.011991110940774282
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,512,128,0.0028204443968004654
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,512,256,0.02281155520015293
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,512,128,0.011596444580290051
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,512,64,0.002640888922744327
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,512,128,0.02292888859907786
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,512,32,0.0026906666656335196
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,512,64,0.011569778124491373
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,512,32,0.011727111207114326
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,256,16384,0.008832000195980072
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,256,65536,0.06560533576541476
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,256,16384,0.024884444144037034
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,256,12288,0.007994666695594788
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,256,65536,0.1223413281970554
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,256,65536,0.019761777586407132
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,256,16384,0.04727200004789564
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,256,12288,0.0211217784219318
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,256,10240,0.00738933351304796
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,256,12288,0.042111999458736844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,256,10240,0.019241778386963736
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,256,8192,0.007045333584149678
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,256,10240,0.03837600019243028
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,256,8192,0.01789155602455139
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,256,7168,0.006811555474996567
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,256,8192,0.03459555572933621
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,256,7168,0.017099555995729234
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,256,6144,0.006608888920810487
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,256,7168,0.034098668230904475
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,256,6144,0.016231111354298063
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,256,5120,0.0070426662762959795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,256,6144,0.032602667808532715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,256,5120,0.015576889117558798
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,256,4096,0.006578666468461354
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,256,5120,0.03239377670817905
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,256,4096,0.014760888285107084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,256,3584,0.007062222394678328
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,256,4096,0.03070311082734002
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,256,3584,0.014712888333532544
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,256,3072,0.00675999994079272
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,256,3584,0.02977777851952447
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,256,3072,0.014149333039919535
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,256,2560,0.006088888893524806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,256,3072,0.028577778074476454
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,256,2560,0.013960889644092984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,256,2048,0.005467555589146084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,256,2560,0.029007110330793593
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,256,2048,0.01352266636159685
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,256,1536,0.004780444420046276
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,256,2048,0.028547555208206177
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,256,1536,0.013177778157922955
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,256,1024,0.004077333129114575
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,256,1536,0.027344889110989038
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,256,1024,0.012785777449607849
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,256,768,0.003687111039956411
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,256,1024,0.025623111261261836
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,256,768,0.012641777594884237
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,256,512,0.0033884443756606844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,256,768,0.025461332665549383
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,256,512,0.01240711079703437
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,256,256,0.0029964444951878656
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,256,512,0.024028443627887305
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,256,256,0.011929777761300405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,256,128,0.0028311111446883944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,256,256,0.022836445106400385
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,256,128,0.011583111352390714
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,256,64,0.0026657777941889236
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,256,128,0.023021333747439917
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,256,32,0.0026968889352348116
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,256,64,0.011519110865063138
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,256,32,0.011726222104496427
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,128,65536,0.013450666434235044
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,128,65536,0.06423822376463148
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,128,16384,0.0070471109615431885
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,128,12288,0.02095111045572493
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,128,16384,0.023784000012609694
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,128,12288,0.006717333363162146
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,128,65536,0.1220995585123698
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,128,10240,0.0068862222962909276
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,128,16384,0.04516444603602091
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,128,10240,0.0192275560564465
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,128,12288,0.04099644554985894
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,128,8192,0.006728888799746831
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,128,10240,0.038112001286612615
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,128,8192,0.017708443933063082
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,128,7168,0.006685333533419504
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,128,8192,0.03453333179155985
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,128,7168,0.01718133356836107
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,128,6144,0.006498666687144174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,128,7168,0.034182220697402954
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,128,6144,0.016318221886952717
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,128,5120,0.006867555694447623
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,128,6144,0.03231199913554721
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,128,5120,0.015415999624464246
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,128,4096,0.00646311127477222
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,128,5120,0.03204355637232462
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,128,3584,0.02979555394914415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,128,4096,0.014716444744004143
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,128,3584,0.006948444578382704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,128,4096,0.03073955575625102
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,128,3584,0.014559111661381192
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,128,3072,0.006599999964237213
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,128,3072,0.014169777433077494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,128,2560,0.006091555373536215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,128,3072,0.02888622217708164
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,128,2560,0.013797333670987023
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,128,2048,0.005379555539952383
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,128,2560,0.029185778564876978
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,128,2048,0.01353422221210268
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,128,1536,0.0047155556579430895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,128,2048,0.028090665737787884
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,128,1536,0.013234666652149625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,128,1024,0.004022222426202563
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,128,1536,0.027560000618298847
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,128,512,0.0033297776761982176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,128,1024,0.012871111432711283
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,128,768,0.0036355555057525635
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,128,1024,0.025581333372328017
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,128,768,0.012658666405412885
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,128,768,0.025395555628670588
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,128,512,0.012215110990736218
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,128,256,0.0029813332690132987
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,128,512,0.023988443944189284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,128,256,0.011955555942323474
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,128,128,0.0027466666781240036
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,128,256,0.022982221510675218
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,128,128,0.011705777711338468
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,128,64,0.0026604444202449587
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,48,128,128,0.0222746663623386
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,128,64,0.011424000064531961
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,128,32,0.0026293332791990707
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,128,32,0.011674666570292579
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,64,16384,0.023855111665195886
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,64,65536,0.012675555215941535
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,64,16384,0.006683555742104848
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,64,65536,0.06428888771269056
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,64,12288,0.006502222269773483
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,64,10240,0.006678222368160884
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,64,12288,0.020607999629444547
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,64,8192,0.006599999964237213
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,64,10240,0.019156444403860304
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,64,7168,0.006500444478458828
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,64,8192,0.017841777867741056
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,64,6144,0.006406222366624408
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,64,7168,0.01703288820054796
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,64,5120,0.006829333388143116
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,64,6144,0.016102222932709586
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,64,5120,0.015340444114473132
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,64,4096,0.00638488887084855
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,64,3584,0.007080888582600488
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,64,4096,0.01475288967291514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,64,3072,0.006518222391605377
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,64,3584,0.014479999740918478
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,64,2560,0.006033777776691649
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,64,3072,0.01402666668097178
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,64,2048,0.0053377776510185665
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,64,2560,0.013751111096805997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,64,1536,0.0047128887640105355
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,64,2048,0.013487111363146039
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,64,1024,0.003974222060706881
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,64,1536,0.013099555340078143
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,64,768,0.003640000190999773
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,64,1024,0.012703110774358114
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,64,512,0.0033084443873829315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,64,768,0.012461333639091916
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,64,256,0.002951111023624738
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,64,512,0.012272888587580787
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,64,128,0.002777777819169892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,64,256,0.011808888779746162
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,64,64,0.0026488889836602737
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,64,128,0.011593778100278644
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,64,64,0.011309333145618439
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,64,32,0.002640888922744327
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,32,16384,0.023896889554129705
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,32,65536,0.011680888632933298
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,64,32,0.01164177805185318
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,32,16384,0.006592888798978593
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,32,65536,0.06367733081181844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,32,12288,0.006443555570311016
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,32,10240,0.006623111251327727
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,32,12288,0.020711110697852243
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,32,8192,0.006552888701359431
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,32,10240,0.019362666540675692
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,32,7168,0.006462222172154321
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,32,5120,0.015499555402331881
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,32,8192,0.017851556340853374
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,32,6144,0.006325333482689328
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,32,7168,0.01698844465944502
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,32,6144,0.016247999336984422
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,32,3584,0.014449778530332776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,32,5120,0.006744888921578725
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,32,3072,0.01407911048995124
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,32,4096,0.006346666564544042
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,32,4096,0.014755555325084262
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,32,3584,0.006706666615274217
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,32,3072,0.006337777607970768
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,32,2560,0.005976000179847081
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,32,2560,0.013815999858909182
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,32,2048,0.0052871112194326185
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,32,1536,0.0046977777447965415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,32,2048,0.013381332986884646
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,32,1536,0.013103110922707452
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,32,1024,0.003951999876234266
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,32,768,0.0035751110149754416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,32,1024,0.012697777814335294
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,32,768,0.012524444195959302
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,32,512,0.0032622222271230486
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,32,64,0.0025920000755124623
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,32,256,0.0029102222373088202
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,32,512,0.012291555603345236
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,32,128,0.0027537778433826235
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,32,256,0.01181244436237547
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,32,128,0.01163200040658315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,48,32,32,0.002611555573013094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,32,64,0.011392888923486074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,48,32,32,0.011640888949235281
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,65536,12288,0.2553759945763482
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,65536,10240,0.22024710973103842
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,65536,16384,0.3246106571621365
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,65536,12288,0.24129067526923287
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,65536,16384,0.29623643557230633
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,65536,8192,0.16415288713243273
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,65536,12288,0.39514933692084414
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,65536,10240,0.20778222878774008
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,65536,10240,0.3298533227708605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,65536,8192,0.2657866742875841
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,65536,8192,0.1709137757619222
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,65536,7168,0.1618728902604845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,65536,16384,0.5266124407450358
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,65536,6144,0.1418693330552843
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,65536,7168,0.15553866492377386
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,65536,5120,0.12517866823408338
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,65536,7168,0.23225866423712838
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,65536,6144,0.13797688484191895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,65536,6144,0.19986044036017525
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,65536,5120,0.12074932787153457
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,65536,4096,0.09779022137324016
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,65536,3584,0.08984000153011745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,65536,4096,0.096705781088935
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,65536,3072,0.07660622066921659
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,65536,3584,0.09556266996595596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,65536,3584,0.13371377521091038
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,65536,4096,0.13598667250739202
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,65536,5120,0.16823111640082467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,65536,2560,0.06805244419309828
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,65536,3072,0.10299911101659139
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,65536,3072,0.08598577976226807
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,65536,2560,0.0739920006857978
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,65536,1536,0.04764266808827718
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,65536,2048,0.06316622098286946
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,65536,2560,0.08599911133448283
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,65536,1536,0.056193775600857206
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,65536,2048,0.06845866971545748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,65536,1024,0.040941334433025785
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,65536,1536,0.05367733372582329
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,65536,768,0.037075555986828275
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,65536,512,0.03273955649799771
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,65536,2048,0.06999378071890937
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,65536,768,0.029366221692827012
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,65536,1024,0.047543111774656505
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,65536,256,0.012025777664449481
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,65536,1024,0.037816888756222196
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,65536,256,0.028396444188223943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,65536,128,0.007928888830873702
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,65536,512,0.021243555678261652
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,65536,256,0.03279377685652839
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,65536,768,0.04248355494605171
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,65536,128,0.027331555883089702
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,65536,64,0.006226666685607698
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,65536,512,0.03723288906945123
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,65536,128,0.029055999384986028
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,65536,32,0.00498488876554701
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,65536,64,0.025635555386543274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,65536,32,0.025231111380789015
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,16384,12288,0.07588799794514973
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,16384,16384,0.09771111276414658
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,16384,16384,0.10344266891479492
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,16384,12288,0.08559911118613349
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,16384,12288,0.09577333264880711
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,16384,10240,0.06670044528113471
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,16384,16384,0.12601599428388807
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,16384,10240,0.08134133285946317
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,16384,8192,0.05789599816004435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,16384,10240,0.07623910903930664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,16384,8192,0.06680444214079115
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,16384,7168,0.05133510960472954
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,16384,7168,0.061308443546295166
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,16384,8192,0.06537689103020562
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,16384,7168,0.05853422482808431
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,16384,6144,0.045796444018681846
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,16384,6144,0.05021333363321093
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,16384,5120,0.03998400105370416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,16384,6144,0.056333336565229625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,16384,5120,0.05219022101826138
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,16384,4096,0.03537155522240533
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,16384,4096,0.03527466787232293
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,16384,5120,0.04283377859327528
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,16384,4096,0.04686044322119819
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,16384,65536,0.3337911234961615
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,16384,3072,0.044024000565210976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,16384,3584,0.03274044394493103
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,16384,3072,0.030247999562157526
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,16384,3584,0.030979557169808283
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,16384,3584,0.04600444436073303
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,16384,2560,0.02829866607983907
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,16384,2048,0.02530044482813941
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,16384,2560,0.04228711128234863
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,16384,3072,0.027286221583684284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,16384,2048,0.019499555230140686
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,16384,2560,0.023864888482623633
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,16384,1536,0.02239111065864563
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,16384,2048,0.038132445679770574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,16384,1024,0.02015022271209293
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,16384,1536,0.03496711121665107
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,16384,1536,0.01550222271018558
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,16384,768,0.01016977760526869
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,16384,1024,0.031974222924974226
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,16384,1024,0.011232888533009423
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,16384,512,0.007248000138335758
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,16384,768,0.019394665956497192
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,16384,512,0.01723022262255351
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,16384,768,0.0302817788388994
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,16384,256,0.00500177757607566
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,16384,256,0.014440889159838358
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,16384,512,0.02814488940768772
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,16384,256,0.024106666445732117
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,16384,128,0.003703111161788305
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,16384,128,0.014173333843549093
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,16384,64,0.0033048888047536216
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,16384,128,0.02309155629740821
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,16384,64,0.015082667271296183
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,16384,32,0.014848000473446317
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,16384,32,0.0034924443397257063
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,16384,65536,0.3210417694515652
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,12288,16384,0.08268444405661689
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,12288,16384,0.08798399898740981
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,12288,16384,0.10604355732599895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,12288,12288,0.06613155868318346
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,12288,12288,0.08060444725884332
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,12288,12288,0.07175199853049384
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,12288,10240,0.05875555674235026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,12288,10240,0.06886221965154012
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,12288,10240,0.06807733244366117
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,12288,8192,0.04955377843644884
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,12288,8192,0.0546648899714152
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,12288,8192,0.058041777875688344
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,12288,65536,0.2782951196034749
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,12288,7168,0.048545777797698975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,12288,7168,0.04391111267937554
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,12288,6144,0.041272001134024724
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,12288,7168,0.05488711264398363
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,12288,6144,0.04190488987498813
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,12288,6144,0.05306844578848945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,12288,5120,0.035996443695492215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,12288,65536,0.28076622221204967
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,12288,5120,0.04801422357559204
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,12288,4096,0.04358844293488396
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,12288,4096,0.033597333563698664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,12288,5120,0.03553244471549988
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,12288,3584,0.04258222381273905
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,12288,3584,0.03125066558519999
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,12288,3584,0.025628444221284654
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,12288,4096,0.02901600135697259
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,12288,3072,0.04091822107632955
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,12288,2560,0.027264888087908428
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,12288,2560,0.03842844565709432
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,12288,2048,0.01626666718059116
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,12288,3072,0.028283556302388508
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,12288,2048,0.023623110519515142
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,12288,1536,0.012878222597969903
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,12288,3072,0.022721777359644573
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,12288,2560,0.019498666127522785
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,12288,1536,0.021428444319301184
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,12288,1024,0.00960177762640847
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,12288,1536,0.03342666559749179
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,12288,2048,0.035911109712388777
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,12288,1024,0.019433778193261888
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,12288,768,0.008061333662933773
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,12288,512,0.0064080001579390625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,12288,768,0.01829866733815935
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,12288,512,0.01555199921131134
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,12288,1024,0.031155556440353394
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,12288,768,0.029746668206320867
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,12288,256,0.0041039999988343985
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,12288,512,0.027783112393485174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,12288,256,0.014120888378885059
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,12288,128,0.003548444559176763
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,12288,256,0.023401778605249193
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,12288,128,0.014035556051466199
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,12288,64,0.0031964443624019623
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,12288,128,0.023567110300064087
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,12288,64,0.014530667000346713
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,12288,32,0.003237333355678452
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,12288,32,0.014318222800890604
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,16384,65536,0.5350453588697646
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,10240,16384,0.07291022274229261
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,10240,16384,0.07800533374150594
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,10240,12288,0.05660621987448799
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,12288,65536,0.39962132771809894
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,10240,16384,0.09070310990015666
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,10240,12288,0.06471288866466947
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,10240,12288,0.07081955671310425
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,10240,10240,0.05116711060206095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,10240,10240,0.06127555502785576
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,10240,10240,0.054732445213529796
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,10240,8192,0.04360888732804192
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,10240,8192,0.053856889406840004
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,10240,8192,0.044109334548314415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,10240,7168,0.04018400112787882
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,10240,7168,0.03899111019240485
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,10240,6144,0.035734222994910345
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,10240,7168,0.051938666237725147
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,10240,6144,0.034154666794670954
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,10240,5120,0.0324924455748664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,10240,6144,0.047451555728912354
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,10240,5120,0.029045333464940388
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,10240,5120,0.04478311207559374
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,10240,4096,0.029046220911873713
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,10240,4096,0.04099199838108487
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,10240,4096,0.023982221881548565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,10240,3584,0.027344889110989038
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,10240,65536,0.24084621005588105
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,10240,65536,0.24024534225463867
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,10240,3072,0.025379555092917547
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,10240,3584,0.04097955425580343
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,10240,3072,0.03952088952064514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,10240,3072,0.018876444962289598
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,10240,3584,0.021516443954573736
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,10240,2560,0.023641778363121882
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,10240,2560,0.036751111348470054
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,10240,2048,0.02204888893498315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,10240,1536,0.010838222172525195
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,10240,2048,0.03436533278889126
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,10240,1536,0.020454222957293194
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,10240,1024,0.00832799987660514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,10240,1536,0.03234666585922241
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,10240,2560,0.016399110356966656
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,10240,1024,0.0188044442070855
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,10240,1024,0.03085600005255805
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,10240,768,0.007167999943097432
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,10240,2048,0.01351466692156262
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,10240,768,0.016739555531077914
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,10240,512,0.00590133335855272
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,10240,768,0.029789331886503432
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,10240,512,0.015812445018026564
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,10240,256,0.003834666477309333
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,10240,512,0.024732444021436904
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,10240,256,0.014084445105658637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,10240,128,0.003370666669474708
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,10240,256,0.0241333329015308
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,10240,128,0.013930666777822705
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,10240,64,0.0031111112071408164
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,10240,128,0.023399111297395494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,10240,64,0.01423466702302297
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,10240,32,0.003168888803985384
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,10240,32,0.014051554931534661
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,8192,16384,0.05904089079962837
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,8192,16384,0.07096978028615315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,8192,12288,0.054246223635143705
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,8192,12288,0.04657511247528923
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,8192,12288,0.06038933330112033
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,8192,16384,0.06929066446092394
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,8192,10240,0.04116977916823493
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,10240,65536,0.38131022453308105
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,8192,10240,0.042585776911841504
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,8192,8192,0.03593511051601834
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,8192,8192,0.05096977618005541
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,8192,10240,0.055161777469846934
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,8192,7168,0.03272622161441379
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,8192,8192,0.03533422284656101
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,8192,7168,0.048179556926091514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,8192,65536,0.1791679991616143
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,8192,7168,0.030918220678965252
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,8192,6144,0.03043022089534336
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,8192,6144,0.027105778455734253
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,8192,4096,0.023834667272037927
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,8192,5120,0.028016000986099243
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,8192,6144,0.04540266593297323
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,8192,5120,0.04239111145337423
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,8192,4096,0.03816711240344577
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,8192,5120,0.023548444112141926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,8192,3584,0.022876444790098403
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,8192,4096,0.019814221395386588
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,8192,3072,0.01555288831392924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,8192,3584,0.0389395554860433
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,8192,3072,0.02090577781200409
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,8192,2560,0.0199946661790212
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,8192,3072,0.03790666659673055
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,8192,2560,0.03515911102294922
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,8192,3584,0.017656889226701524
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,8192,2048,0.01881511178281572
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,8192,1536,0.009266667068004608
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,8192,65536,0.20791910754309761
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,8192,2560,0.013495999905798169
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,8192,2048,0.0331857767370012
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,8192,2048,0.011296889020336999
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,8192,1536,0.017394666870435078
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,8192,1024,0.007361777954631382
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,8192,1536,0.0318159990840488
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,8192,1024,0.029140445921156142
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,8192,768,0.006457777900828256
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,8192,1024,0.016150222884284127
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,8192,768,0.014477334088749357
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,8192,512,0.00519466648499171
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,8192,768,0.02799644403987461
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,8192,512,0.013343999783198038
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,8192,256,0.0036764442920684814
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,8192,512,0.024375110864639282
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,8192,256,0.01255022237698237
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,8192,128,0.00327822214199437
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,8192,256,0.02365600069363912
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,8192,128,0.01236977759334776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,8192,64,0.003104888937539525
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,8192,128,0.02277066641383701
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,8192,64,0.012543111211723752
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,8192,32,0.0032337777730491427
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,8192,32,0.012383999923865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,7168,16384,0.05576533079147339
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,7168,16384,0.07090844710667928
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,7168,16384,0.059528887271881104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,7168,12288,0.04518844352828132
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,7168,12288,0.04562044474813673
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,7168,10240,0.03842844565709432
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,7168,12288,0.05695555607477824
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,7168,10240,0.03901955485343933
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,7168,10240,0.0551111102104187
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,7168,8192,0.034486220942603216
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,7168,7168,0.030489779180950586
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,7168,65536,0.16466222869025335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,7168,8192,0.048058665461010404
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,7168,8192,0.03185155656602647
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,7168,7168,0.04697155621316698
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,7168,7168,0.028304000695546467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,8192,65536,0.2628266546461317
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,7168,6144,0.028939555088678997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,7168,6144,0.024648000796635944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,7168,6144,0.044496887260013156
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,7168,5120,0.02639111048645443
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,7168,5120,0.021536888347731695
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,7168,4096,0.023223110371165808
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,7168,4096,0.03728266557057699
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,7168,3584,0.022260444031821355
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,7168,5120,0.04189866781234741
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,7168,4096,0.018152889278199937
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,7168,3584,0.03782577647103204
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,7168,3072,0.02037333283159468
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,7168,3584,0.016156444946924847
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,7168,2560,0.019091554813914828
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,7168,3072,0.03618755605485704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,7168,3072,0.014282666974597506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,7168,2560,0.03466044531928168
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,7168,2048,0.010499555203649731
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,7168,2048,0.01826755536927117
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,7168,1536,0.008851555486520132
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,7168,2560,0.012317332956526013
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,7168,1536,0.016760889026853774
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,7168,2048,0.033011555671691895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,7168,1024,0.006924444602595435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,7168,1536,0.031168000565634832
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,7168,1024,0.014260444376203748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,7168,768,0.006120000034570694
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,7168,512,0.0047475554876857335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,7168,1024,0.029037333197063867
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,7168,768,0.014269333746698169
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,7168,65536,0.19891999827490914
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,7168,768,0.026182222697469924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,7168,512,0.012924444344308643
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,7168,256,0.0034799998005231223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,7168,512,0.024329778220918443
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,7168,256,0.012460444536474017
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,7168,128,0.003208000005947219
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,7168,256,0.023756444454193115
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,7168,128,0.012379555238617791
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,7168,64,0.002975111206372579
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,7168,128,0.023240000009536743
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,7168,64,0.012286221815480126
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,7168,32,0.003030222323205736
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,7168,32,0.012219555675983429
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,6144,16384,0.05238666799333361
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,7168,65536,0.2446524567074246
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,6144,16384,0.06702399916119046
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,6144,12288,0.04092622134420607
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,6144,12288,0.04003644320699904
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,6144,16384,0.05336266756057739
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,6144,12288,0.05661866399976942
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,6144,10240,0.03525244527392917
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,6144,10240,0.03765777746836344
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,6144,10240,0.05188622077306112
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,6144,8192,0.03187911046875848
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,6144,7168,0.029294222593307495
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,6144,8192,0.02881422307756212
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,6144,8192,0.04738044407632616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,6144,7168,0.046816888782713145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,6144,6144,0.02780888809098138
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,6144,5120,0.02410488824049632
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,6144,65536,0.15657867325676814
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,6144,6144,0.02272533377011617
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,6144,7168,0.025599110457632277
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,6144,6144,0.042097777128219604
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,6144,4096,0.022083555658658344
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,6144,5120,0.04014933440420363
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,6144,4096,0.0366195539633433
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,6144,5120,0.019692444139056735
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,6144,3584,0.02110311057832506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,6144,3584,0.03712888889842563
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,6144,3072,0.019663110375404358
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,6144,4096,0.016534222496880423
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,6144,2560,0.011428444749779172
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,6144,3072,0.03513066636191474
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,6144,2560,0.018583110637134977
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,6144,3584,0.014533332652515836
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,6144,2048,0.009955555200576782
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,6144,3072,0.013000888956917657
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,6144,2560,0.03432888785998026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,6144,2048,0.017665778597195942
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,6144,1536,0.008275555239783393
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,6144,1024,0.0063999998900625445
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,6144,65536,0.19305866294436982
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,6144,2048,0.032538665665520564
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,6144,1536,0.016377778516875375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,6144,1024,0.013975999421543546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,6144,1536,0.03126666612095303
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,6144,768,0.005718222094906702
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,6144,1024,0.02899555696381463
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,6144,768,0.013335111240545908
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,6144,512,0.0041573333243529005
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,6144,768,0.026312000221676294
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,6144,512,0.013035555680592855
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,6144,256,0.0034595556143257353
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,6144,512,0.024180443750487432
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,6144,256,0.012318222059143914
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,6144,128,0.003156444471743372
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,6144,256,0.023209777143266466
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,6144,128,0.012153777811262341
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,6144,64,0.002941333378354708
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,6144,128,0.023383999864260357
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,6144,32,0.003002666764789157
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,6144,64,0.012239110966523489
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,6144,32,0.012125333150227865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,5120,16384,0.04622044497066074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,5120,16384,0.06658755408393012
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,5120,12288,0.036870221296946205
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,5120,16384,0.04814133379194471
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,5120,12288,0.0567528870370653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,5120,12288,0.037864890363481306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,5120,10240,0.03515733281771342
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,5120,10240,0.05201511250601875
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,5120,8192,0.04666755596796671
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,5120,10240,0.032427555984920926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,5120,8192,0.030887110365761652
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,5120,8192,0.023953777220514085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,5120,65536,0.14038933648003474
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,5120,7168,0.027279999521043565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,5120,6144,0.025684444440735713
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,5120,7168,0.021191999316215515
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,5120,6144,0.041746666034062706
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,5120,7168,0.04584266742070516
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,6144,65536,0.19571110937330458
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,5120,5120,0.023253333237436082
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,5120,6144,0.018770666586028207
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,5120,5120,0.038980444272359215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,5120,4096,0.021154666940371197
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,5120,5120,0.016490666402710807
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,5120,4096,0.03602133194605509
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,5120,65536,0.1724853383170234
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,5120,3584,0.019919999771647982
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,5120,3072,0.010957332948843638
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,5120,4096,0.014154665999942355
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,5120,3584,0.03702133231692844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,5120,3072,0.01908266709910499
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,5120,2560,0.009591110878520543
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,5120,3584,0.012311110893885294
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,5120,2560,0.01795377830664317
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,5120,3072,0.034691555632485285
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,5120,2048,0.008455110920800103
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,5120,2560,0.034076445632510714
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,5120,2048,0.01732711162832048
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,5120,1536,0.007346666521496243
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,5120,2048,0.032095111078686185
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,5120,1536,0.015210666590266757
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,5120,1024,0.006085333310895496
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,5120,1536,0.0310808883772956
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,5120,1024,0.013806222213639153
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,5120,768,0.0052168890833854675
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,5120,1024,0.0262044436401791
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,5120,768,0.013072888884279462
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,5120,512,0.003959111041492886
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,5120,768,0.02643288837538825
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,5120,512,0.01296533313062456
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,5120,256,0.003321777822242843
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,5120,512,0.024687111377716064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,5120,256,0.012266666524940066
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,5120,128,0.003077333379122946
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,5120,128,0.023217777411142986
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,5120,256,0.023437332775857713
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,5120,128,0.012133333418104382
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,5120,64,0.0029279999434947968
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,5120,32,0.002924444360865487
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,5120,64,0.01202755586968528
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,5120,32,0.011986666255527072
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,4096,16384,0.041887111133999295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,4096,16384,0.06519911024305555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,4096,12288,0.03397333290841844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,4096,16384,0.03830044468243917
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,4096,12288,0.030340443054835003
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,4096,12288,0.0520995557308197
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,4096,10240,0.030155556069480047
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,4096,10240,0.02609511050913069
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,4096,10240,0.049810667832692467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,4096,8192,0.044333332114749484
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,5120,65536,0.18531644344329834
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,4096,8192,0.02754577828778161
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,4096,8192,0.02330933345688714
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,4096,7168,0.024193776978386775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,4096,65536,0.11614044507344563
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,4096,7168,0.04427999920315213
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,4096,6144,0.0226248883538776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,4096,6144,0.03956800036960178
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,4096,5120,0.02117777698569828
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,4096,7168,0.020224889119466145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,4096,6144,0.018072888255119324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,4096,5120,0.03769866625467936
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,4096,5120,0.016240888171725802
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,4096,4096,0.01956622302532196
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,4096,3584,0.012831110921170978
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,4096,4096,0.0351839992735121
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,4096,3584,0.019183110859658983
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,4096,65536,0.1653840012020535
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,4096,4096,0.014041778114106921
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,4096,3072,0.011710222396585675
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,4096,3584,0.036213331752353244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,4096,3072,0.017920888132519193
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,4096,3072,0.03369599911901686
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,4096,2560,0.008565333154466417
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,4096,2560,0.01718933383623759
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,4096,2048,0.007618666523032718
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,4096,2048,0.016604445046848722
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,4096,2560,0.03317066695955064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,4096,1536,0.006641777853171031
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,4096,2048,0.03188622328970168
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,4096,1536,0.014746665954589844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,4096,1024,0.005716444303592046
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,4096,1536,0.028352889749738906
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,4096,1024,0.01349422252840466
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,4096,768,0.004691555682155821
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,4096,1024,0.026573333475324843
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,4096,768,0.013020444247457715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,4096,512,0.0037173334923055437
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,4096,768,0.02546399997340308
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,4096,512,0.012709333664841123
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,4096,256,0.0032951111594835916
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,4096,512,0.024337778488794964
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,4096,256,0.012076444096035428
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,4096,128,0.0029813332690132987
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,4096,256,0.023745778534147475
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,4096,128,0.012006222373909421
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,4096,64,0.0028684443483750024
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,4096,128,0.022700443863868713
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,4096,64,0.011785777906576792
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,4096,32,0.0030657777355776895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,4096,32,0.011876444849703046
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,4096,65536,0.13496088981628418
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3584,16384,0.038951999611324735
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3584,12288,0.032720890310075544
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,3584,16384,0.05954044395022922
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3584,16384,0.0348639984925588
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3584,12288,0.0277502222193612
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,3584,12288,0.05331822236378988
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3584,10240,0.029304888513353135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3584,8192,0.026039999392297532
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3584,10240,0.025625778569115534
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,3584,10240,0.04888977938228183
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,3584,8192,0.04394844505521986
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3584,7168,0.023461333579487268
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,3584,7168,0.04264711009131538
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3584,8192,0.0206995556751887
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3584,6144,0.022237333986494277
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3584,7168,0.018801778554916382
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3584,65536,0.10874311129252116
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,3584,6144,0.039064000050226845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3584,5120,0.02051377793153127
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3584,4096,0.013044444223244986
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3584,6144,0.01683200067943997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,3584,5120,0.03828533490498861
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3584,4096,0.019111111760139465
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3584,3584,0.012095111111799875
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,3584,4096,0.03525066706869338
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3584,3584,0.018728888697094388
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3584,5120,0.01479111115137736
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3584,3072,0.011158222125636207
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,3584,3584,0.03513866662979126
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3584,3072,0.01756888959142897
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3584,2560,0.008237332933478886
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,3584,3072,0.03367111086845398
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,3584,65536,0.15503999922010633
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3584,2560,0.017112889223628573
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3584,2048,0.007214222517278459
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,3584,2560,0.03331911232736375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3584,2048,0.015133332875039844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3584,1536,0.006350222147173352
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,3584,2048,0.03179644544919332
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3584,1536,0.014130666851997375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3584,1024,0.005212444398138259
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,3584,1024,0.026400888959566753
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3584,768,0.01298577752378252
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,3584,1536,0.030360000001059637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3584,1024,0.013398222625255585
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3584,768,0.004541333350870344
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3584,512,0.003604444364706675
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,3584,768,0.025224000215530396
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3584,512,0.012644444074895648
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3584,256,0.0032906666811969546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,3584,512,0.024626667300860088
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3584,256,0.012075555821259817
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3584,128,0.002924444360865487
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,3584,256,0.023549333214759827
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3584,128,0.011968889170222811
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3584,64,0.0028088889602157804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,3584,128,0.022962666220135156
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3584,32,0.011678222152921887
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3584,64,0.011794666449228922
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3584,32,0.002895111011134254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3584,65536,0.1224293311436971
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3072,16384,0.03808177842034234
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,3072,16384,0.05844889084498087
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3072,16384,0.03143822153409322
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3072,12288,0.02511288887924618
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3072,12288,0.03130755490726895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,3072,12288,0.05009599857860141
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3072,10240,0.028731554746627808
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,3072,10240,0.04785511228773329
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3072,8192,0.024549333585633173
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3072,10240,0.02224622170130412
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3072,7168,0.017267554998397827
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,3072,8192,0.04260711206330193
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3072,7168,0.022991110881169636
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3072,8192,0.019056888090239633
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,3072,7168,0.041859553919898145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3072,6144,0.02180800007449256
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3072,5120,0.013703111145231458
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,3072,6144,0.038952890369627215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3072,5120,0.020267556111017864
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3072,4096,0.012330667012267642
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3072,4096,0.018816888332366943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,3072,5120,0.03717244333691067
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3072,3584,0.011490667031870948
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3072,65536,0.11363466580708821
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3072,6144,0.01566488875283135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,3072,4096,0.03465777635574341
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,3072,65536,0.145468446943495
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3072,3584,0.018087110585636564
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3072,3072,0.010688888529936472
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,3072,3584,0.034754667017194964
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3072,3072,0.01757422255145179
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3072,2560,0.007872888611422645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,3072,3072,0.03324177861213684
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3072,2560,0.016545777519543965
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3072,2048,0.00700711127784517
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,3072,2560,0.03287022312482198
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3072,1536,0.014277334014574686
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3072,2048,0.015441776977645027
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3072,1536,0.006248000181383557
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,3072,2048,0.03067200051413642
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3072,1024,0.004555555681387584
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,3072,1536,0.027518222729365032
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3072,1024,0.013367111484209696
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3072,768,0.0041893331540955436
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,3072,1024,0.026541334059503343
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3072,768,0.01292711082432005
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3072,512,0.0036231109665499795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,3072,768,0.025676444172859192
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3072,512,0.012595555848545499
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3072,256,0.003232888877391815
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,3072,512,0.024542222420374554
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3072,256,0.012063110868136087
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3072,128,0.002926222152180142
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,3072,256,0.02279111080699497
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3072,65536,0.10172177685631646
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3072,128,0.011902222202883827
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3072,64,0.002761777697337998
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,3072,128,0.022837332553333704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3072,64,0.011728888584507836
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,3072,32,0.0028640000770489373
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,3072,32,0.011746666497654386
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2560,12288,0.03054844339688619
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2560,16384,0.03544355432192484
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,2560,12288,0.04786666565471225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,2560,16384,0.055071999629338585
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2560,16384,0.02817866537306044
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2560,10240,0.027477333943049114
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2560,12288,0.021711111068725586
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,2560,10240,0.04567289021280077
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2560,8192,0.024205333656734888
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2560,10240,0.019118222925398085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,2560,8192,0.0417191121313307
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2560,7168,0.022655111220147874
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,2560,7168,0.03917600048912896
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2560,8192,0.016382222374280293
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2560,6144,0.013396444420019785
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2560,6144,0.021817778547604878
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2560,7168,0.014932443698247274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2560,5120,0.01216533366176817
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2560,5120,0.01982311076588101
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,2560,6144,0.03818133473396301
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2560,65536,0.1083279980553521
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2560,4096,0.01094133324093289
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,2560,5120,0.036935110886891685
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2560,4096,0.01868799991077847
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2560,3584,0.010227555202113258
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,2560,4096,0.03444800111982558
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2560,3584,0.018132444885041978
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2560,3072,0.009846222069528367
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,2560,3584,0.034471998612085976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2560,3072,0.016851555969980028
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2560,2560,0.007616000043021307
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,2560,3072,0.032491554816563926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2560,2560,0.01533600025706821
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2560,2048,0.006828444285525217
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,2560,2560,0.03234577841228909
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2560,2048,0.014903111590279473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2560,1536,0.005931555396980709
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,2560,2048,0.030926220946841772
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2560,1536,0.014025777578353882
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2560,1024,0.004412444515360726
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,2560,1536,0.027878221538331773
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,2560,65536,0.1402791076236301
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2560,1024,0.013457777599493662
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2560,768,0.003923555629120933
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,2560,1024,0.026424888107511733
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2560,768,0.012930666406949362
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2560,512,0.003640888879696528
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,2560,768,0.025576000412305195
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2560,512,0.012503999802801343
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2560,256,0.0032337777730491427
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,2560,512,0.02440888848569658
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2560,256,0.012127999630239276
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2560,128,0.0029102222373088202
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,2560,256,0.022886221607526142
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2560,128,0.011872000164455838
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2560,64,0.0027884443600972495
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,2560,128,0.02291822267903222
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2560,32,0.0028524444335036804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2560,64,0.011795555551846823
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2560,65536,0.09720622168646918
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2560,32,0.011619555453459421
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,2048,16384,0.05925866630342272
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2048,16384,0.03492000036769443
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2048,12288,0.028591109646691218
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2048,16384,0.02233955595228407
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2048,12288,0.018075555562973022
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2048,10240,0.015992000699043274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,2048,12288,0.04825777808825175
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2048,10240,0.02641777859793769
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2048,8192,0.014163555370436775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,2048,10240,0.04628711276584201
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2048,8192,0.02323466704951392
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2048,7168,0.013286222186353473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,2048,8192,0.0411155554983351
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2048,7168,0.02198666665289137
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2048,6144,0.012520888613329994
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2048,6144,0.02088800072669983
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,2048,7168,0.04129599862628513
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2048,65536,0.10018755329979791
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2048,5120,0.011286222272449069
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,2048,6144,0.03918222255176968
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2048,5120,0.018943111101786297
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,2048,5120,0.03654933306905959
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2048,4096,0.00997688869635264
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2048,4096,0.018107554978794523
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2048,3584,0.009135999613338048
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,2048,4096,0.034416887495252825
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2048,3584,0.016685333516862657
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2048,3072,0.008837333156002892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,2048,3584,0.034203555848863386
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2048,3072,0.015627556376987033
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2048,2560,0.00721955547730128
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,2048,3072,0.03286755416128371
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2048,2560,0.014722666806644864
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2048,2048,0.00654666663871871
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,2048,65536,0.1513484451505873
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,2048,2560,0.030336889955732558
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2048,2048,0.014621333943472968
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2048,1536,0.005410666680998272
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,2048,2048,0.029471112622155085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2048,1536,0.01388088862101237
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2048,1024,0.0042568888101312844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,2048,1536,0.02770400047302246
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2048,1024,0.01344622257683012
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2048,768,0.0038444445365005066
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2048,65536,0.07030666536755033
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,2048,1024,0.026308443811204698
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2048,768,0.01291911138428582
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2048,512,0.0034266666819651923
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,2048,768,0.025263999899228413
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2048,512,0.012544000314341651
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2048,256,0.003118222165438864
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,2048,512,0.023796444137891132
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2048,256,0.012138666378127204
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2048,128,0.002867555452717675
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,2048,256,0.023568888505299885
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2048,128,0.011794666449228922
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2048,64,0.0027840000887711844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,2048,128,0.022829333941141765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,2048,32,0.0029306666304667792
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2048,64,0.011778666741318174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,2048,32,0.011547555526097616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1536,16384,0.018904889623324077
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1536,16384,0.03052799900372823
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1536,12288,0.02758399976624383
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,1536,16384,0.05502400133344862
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1536,10240,0.013587555951542325
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,1536,12288,0.04849955439567566
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1536,10240,0.024922667278183833
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1536,12288,0.015473778049151102
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1536,8192,0.012057777908113269
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1536,8192,0.022447110878096685
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,1536,10240,0.044638223118252225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1536,7168,0.011504888534545898
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,1536,8192,0.040058665805392794
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1536,7168,0.020639111598332722
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1536,6144,0.010392888552612728
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,1536,7168,0.04022044605678982
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1536,6144,0.020226667324701946
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1536,5120,0.00980177770058314
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1536,65536,0.09623555342356364
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,1536,6144,0.03798222210672166
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1536,5120,0.01830311119556427
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1536,4096,0.008982222113344405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,1536,5120,0.03617422117127313
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1536,4096,0.016686222619480558
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1536,3584,0.008767999708652496
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,1536,4096,0.033527112669414945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1536,3584,0.016195555528004963
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1536,3072,0.008401778009202745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,1536,3584,0.03165333469708761
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1536,3072,0.015211555692884656
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1536,2560,0.007063111497296228
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,1536,65536,0.13848533895280626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,1536,3072,0.03022577696376377
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1536,2560,0.01459822224246131
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1536,2048,0.005937777873542573
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,1536,2560,0.029562665356530085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1536,2048,0.01447111037042406
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1536,1536,0.0052266667286554975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,1536,2048,0.028756442997190688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1536,1536,0.013711111413107978
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1536,1024,0.004171555654870139
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,1536,1536,0.027248000105222065
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1536,1024,0.01331733332739936
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1536,768,0.0037964445849259696
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,1536,1024,0.0258906665775511
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1536,768,0.012727999852763282
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1536,512,0.0034506666577524612
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,1536,768,0.024824889169798955
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1536,512,0.01257333325015174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1536,65536,0.05674844317966037
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1536,256,0.003087999920050303
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,1536,512,0.024384000235133704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1536,256,0.011954666839705573
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1536,128,0.002855111120475663
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,1536,256,0.022652443912294176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1536,128,0.011879111329714457
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1536,64,0.0027253333893087176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,1536,128,0.022836445106400385
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1536,64,0.011607111328177981
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1536,32,0.0027955555253558685
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1536,32,0.011557333171367645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1024,16384,0.024255111813545227
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1024,12288,0.012530666258600024
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,1024,16384,0.046068443192376025
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1024,12288,0.02155911094612545
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1024,10240,0.011527111132939657
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1024,16384,0.014897776974572076
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1024,10240,0.020290666156344943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,1024,12288,0.0410186681482527
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1024,8192,0.010291555689440833
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,1024,10240,0.03973689013057285
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1024,8192,0.01889600025282966
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1024,7168,0.009520889156394536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,1024,8192,0.036779556009504534
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1024,65536,0.060439997249179415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1024,7168,0.01809155609872606
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1024,6144,0.00868088917599784
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1024,6144,0.016708445217874315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,1024,7168,0.03678133421474033
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1024,5120,0.00868266655339135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,1024,6144,0.035156445370780096
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1024,5120,0.015619556109110514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,1024,65536,0.11504088507758246
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1024,4096,0.008124444219801161
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,1024,5120,0.031898667414983116
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1024,4096,0.01516177753607432
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1024,3584,0.008173333273993598
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,1024,4096,0.02920711040496826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1024,3584,0.014888889259762235
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1024,3072,0.007174222005738153
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,1024,3584,0.030397332376903955
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1024,65536,0.04048266675737169
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1024,3072,0.014577777849303352
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1024,2560,0.006146666490369373
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,1024,3072,0.029529776838090684
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1024,2560,0.014144000079896716
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1024,2048,0.005482666608360078
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,1024,2560,0.02884977724817064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1024,2048,0.013886222408877479
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1024,1536,0.004807999978462855
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,1024,2048,0.0281368891398112
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1024,1536,0.013362666798962487
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1024,1024,0.004063999901215236
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,1024,1536,0.026946667167875502
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1024,1024,0.01292799992693795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,1024,768,0.024506666594081458
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1024,768,0.003730666720204883
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,1024,1024,0.025340444511837427
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1024,768,0.012662221988042196
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1024,512,0.003407111184464561
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1024,512,0.012288000020715924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1024,256,0.003075555587808291
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,1024,512,0.023435556226306494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1024,256,0.011931555966536203
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1024,128,0.0027911110470692315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,1024,256,0.02330933345688714
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1024,128,0.011734222372372946
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1024,64,0.002720000015364753
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,1024,128,0.02237066626548767
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1024,64,0.011539555258221097
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,1024,32,0.0027226667023367355
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,1024,32,0.011404444773991903
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,768,16384,0.0119991112086508
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,768,16384,0.024029332730505202
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,768,12288,0.010468444062603844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,768,12288,0.021154666940371197
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,768,16384,0.045124444696638316
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,768,10240,0.010037333601050908
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,768,12288,0.04080444574356079
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,768,10240,0.01975288821591271
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,768,65536,0.05981689029269748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,768,8192,0.009154666629102495
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,768,10240,0.03934577769703335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,768,8192,0.018791110979186166
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,768,7168,0.008708444734414419
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,768,7168,0.01683200067943997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,768,8192,0.03495644529660543
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,768,6144,0.00814488861295912
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,768,7168,0.03625333309173584
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,768,6144,0.016189333465364244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,768,5120,0.007995555798212687
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,768,65536,0.11085777812533909
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,768,6144,0.034458667039871216
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,768,65536,0.033640000555250384
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,768,5120,0.015584000282817416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,768,4096,0.007152000235186682
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,768,5120,0.031088001198238794
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,768,4096,0.014950222439236112
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,768,3584,0.007530666887760162
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,768,4096,0.029226667351192896
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,768,3584,0.014836443795098199
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,768,3072,0.0068604445291890045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,768,3584,0.030127998855378892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,768,3072,0.01440000037352244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,768,2560,0.006385777973466449
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,768,3072,0.02960977620548672
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,768,2048,0.013727111121018728
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,768,2560,0.014190221826235453
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,768,2048,0.00546577779783143
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,768,2560,0.02866311205757989
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,768,1536,0.0047439999050564235
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,768,2048,0.027820444769329492
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,768,1536,0.013353777428468069
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,768,1024,0.004096888833575779
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,768,1536,0.026628444592158
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,768,1024,0.012879999975363413
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,768,768,0.003690666622585721
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,768,1024,0.02572888963752323
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,768,768,0.012663111090660095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,768,512,0.0034186666210492453
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,768,768,0.024587555064095393
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,768,512,0.012303111453851065
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,768,256,0.003032888803217146
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,768,256,0.011807111402352652
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,768,512,0.023753777146339417
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,768,128,0.0028204443968004654
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,768,256,0.02329866588115692
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,768,128,0.011785777906576792
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,768,64,0.0026453334010309647
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,768,128,0.02219644520017836
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,768,32,0.0026906666656335196
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,768,64,0.011487111449241638
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,768,32,0.01129155523247189
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,512,16384,0.010419555836253695
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,512,16384,0.023531556129455566
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,512,12288,0.009219555391205682
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,512,16384,0.04509688748253716
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,512,10240,0.008746667040718926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,512,12288,0.020958221620983545
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,512,65536,0.05736800034840902
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,512,12288,0.040411555104785495
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,512,10240,0.018518222702874076
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,512,8192,0.008365333080291748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,512,10240,0.03858044412400987
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,512,8192,0.01755466726091173
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,512,7168,0.007981333467695449
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,512,65536,0.024933333198229473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,512,7168,0.034407999780442976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,512,65536,0.11069511042700873
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,512,7168,0.016549333930015564
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,512,8192,0.0343413319852617
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,512,6144,0.0073831114504072405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,512,6144,0.015807999504937064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,512,5120,0.00740444411834081
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,512,6144,0.03275110986497667
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,512,5120,0.015231110983424716
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,512,4096,0.006628444625271692
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,512,5120,0.031104889180925157
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,512,4096,0.014724445011880664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,512,3584,0.007045333584149678
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,512,4096,0.02917511264483134
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,512,3584,0.014555555250909595
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,512,3072,0.006707555717892117
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,512,3584,0.03003822101487054
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,512,3072,0.014130666851997375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,512,2560,0.006114666660626729
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,512,3072,0.02907555633121067
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,512,2560,0.013772444592581855
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,512,2048,0.005411555369695027
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,512,2560,0.02826666831970215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,512,2048,0.013431111143694984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,512,1536,0.004767111192146937
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,512,2048,0.02787644333309597
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,512,1536,0.013217777841620974
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,512,1024,0.00407288885778851
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,512,1536,0.026867555247412786
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,512,1024,0.01291911138428582
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,512,768,0.0037253333462609183
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,512,1024,0.025013332565625507
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,512,768,0.01240355521440506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,512,512,0.0033582221302721235
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,512,768,0.024792000651359558
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,512,512,0.012148444851239523
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,512,256,0.003031999907559819
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,512,512,0.02327644493844774
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,512,256,0.011769777370823754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,512,128,0.0028408887899584244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,512,256,0.022658665974934895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,512,128,0.011702222128709158
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,512,64,0.0027093332674768236
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,512,32,0.0026648888985315957
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,512,128,0.022848889231681824
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,512,64,0.011335111326641507
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,512,32,0.011228444675604502
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,256,16384,0.008108444511890411
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,256,65536,0.05702133311165703
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,256,16384,0.022375999225510493
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,256,12288,0.007253333098358578
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,256,65536,0.1097724437713623
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,256,65536,0.016973333226309884
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,256,16384,0.04384977618853251
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,256,12288,0.019740444090631273
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,256,10240,0.0074444446298811175
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,256,12288,0.03846222162246704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,256,10240,0.018260444204012554
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,256,8192,0.00692622239391009
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,256,10240,0.03703022334310744
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,256,8192,0.017138666576809354
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,256,7168,0.00674311113026407
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,256,6144,0.015711999601787992
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,256,8192,0.03371466530693902
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,256,6144,0.0064479998416370815
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,256,7168,0.01648622254530589
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,256,7168,0.03360088997417026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,256,5120,0.006835555450783835
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,256,6144,0.03228444523281521
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,256,5120,0.015227556228637695
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,256,4096,0.006507555643717448
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,256,5120,0.030780444542566936
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,256,4096,0.014550222290886773
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,256,3584,0.0069448889957533936
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,256,4096,0.029104888439178467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,256,3584,0.014479110638300577
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,256,3072,0.006585777633719974
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,256,3584,0.03015377786424425
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,256,3072,0.014030221435758801
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,256,2560,0.006071110980378256
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,256,3072,0.029304001066419814
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,256,2560,0.013816888961527081
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,256,2048,0.005384888913896348
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,256,2560,0.028322666883468628
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,256,2048,0.013449778159459433
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,256,1536,0.0047120000753137796
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,256,2048,0.027535999814669292
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,256,1536,0.013068444199032254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,256,1024,0.004040889028045866
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,256,1536,0.02648266653219859
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,256,1024,0.012854222622182635
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,256,768,0.003670222229427762
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,256,1024,0.02555822167131636
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,256,768,0.012464889221721225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,256,512,0.0033440000067154565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,256,768,0.024803555674023096
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,256,512,0.012263110942310758
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,256,256,0.0029644444584846497
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,256,512,0.02334311107794444
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,256,256,0.011730666789743634
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,256,128,0.0027804445061418745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,256,256,0.023102222217453852
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,256,128,0.011550222006109027
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,256,64,0.0026746667507621977
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,256,128,0.022798221972253587
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,256,32,0.0026693333768182327
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,256,64,0.011382222175598145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,256,32,0.011236444115638733
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,128,65536,0.013266666895813413
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,128,65536,0.055974225203196205
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,128,16384,0.006743999818960826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,128,16384,0.022237333986494277
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,128,65536,0.10872177945242988
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,128,12288,0.006504000061088138
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,128,16384,0.04230755567550659
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,128,12288,0.019495111372735765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,128,10240,0.006732444382376141
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,128,12288,0.03818666603830125
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,128,10240,0.01832355558872223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,128,10240,0.03671377897262573
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,128,8192,0.006588444527652528
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,128,8192,0.01686755485004849
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,128,7168,0.006504888749784893
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,128,8192,0.03340889016787211
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,128,7168,0.016390222642156813
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,128,6144,0.0064399999876817065
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,128,7168,0.033555554019080266
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,128,6144,0.015722667177518208
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,128,5120,0.006806222101052602
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,128,6144,0.03240977724393209
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,128,5120,0.015056888262430826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,128,4096,0.006476444502671559
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,128,5120,0.030656887425316706
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,128,4096,0.014684443672498068
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,128,3584,0.0068693334857622785
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,128,4096,0.02918133470747206
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,128,3584,0.01440088947614034
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,128,3072,0.006492444624503453
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,128,3584,0.0299911101659139
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,128,3072,0.014162666267818876
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,128,2560,0.006048000107208888
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,128,3072,0.028746667835447524
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,128,2560,0.013752000199423896
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,128,2048,0.005323555734422471
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,128,2560,0.028344889481862385
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,128,2048,0.013396444420019785
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,128,1536,0.004678222040335338
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,128,2048,0.027680888772010803
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,128,1536,0.013026666310098437
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,128,1024,0.00398577791121271
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,128,1536,0.026342221432262
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,128,1024,0.012687111066447364
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,128,768,0.003638222399685118
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,128,1024,0.025440000825458105
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,128,768,0.012410666379663678
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,128,512,0.023319111929999456
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,128,512,0.003314666656984223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,128,768,0.02444000045458476
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,128,512,0.012095111111799875
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,128,256,0.002945777856641345
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,128,128,0.02273955610063341
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,128,32,0.0026391111314296722
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,128,256,0.011781333221329583
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,128,128,0.0027457777824666765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,128,128,0.011462222370836469
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,32,128,256,0.02310577697224087
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,128,64,0.0026524443593290118
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,128,64,0.011258666714032492
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,128,32,0.011234666738245221
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,64,65536,0.012510221865442065
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,64,65536,0.055032001601325146
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,64,10240,0.018216000662909616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,64,16384,0.006511111226346757
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,64,12288,0.006454222318198945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,64,16384,0.022064000368118286
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,64,12288,0.019524445136388142
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,64,10240,0.006604444649484422
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,64,8192,0.006513777706358168
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,64,8192,0.016833777228991192
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,64,7168,0.0063795554969045855
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,64,6144,0.006211555666393704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,64,7168,0.016334222422705758
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,64,6144,0.015576889117558798
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,64,5120,0.006559111177921295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,64,4096,0.006153777655627992
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,64,5120,0.015103111664454142
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,64,4096,0.014544889330863953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,64,3584,0.006601777755551868
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,64,3072,0.006294222341643439
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,64,3584,0.014150222142537435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,64,3072,0.013903111219406128
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,64,2560,0.006022222340106964
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,64,2560,0.01366666704416275
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,64,2048,0.005331555588377847
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,64,2048,0.01333333303531011
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,64,768,0.012344888514942594
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,64,1536,0.004650666481918759
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,64,1024,0.003949333396222857
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,64,1536,0.012996444271670448
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,64,1024,0.0125946667459276
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,64,768,0.003645333151022593
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,64,512,0.0032515554792351196
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,64,512,0.012061333490742577
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,64,256,0.0028906667398081887
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,64,256,0.011687999798191918
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,64,128,0.0027146666414207886
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,64,128,0.011429333024554782
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,64,64,0.0025902222841978073
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,64,64,0.011232888533009423
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,64,32,0.002570666579736604
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,64,32,0.010996444357766045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,32,65536,0.012149333126015134
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,32,65536,0.05494755506515503
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,32,16384,0.006899555524190267
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,32,16384,0.022107554806603327
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,32,12288,0.006759111252095964
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,32,10240,0.006539555473460092
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,32,12288,0.01958488921324412
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,32,10240,0.01816177699300978
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,32,8192,0.0064533332155810455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,32,8192,0.016872000363137987
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,32,7168,0.006255110932721033
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,32,6144,0.006120000034570694
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,32,7168,0.016269332832760282
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,32,6144,0.015659555792808533
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,32,5120,0.00647288892004225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,32,4096,0.006081777728266186
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,32,5120,0.015087111128701104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,32,4096,0.014539556370841132
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,32,3584,0.006542222367392645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,32,3584,0.01421955558988783
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,32,3072,0.006243555496136348
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,32,3072,0.013912889692518445
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,32,2560,0.005942222144868638
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,32,2560,0.013668444421556262
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,32,2048,0.0053119998839166425
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,32,2048,0.013416888813177744
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,32,1536,0.004616888860861461
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,32,1024,0.003921777837806278
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,32,1536,0.013014222184816996
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,32,768,0.003567111161020067
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,32,1024,0.012527999778588613
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,32,512,0.003248888792263137
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,32,768,0.012324444121784635
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,32,64,0.002558222247494592
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,32,512,0.011957333319716983
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,32,256,0.0028782222006056043
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,32,256,0.011563555234008364
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,32,128,0.0027351110345787476
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,32,128,0.011428444749779172
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,32,64,0.01111288865407308
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,32,32,32,0.0025253333151340485
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,32,32,32,0.011014222270912595
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,65536,10240,0.16972088813781738
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,65536,12288,0.19848444726732042
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,65536,12288,0.24078665839301217
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,65536,16384,0.26125600602891713
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,65536,10240,0.20593067010243735
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,65536,12288,0.39328977796766496
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,65536,8192,0.140883551703559
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,65536,16384,0.29390043682522243
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,65536,10240,0.3285306559668647
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,65536,8192,0.1690746678246392
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,65536,8192,0.26383201281229657
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,65536,7168,0.13508444362216526
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,65536,16384,0.523154682583279
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,65536,7168,0.15392533938090006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,65536,6144,0.11702577273050944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,65536,6144,0.13619555367363825
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,65536,5120,0.09455733166800605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,65536,7168,0.2307973437839084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,65536,5120,0.1669680012596978
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,65536,5120,0.12019733587900798
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,65536,4096,0.07926844225989448
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,65536,4096,0.10161599848005508
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,65536,6144,0.19830311669243705
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,65536,4096,0.13461600409613714
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,65536,3584,0.07685599724451701
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,65536,3072,0.06512533293830024
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,65536,3584,0.11694578329722087
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,65536,3584,0.09343733390172322
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,65536,3072,0.08224533663855658
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,65536,2560,0.06012800004747179
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,65536,3072,0.11573599444495307
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,65536,2048,0.049511998891830444
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,65536,2560,0.07230755355623034
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,65536,2048,0.06375377708011203
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,65536,1536,0.04273955689536201
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,65536,1536,0.05285511083073086
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,65536,1024,0.036791112687852644
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,65536,1536,0.05442755421002706
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,65536,2560,0.0961466696527269
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,65536,1024,0.045644445551766284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,65536,768,0.03400889039039612
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,65536,1024,0.037131554550594754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,65536,256,0.027109333210521277
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,65536,768,0.02833777666091919
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,65536,768,0.04109777675734626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,65536,512,0.02014222244421641
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,65536,128,0.00757155567407608
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,65536,256,0.01165066659450531
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,65536,256,0.03220266766018338
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,65536,128,0.025331555141343012
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,65536,512,0.03587999939918518
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,65536,64,0.005668444352017508
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,65536,32,0.004568888909286923
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,65536,2048,0.06931733422809176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,65536,64,0.022836445106400385
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,65536,128,0.028817776176664565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,65536,512,0.031158222092522517
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,65536,32,0.02312622302108341
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,16384,12288,0.0628568861219618
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,16384,16384,0.07856711414125231
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,16384,16384,0.10364533132976955
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,16384,12288,0.0847288899951511
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,16384,12288,0.094468441274431
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,16384,10240,0.05539377861552768
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,16384,16384,0.12403555711110432
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,16384,10240,0.07964266671074761
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,16384,10240,0.07286044624116686
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,16384,8192,0.06433333290947808
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,16384,8192,0.04762755499945747
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,16384,8192,0.06661778026156955
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,16384,7168,0.05697155661053128
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,16384,7168,0.04376177655325996
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,16384,7168,0.060538669427235924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,16384,6144,0.049555556641684644
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,16384,6144,0.039613333013322614
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,16384,5120,0.035981333918041654
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,16384,6144,0.058202664057413735
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,16384,65536,0.26631556616889107
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,16384,4096,0.03195110956827799
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,16384,5120,0.05305777655707466
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,16384,5120,0.04212622178925408
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,16384,4096,0.04753155509630839
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,16384,4096,0.03478488988346524
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,16384,3584,0.0307262208726671
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,16384,3584,0.03385422295994229
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,16384,3072,0.028479999966091577
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,16384,3584,0.04535999894142151
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,16384,2560,0.026403556267420452
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,16384,2048,0.02366311185889774
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,16384,3072,0.026844445202085707
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,16384,2048,0.036567999256981745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,16384,2560,0.04025600022739834
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,16384,3072,0.04272622201177809
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,16384,2560,0.023205333285861548
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,16384,65536,0.31424445576137966
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,16384,2048,0.019278221660190158
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,16384,1536,0.02141511109140184
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,16384,1024,0.019324445062213473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,16384,1536,0.033244444264305964
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,16384,768,0.00905866672595342
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,16384,768,0.018312000566058688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,16384,512,0.0070773329999711775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,16384,256,0.005072888814740711
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,16384,768,0.029304001066419814
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,16384,512,0.017238222890430026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,16384,1024,0.01109777804878023
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,16384,512,0.02753777801990509
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,16384,1536,0.015346666177113851
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,16384,256,0.01366488883892695
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,16384,128,0.003527111063400904
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,16384,256,0.024166221419970196
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,16384,128,0.013402666482660504
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,16384,1024,0.03128622306717767
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,16384,64,0.003253333270549774
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,16384,128,0.02316177719169193
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,16384,32,0.0033093332830402586
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,16384,64,0.013428444663683573
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,16384,32,0.01405422223938836
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,12288,16384,0.06458666589525011
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,12288,16384,0.09392711189058091
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,12288,12288,0.0794008904033237
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,12288,12288,0.052240888277689614
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,12288,16384,0.10443022516038682
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,12288,12288,0.07653155591752794
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,12288,10240,0.04631111025810242
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,12288,10240,0.06714844703674316
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,12288,10240,0.07008800241682264
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,12288,8192,0.04050844576623704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,12288,8192,0.0535093347231547
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,16384,65536,0.5318329069349501
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,12288,8192,0.061062223381466336
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,12288,7168,0.03705777724583944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,12288,7168,0.05243911014662849
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,12288,7168,0.058074666394127734
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,12288,6144,0.036334223217434354
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,12288,6144,0.05372711022694906
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,12288,6144,0.041456000672446355
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,12288,65536,0.21009155114491782
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,12288,5120,0.031114667654037476
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,12288,5120,0.04983911249372694
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,12288,5120,0.03501244386037191
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,12288,4096,0.029763556189007227
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,12288,4096,0.04601422283384535
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,12288,3584,0.027831110689375136
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,12288,3584,0.043284446001052856
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,12288,4096,0.02863466739654541
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,12288,3584,0.025215999947653875
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,12288,3072,0.025041777226659987
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,12288,2560,0.023193778263198003
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,12288,2560,0.037851555479897395
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,12288,2048,0.021809778279728357
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,12288,3072,0.04038933250639174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,12288,3072,0.02234755622016059
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,12288,2048,0.03467466764979892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,12288,2560,0.01924800044960446
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,12288,1536,0.020247111717859905
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,12288,1024,0.009353777600659264
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,12288,1536,0.03225422236654494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,12288,2048,0.016082665986484952
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,12288,1024,0.018731556004948087
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,12288,768,0.007840889195601145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,12288,768,0.017672888106769986
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,12288,1536,0.012823999755912356
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,12288,1024,0.029806223180558946
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,12288,512,0.006298666612969504
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,12288,768,0.028784000211291846
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,12288,512,0.01442577772670322
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,12288,256,0.004310222135649787
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,12288,512,0.02541333271397485
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,12288,256,0.013447110851605734
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,12288,128,0.003400000019205941
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,12288,256,0.023971555961502924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,12288,128,0.013255111045307584
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,12288,64,0.0030817778574095834
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,12288,128,0.02351466649108463
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,12288,32,0.003109333415826162
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,12288,64,0.01308444473478529
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,12288,32,0.013723555538389417
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,12288,65536,0.2856302261352539
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,10240,16384,0.058014220661587186
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,10240,12288,0.04705866509013706
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,10240,16384,0.09190666675567627
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,10240,16384,0.08093777630064222
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,10240,12288,0.06860711177190144
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,10240,12288,0.06735733482572767
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,10240,10240,0.04267377654711405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,10240,10240,0.058303111129336886
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,10240,10240,0.06227466795179579
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,10240,8192,0.03766399953100417
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,10240,8192,0.04136000076929728
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,10240,8192,0.055267555846108325
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,10240,65536,0.18552533785502115
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,10240,7168,0.03459377752410041
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,10240,7168,0.038464887274636164
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,10240,6144,0.03206488821241591
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,10240,7168,0.05239911211861504
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,10240,6144,0.03372711274358962
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,10240,5120,0.029727111260096233
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,10240,6144,0.048420445786582104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,10240,65536,0.23171022203233507
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,10240,5120,0.028622223271263972
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,10240,5120,0.04538311229811775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,10240,3584,0.025219556358125474
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,10240,4096,0.02698755595419142
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,10240,4096,0.023802666200531855
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,12288,65536,0.3982275591956244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,10240,3584,0.03972266780005561
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,10240,3072,0.025007999605602686
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,10240,4096,0.041604446040259466
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,10240,3584,0.021141333712471858
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,10240,2560,0.022426666484938726
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,10240,3072,0.037216001086764865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,10240,2560,0.035565333233939275
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,10240,2048,0.021758221917682227
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,10240,3072,0.018616000811258953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,10240,1536,0.010630221830474006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,10240,2048,0.03343377841843499
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,10240,1536,0.01956622302532196
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,10240,1024,0.008255999949243333
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,10240,1536,0.03169066707293192
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,10240,768,0.007054222126801808
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,10240,1024,0.017972444494565327
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,10240,768,0.016295111841625638
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,10240,1024,0.02945244312286377
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,10240,2048,0.013376000026861826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,10240,512,0.005807111246718301
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,10240,256,0.0037813331517908308
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,10240,2560,0.016132444143295288
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,10240,768,0.028676443629794653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,10240,512,0.014144889182514615
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,10240,512,0.024501333634058636
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,10240,128,0.003253333270549774
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,10240,256,0.013396444420019785
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,10240,256,0.023949333363109167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,10240,128,0.013159111142158508
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,10240,64,0.0029928889125585556
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,10240,128,0.023541332946883306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,10240,32,0.0030097777230872046
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,10240,64,0.013072888884279462
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,10240,32,0.013614222407341003
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,8192,16384,0.04814222123887804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,8192,16384,0.07183288865619235
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,8192,12288,0.039444443252351545
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,8192,16384,0.06743999984529284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,8192,12288,0.052520000272327
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,10240,65536,0.372767104042901
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,8192,12288,0.06323289208941989
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,8192,10240,0.0355404449833764
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,8192,8192,0.03159733282195197
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,8192,10240,0.057054221630096436
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,8192,10240,0.04188977678616842
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,8192,8192,0.03847555650605096
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,8192,8192,0.050513777467939586
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,8192,7168,0.029305779271655615
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,8192,7168,0.04833244283994039
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,8192,6144,0.02716711163520813
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,8192,7168,0.030610667334662542
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,8192,6144,0.026735110415352717
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,8192,5120,0.042973332934909396
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,8192,6144,0.04678844412167867
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,8192,5120,0.02512888941499922
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,8192,4096,0.02204711072974735
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,8192,5120,0.023211555348502264
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,8192,4096,0.01945955554644267
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,8192,3584,0.02088622252146403
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,8192,4096,0.03937600056330363
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,8192,3584,0.01736000014675988
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,8192,3072,0.01995644470055898
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,8192,3584,0.03822933302985297
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,8192,65536,0.14434666103786892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,8192,2560,0.01872977779971229
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,8192,65536,0.2043022182252672
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,8192,3072,0.03571733170085483
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,8192,3072,0.015319999721315173
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,8192,2048,0.017814222309324477
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,8192,1536,0.009174221919642555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,8192,2560,0.035022222333484225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,8192,2560,0.013345777988433838
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,8192,2048,0.03212711215019226
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,8192,1536,0.01664355562792884
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,8192,2048,0.011166222393512726
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,8192,1024,0.007144889069928064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,8192,1536,0.030751112434599135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,8192,1024,0.01464000013139513
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,8192,768,0.006219555520349079
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,8192,1024,0.028239998552534316
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,8192,768,0.013215999636385175
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,8192,512,0.005223999834722943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,8192,768,0.02548622257179684
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,8192,256,0.00351200004418691
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,8192,512,0.0124195557501581
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,8192,512,0.02394666605525547
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,8192,256,0.012102222277058495
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,8192,128,0.0032097777972618737
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,8192,256,0.023464888334274292
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,8192,128,0.011717333561844297
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,8192,64,0.003024000053604444
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,8192,32,0.0030817778574095834
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,8192,64,0.011651555697123209
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,8192,128,0.023408000667889912
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,8192,32,0.012285333540704517
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,7168,16384,0.04409866531689962
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,7168,16384,0.07136088609695435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,7168,12288,0.03945688737763299
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,7168,16384,0.0610791113641527
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,7168,12288,0.04768800073199802
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,8192,65536,0.26013954480489093
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,7168,12288,0.06044266621271769
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,7168,10240,0.03321511215633816
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,7168,8192,0.029684444268544514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,7168,10240,0.05500088797675239
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,7168,10240,0.0436480012204912
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,7168,8192,0.03741511040263706
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,7168,8192,0.04890666736496819
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,7168,7168,0.027659555276234944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,7168,7168,0.030722667773564655
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,7168,6144,0.0259253333012263
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,7168,7168,0.046816001335779824
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,7168,65536,0.13140622774759927
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,7168,6144,0.02734222180313534
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,7168,4096,0.020870221985710993
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,7168,5120,0.02388800018363529
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,7168,6144,0.04499199986457825
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,7168,5120,0.041788445578681103
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,7168,4096,0.038198222716649376
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,7168,5120,0.02124888863828447
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,7168,3584,0.02083822256988949
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,7168,3584,0.03694933321740892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,7168,3072,0.01900800069173177
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,7168,4096,0.017955554856194388
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,7168,3584,0.016007110476493835
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,7168,3072,0.014106666048367819
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,7168,3072,0.03507644600338406
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,7168,2560,0.017980444762441847
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,7168,2048,0.010430222584141625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,7168,2560,0.033999111917283796
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,7168,2560,0.012181333369678922
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,7168,2048,0.01737244427204132
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,7168,1536,0.00868977771864997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,7168,2048,0.032050665881898664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,7168,1536,0.016354666815863717
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,7168,1024,0.006888888776302338
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,7168,1536,0.03011288907792833
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,7168,768,0.006056889063782162
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,7168,1024,0.013282666603724161
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,7168,768,0.012630222572220696
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,7168,1024,0.027727999620967444
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,7168,512,0.0047253333032131195
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,7168,768,0.025799110531806946
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,7168,512,0.012402666939629449
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,7168,256,0.003367111086845398
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,7168,512,0.02437422176202138
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,7168,256,0.011976000335481433
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,7168,128,0.003113777687152227
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,7168,256,0.023039110832744177
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,7168,128,0.011710222396585675
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,7168,64,0.0029528888149393927
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,7168,32,0.0122097780307134
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,7168,32,0.0029395555870400537
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,7168,64,0.011786667009194693
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,7168,65536,0.18651199340820312
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,7168,128,0.02267644471592373
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,6144,16384,0.041499555110931396
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,6144,16384,0.06999555561277601
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,7168,65536,0.24242401123046875
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,6144,16384,0.05233066611819797
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,6144,12288,0.034713778230879046
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,6144,10240,0.03175110949410333
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,6144,12288,0.0603368878364563
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,6144,10240,0.05393955442640516
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,6144,10240,0.03423733181423611
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,6144,12288,0.04038222299681769
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,6144,8192,0.02980177932315403
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,6144,8192,0.028498666154013738
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,6144,8192,0.04820444517665439
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,6144,65536,0.12308443917168511
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,6144,7168,0.026783999469545152
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,6144,7168,0.04671288861168755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,6144,7168,0.02494311167134179
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,6144,6144,0.02489600082238515
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,6144,5120,0.0217884447839525
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,6144,6144,0.04348355531692505
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,6144,5120,0.04100622071160211
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,6144,6144,0.02202933364444309
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,6144,4096,0.020239111449983384
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,6144,3584,0.019461333751678467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,6144,4096,0.037851555479897395
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,6144,5120,0.01926488843229082
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,6144,3584,0.036341332727008395
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,6144,4096,0.016224000189039443
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,6144,65536,0.1871617767545912
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,6144,3072,0.018608000543382432
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,6144,2560,0.011211555865075854
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,6144,3584,0.014392000105645923
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,6144,2560,0.01770666738351186
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,6144,3072,0.03519644339879354
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,6144,2048,0.00979377743270662
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,6144,2560,0.033266666862699724
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,6144,2048,0.01696622206105126
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,6144,3072,0.012849777936935425
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,6144,1536,0.008185777399275038
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,6144,1536,0.015992000699043274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,6144,1024,0.006883555402358373
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,6144,2048,0.03165955675972833
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,6144,1024,0.013034666577974955
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,6144,768,0.01257777793539895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,6144,768,0.005623111294375525
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,6144,1536,0.030697776211632624
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,6144,1024,0.02587911155488756
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,6144,512,0.0041617775956789655
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,6144,768,0.025117332736651104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,6144,512,0.012255111502276527
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,6144,256,0.0033235556135574975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,6144,256,0.011906666888131036
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,6144,512,0.023716444770495098
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,6144,128,0.003064888839920362
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,6144,256,0.02310666607485877
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,6144,128,0.011592888997660743
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,6144,64,0.0028577778074476454
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,6144,128,0.023099554909600153
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,6144,32,0.002961777771512667
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,6144,64,0.011679111255539788
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,6144,32,0.012088889049159156
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,5120,16384,0.03730666637420654
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,5120,16384,0.06856622298558553
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,5120,12288,0.03877244393030802
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,5120,16384,0.04963199959860908
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,5120,12288,0.03204977843496535
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,5120,12288,0.05891466803020901
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,5120,10240,0.02865866820017497
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,5120,10240,0.03137333194414774
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,5120,8192,0.02590755621592204
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,5120,10240,0.05289777782228258
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,5120,8192,0.023571555813153584
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,5120,8192,0.046247111426459424
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,5120,65536,0.10655022329754299
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,5120,7168,0.024737776981459722
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,5120,7168,0.04570933183034261
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,5120,6144,0.022796443767017786
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,6144,65536,0.19329688284132215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,5120,7168,0.0207324441936281
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,5120,5120,0.020765332712067496
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,5120,6144,0.04285066657596164
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,5120,6144,0.018522666560279
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,5120,5120,0.04086577892303467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,5120,4096,0.019426667028003268
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,5120,4096,0.0373671121067471
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,5120,3584,0.018557333283954196
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,5120,65536,0.16752444373236763
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,5120,5120,0.016229333149062265
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,5120,4096,0.013598221871587964
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,5120,3584,0.03603377938270569
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,5120,3072,0.010972444381978778
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,5120,3072,0.01793866687350803
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,5120,3584,0.01223111069864697
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,5120,2560,0.009733333355850643
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,5120,3072,0.03419288992881775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,5120,2560,0.017071111334694754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,5120,2048,0.008475555313958062
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,5120,2048,0.016389333539538913
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,5120,2560,0.03299199872546726
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,5120,1536,0.007269333634111617
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,5120,2048,0.03128355410363939
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,5120,1536,0.015064888530307345
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,5120,1024,0.006008000009589725
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,5120,1536,0.02973955538537767
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,5120,1024,0.012990222209029727
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,5120,768,0.005294222384691238
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,5120,1024,0.026494221554862127
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,5120,768,0.012495111260149213
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,5120,512,0.0038373333712418876
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,5120,512,0.012284444438086616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,5120,768,0.024570667081409033
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,5120,256,0.0032871110985676446
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,5120,512,0.023564444647894964
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,5120,256,0.011915555430783166
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,5120,128,0.002974222310715251
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,5120,256,0.023240000009536743
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,5120,64,0.002797333316670524
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,5120,128,0.01163733336660597
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,5120,128,0.023256889647907678
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,5120,64,0.011445333560307821
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,5120,32,0.002842666581273079
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,5120,32,0.01199288914601008
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,5120,65536,0.18639555242326525
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,4096,16384,0.032905777295430504
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,4096,16384,0.06760978036456637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,4096,16384,0.03694933321740892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,4096,12288,0.028104000621371802
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,4096,12288,0.03221155537499322
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,4096,12288,0.05626133415434095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,4096,10240,0.05123022198677063
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,4096,10240,0.025360888904995386
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,4096,8192,0.04453866680463155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,4096,10240,0.027058665951093037
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,4096,8192,0.023629332582155865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,4096,7168,0.021432888176706102
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,4096,8192,0.02250755495495266
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,4096,6144,0.02000711030430264
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,4096,7168,0.044024000565210976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,4096,7168,0.020648888415760465
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,4096,5120,0.0159262220064799
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,4096,6144,0.041238221857282854
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,4096,5120,0.0189057770702574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,4096,6144,0.017874666386180453
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,4096,5120,0.03897510965665182
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,4096,4096,0.013697778185208639
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,4096,4096,0.018076444665590923
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,4096,3584,0.012636444634861417
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,4096,3584,0.017515555024147034
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,4096,4096,0.037068443165885076
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,4096,3072,0.011615110768212212
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,4096,3584,0.03561688794030084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,4096,3072,0.01721600029203627
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,4096,2560,0.00851111114025116
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,4096,2560,0.016420443852742512
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,4096,3072,0.033477332856920027
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,4096,65536,0.08665421936247085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,4096,2048,0.007791111038790808
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,4096,2048,0.015847111741701763
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,4096,2560,0.03253066539764404
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,4096,1536,0.00674933319290479
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,4096,65536,0.16613154941134983
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,4096,2048,0.030862222115198772
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,4096,1536,0.013254221942689685
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,4096,1024,0.005657777604129579
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,4096,1536,0.029566221767001685
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,4096,1024,0.012692444854312472
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,4096,768,0.004355555607212914
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,4096,1024,0.025294222765498694
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,4096,768,0.012486221889654795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,4096,512,0.0035955554081334006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,4096,768,0.0243520008193122
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,4096,256,0.011744889120260874
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,4096,512,0.012103111379676394
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,4096,256,0.0033297776761982176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,4096,512,0.023504000571038988
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,4096,128,0.0029440000653266907
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,4096,256,0.02296977738539378
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,4096,128,0.011601777540312873
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,4096,64,0.0027840000887711844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,4096,128,0.02269777821169959
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,4096,32,0.0029191111938820947
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,4096,64,0.011546666423479715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,4096,32,0.01219377749496036
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3584,16384,0.030753778086768255
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,4096,65536,0.1332577731874254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,3584,16384,0.06238489018546211
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3584,12288,0.026328000757429335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3584,16384,0.03405244482888116
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3584,10240,0.024296889702479046
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3584,12288,0.027076444692081873
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,3584,12288,0.05512711074617174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3584,8192,0.021934222843911912
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3584,10240,0.0236791107389662
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,3584,8192,0.043415112627877124
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,3584,10240,0.048692444960276283
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3584,7168,0.020415110720528495
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3584,8192,0.01975200076897939
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,3584,7168,0.042583111259672374
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3584,6144,0.01936533384852939
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,3584,6144,0.04061155517896017
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3584,7168,0.0181013329161538
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3584,6144,0.016193777322769165
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3584,5120,0.018394667241308425
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3584,4096,0.012929778132173749
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,3584,5120,0.03875733415285746
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3584,5120,0.014556444353527494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3584,4096,0.01753777762254079
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,3584,4096,0.03674755493799845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3584,3584,0.011991110940774282
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3584,3584,0.017278222574128043
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3584,3072,0.011005333728260465
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,3584,3584,0.03512888815667894
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3584,3072,0.016508445143699646
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3584,65536,0.07852355639139812
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,3584,3072,0.03410755594571432
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3584,2560,0.008273777862389883
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,3584,65536,0.16393510500590006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3584,2560,0.016319110989570618
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3584,2048,0.00721777809990777
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3584,1536,0.006396444307433234
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,3584,2560,0.03282399972279867
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3584,2048,0.015556444724400839
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,3584,2048,0.03071555495262146
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,3584,1536,0.02971022327740987
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3584,1536,0.013178666432698568
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3584,1024,0.004996444616052839
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3584,1024,0.01274577776590983
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3584,768,0.004317333300908406
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,3584,1024,0.025422222084469263
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3584,768,0.01240888900227017
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3584,512,0.0035706667436493766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,3584,768,0.024189333120981853
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3584,512,0.012144889268610211
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3584,256,0.0031973332580592898
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,3584,512,0.023390221926901076
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,3584,256,0.023129777775870428
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3584,256,0.011755555868148804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3584,128,0.0029173334025674393
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3584,128,0.011508444117175208
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3584,64,0.002741333304180039
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,3584,128,0.02311822275320689
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3584,32,0.0027439999911520216
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3584,64,0.011389333340856763
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3584,32,0.011896000140243106
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3584,65536,0.12439821826087104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3072,16384,0.028992888000276353
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,3072,16384,0.061483555369906955
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3072,12288,0.024496000674035814
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3072,10240,0.022964444425370958
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3072,12288,0.024643555283546448
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,3072,12288,0.05362933211856418
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3072,16384,0.030590222941504583
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3072,8192,0.02068177858988444
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,3072,10240,0.049433777729670204
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,3072,8192,0.04378933376736111
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3072,10240,0.021743999587164983
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3072,7168,0.019751111666361492
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3072,6144,0.015056000815497505
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3072,8192,0.018591110905011494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,3072,7168,0.042239109675089516
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3072,6144,0.018767999278174508
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3072,7168,0.016895999511082966
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3072,5120,0.013472889032628802
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3072,5120,0.017855111095640395
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,3072,6144,0.04043289025624593
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3072,65536,0.07233422332339816
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3072,4096,0.01220355596807268
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,3072,5120,0.039035555389192365
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3072,4096,0.017321777012613084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3072,3584,0.011275555524561139
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,3072,4096,0.036632001399993896
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3072,3584,0.01685599982738495
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3072,3072,0.01035822182893753
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,3072,3584,0.034613334470325045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3072,3072,0.016344888342751395
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3072,2560,0.00774133370982276
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,3072,3072,0.03344800074895223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3072,2560,0.015522667103343539
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,3072,65536,0.14713066154056126
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3072,2048,0.00702400008837382
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,3072,2560,0.0322675539387597
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3072,2048,0.015066666735543145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3072,1536,0.006234666539563074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,3072,2048,0.028352889749738906
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3072,1536,0.01313511116637124
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3072,1024,0.0047395556337303585
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,3072,1536,0.02715644405947791
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3072,1024,0.012656889028019376
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3072,768,0.0038986665507157645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,3072,1024,0.02472177810139126
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3072,768,0.012496000362767113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3072,512,0.0035457776652442086
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,3072,768,0.02421155571937561
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3072,512,0.012131555212868584
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3072,256,0.0031866667171319327
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,3072,512,0.02346399923165639
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3072,256,0.011686222420798408
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3072,128,0.002887111157178879
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,3072,256,0.023173333870040044
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3072,128,0.011526222030321756
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3072,64,0.0027022223091787765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,3072,128,0.022983110613293115
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3072,32,0.0027928888383838865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3072,64,0.011383111278216044
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,3072,32,0.011854222251309289
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2560,16384,0.02731288969516754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,2560,16384,0.060164444976382785
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2560,16384,0.027344889110989038
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2560,12288,0.023801777097913954
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,3072,65536,0.10572889116075303
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,2560,12288,0.05154133174154493
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2560,12288,0.02178222272131178
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2560,10240,0.02176355487770504
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,2560,10240,0.04745066828197903
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2560,8192,0.020257777637905545
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2560,65536,0.06780178017086454
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,2560,8192,0.04166133205095927
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2560,8192,0.01612622208065457
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2560,10240,0.018830221560266282
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2560,7168,0.019024888674418133
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2560,6144,0.013222222526868185
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2560,6144,0.018386666973431904
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,2560,7168,0.04064800010787116
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2560,7168,0.014616888430383472
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2560,5120,0.011908444265524546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,2560,6144,0.03970044520166185
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2560,5120,0.017695110705163743
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,2560,65536,0.13295555114746094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2560,4096,0.0107004443804423
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2560,3584,0.010102222363154093
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2560,4096,0.016794666647911072
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,2560,5120,0.0377786656220754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,2560,4096,0.03597955571280585
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2560,3072,0.009536888864305284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2560,3584,0.01666755477587382
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2560,3072,0.014952888091405233
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,2560,3584,0.03365688853793674
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2560,2560,0.00739911115831799
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,2560,3072,0.03331022130118476
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2560,2560,0.015411555767059326
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2560,2048,0.006764444626039929
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,2560,2560,0.03240622083346049
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2560,2048,0.01368799971209632
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2560,1536,0.005929777605666055
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,2560,2048,0.028412444723976985
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2560,1536,0.01308444473478529
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2560,768,0.0038684445122877755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2560,1024,0.004234666625658671
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,2560,1536,0.0266302227973938
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2560,1024,0.012713777522246042
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,2560,1024,0.02477600011560652
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2560,768,0.012225777738624148
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2560,512,0.003528000166018804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,2560,768,0.02415200074513753
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2560,65536,0.09521510865953232
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2560,512,0.012155555188655853
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2560,256,0.003144888828198115
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,2560,512,0.02327022287580702
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2560,256,0.011738667057620155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2560,128,0.0028640000770489373
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,2560,256,0.022894221875402663
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2560,128,0.011501332951916588
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2560,64,0.0026977778308921387
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2560,64,0.01144444445768992
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,2560,128,0.02287111183007558
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2560,32,0.0028008888992998335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2560,32,0.011808888779746162
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2048,12288,0.023018666439586218
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2048,16384,0.026086222794320848
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,2048,16384,0.061602665318383105
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,2048,12288,0.05391822258631388
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2048,12288,0.018039999736679923
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2048,16384,0.02180800007449256
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2048,10240,0.020510221521059673
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2048,8192,0.013916444447305469
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,2048,10240,0.046911110480626426
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2048,8192,0.019485332899623446
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2048,7168,0.012991111311647626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,2048,8192,0.042209777567121715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2048,7168,0.018876444962289598
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2048,6144,0.012072888513406118
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2048,10240,0.016488000750541687
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,2048,7168,0.04153777824507819
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2048,6144,0.01796266602145301
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,2048,6144,0.039488888449139066
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2048,5120,0.011004444625642566
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2048,65536,0.05862577756245931
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2048,5120,0.017335111896197002
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2048,4096,0.009891555541091496
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,2048,5120,0.03841422332657708
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2048,4096,0.016542222764756944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2048,3584,0.009217778013812171
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,2048,4096,0.0355022218492296
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2048,3584,0.0158915552828047
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2048,3072,0.008590222232871586
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,2048,3584,0.033844444486829973
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,2048,65536,0.15162400404612222
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2048,3072,0.015425778097576566
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2048,2560,0.007275555696752336
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,2048,3072,0.0326551099618276
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2048,2560,0.01386844449573093
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2048,2048,0.006435555716355641
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,2048,2560,0.029995554023318823
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2048,2048,0.013413333230548434
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2048,1536,0.005527110977305307
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,2048,2048,0.028951999213960435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2048,1536,0.01309866706530253
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2048,1024,0.00416711096962293
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,2048,1536,0.02678577767478095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2048,1024,0.012701333396964602
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2048,768,0.003773333297835456
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,2048,1024,0.02510133385658264
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2048,768,0.012490666574902005
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2048,512,0.003411555455790626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,2048,768,0.024546666277779475
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2048,512,0.01216533366176817
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2048,256,0.003084444544381566
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2048,65536,0.06824444399939643
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,2048,512,0.023232888844278123
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2048,256,0.011740444435013665
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2048,128,0.0028640000770489373
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,2048,256,0.022758222288555566
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2048,128,0.011477332976129321
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2048,64,0.0026728889594475427
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,2048,128,0.02276800076166789
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2048,64,0.011295110815101199
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,2048,32,0.0026835555003749
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,2048,32,0.011982222398122152
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1536,16384,0.025102222959200542
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,1536,16384,0.05832266807556152
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1536,12288,0.02134311033619775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1536,12288,0.014953777194023132
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1536,10240,0.013269333375824822
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1536,16384,0.01822933389080895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,1536,12288,0.0515324440267351
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1536,10240,0.020082667469978333
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1536,8192,0.011810666984981961
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,1536,10240,0.046591109699673124
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1536,8192,0.018990221950742934
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1536,7168,0.011086222198274402
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,1536,8192,0.04197777642144097
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1536,7168,0.01831288966867659
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1536,65536,0.05940177705552843
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1536,6144,0.010288000106811523
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1536,6144,0.01756266587310367
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1536,5120,0.009482666850090027
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,1536,6144,0.038712888956069946
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,1536,7168,0.0415653321478102
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1536,5120,0.016700444949997794
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1536,4096,0.00888177752494812
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,1536,5120,0.03706577751371596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1536,4096,0.0158906661801868
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1536,3584,0.008760000268618265
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,1536,65536,0.1360551118850708
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1536,3584,0.014464888307783337
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,1536,4096,0.03531644410557217
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1536,3072,0.008230222596062554
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,1536,3584,0.03287555442916022
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1536,3072,0.01402844488620758
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1536,2560,0.0069520001610120135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,1536,3072,0.030604445272021826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1536,2560,0.013752889302041797
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1536,2048,0.00601333338353369
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,1536,2560,0.029696000946892634
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1536,2048,0.01331022216214074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1536,1024,0.0041306668685542215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1536,1536,0.004795555439260271
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,1536,2048,0.02755555510520935
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1536,65536,0.05485599901941088
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1536,1536,0.012963555753231049
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,1536,1536,0.02647022240691715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1536,1024,0.012561777399645912
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1536,768,0.0037475555307335327
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,1536,1024,0.025272889269722834
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1536,768,0.012327999704413943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,1536,512,0.023307555251651343
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1536,512,0.003391111062632667
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,1536,768,0.023864888482623633
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1536,512,0.012042666474978128
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1536,256,0.003085333233078321
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1536,256,0.01167377746767468
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1536,128,0.00282577777074443
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,1536,256,0.02306755549377865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1536,128,0.01145244472556644
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1536,64,0.0026808888134029177
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,1536,128,0.02233422299226125
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1536,32,0.002704888789190186
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1536,64,0.011349333657158745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1536,32,0.011717333561844297
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1024,16384,0.014459555347760519
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1024,16384,0.02253155575858222
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1024,12288,0.012334222594896952
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1024,12288,0.020288000504175823
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,1024,16384,0.04866577850447761
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1024,10240,0.011251555548773872
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1024,10240,0.019448000523779128
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,1024,12288,0.04558666547139486
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1024,8192,0.010026666853162978
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,1024,10240,0.041879110866122775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1024,8192,0.018202667435010273
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,1024,8192,0.03872266742918227
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1024,7168,0.009457777771684859
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1024,7168,0.017148445049921673
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1024,6144,0.008564444051848518
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,1024,7168,0.037759111987219915
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1024,6144,0.016009777784347534
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1024,65536,0.05238133337762621
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1024,5120,0.008626666333940294
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,1024,65536,0.11927821901109482
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,1024,6144,0.03583022289805942
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1024,5120,0.015301333533393012
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1024,4096,0.00794666674402025
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,1024,5120,0.03456799851523505
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1024,4096,0.014852444330851236
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1024,3584,0.008173333273993598
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,1024,4096,0.032272001107533775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1024,3584,0.014674666855070325
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1024,65536,0.03725688987308078
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1024,3072,0.007335999773608313
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,1024,3584,0.03047555685043335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1024,3072,0.014254222313563028
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1024,2560,0.006461333483457565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,1024,3072,0.02884799904293484
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1024,2560,0.013986666997273764
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1024,2048,0.005710222240951326
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,1024,2560,0.02755466600259145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1024,2048,0.013628444737858243
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1024,1536,0.004765333400832282
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1024,1536,0.013206221991115145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,1024,2048,0.02740266587999132
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1024,1024,0.004089777668317159
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,1024,1536,0.025740444660186768
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1024,1024,0.012649777862760754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1024,768,0.00370399985048506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1024,768,0.012348444097571902
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,1024,1024,0.024707555770874023
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1024,512,0.0034151110384199354
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,1024,768,0.024162666665183172
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1024,512,0.012088889049159156
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1024,256,0.0030382221771611106
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,1024,512,0.023563555545277063
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1024,256,0.011689778003427716
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1024,128,0.0028124445428450904
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,1024,256,0.023100444012218054
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1024,32,0.0027013334135214486
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1024,128,0.01144977741771274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,1024,64,0.0026933333526055017
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,1024,128,0.022260444031821355
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1024,64,0.011431111229790581
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,1024,32,0.011669333610269757
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,768,16384,0.01183644433816274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,768,16384,0.022351110974947613
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,768,12288,0.01017955525053872
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,768,12288,0.020019556085268658
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,768,16384,0.044010665681627065
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,768,10240,0.009886222581068674
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,768,12288,0.04230044285456339
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,768,8192,0.00922400007645289
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,768,10240,0.01812177730931176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,768,8192,0.016928889685206942
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,768,65536,0.05065422256787618
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,768,10240,0.039396444956461586
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,768,7168,0.008548444343937768
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,768,7168,0.017454221844673157
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,768,8192,0.03550488750139872
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,768,7168,0.03548444310824076
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,768,6144,0.008124444219801161
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,768,6144,0.015992888145976596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,768,5120,0.007813333637184566
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,768,6144,0.032906668053732976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,768,65536,0.1066017746925354
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,768,5120,0.01606399979856279
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,768,4096,0.00721066693464915
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,768,5120,0.031152000029881794
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,768,4096,0.014783110883500842
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,768,3584,0.007490666376219855
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,768,3584,0.02994489007525974
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,768,4096,0.030617776844236586
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,768,3584,0.014519999424616495
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,768,3072,0.006884444504976273
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,768,65536,0.031138665146297876
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,768,2560,0.0063733334342638654
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,768,3072,0.01427288850148519
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,768,3072,0.028402666250864666
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,768,2048,0.026874666412671406
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,768,2560,0.013952889376216464
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,768,2560,0.02754488918516371
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,768,2048,0.005421333428886201
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,768,2048,0.01346577786737018
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,768,1536,0.004770666774776247
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,768,1536,0.013091555900043912
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,768,1536,0.02576622201336755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,768,1024,0.004109333372778363
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,768,1024,0.0125608891248703
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,768,768,0.003718222181002299
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,768,1024,0.02400711178779602
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,768,768,0.012288889123333825
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,768,512,0.0033724444607893624
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,768,768,0.023960888385772705
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,768,512,0.012120000190205045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,768,256,0.003018666679660479
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,768,512,0.02332266668478648
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,768,256,0.011692444483439127
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,768,128,0.0028355556229750314
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,768,256,0.022447999980714586
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,768,128,0.011428444749779172
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,768,64,0.0026577777332729767
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,768,128,0.02286044425434536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,768,32,0.002696000039577484
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,768,64,0.011324444578753578
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,768,32,0.011653333074516721
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,512,16384,0.01019377758105596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,512,12288,0.008946666287051307
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,512,16384,0.022079111801253423
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,512,16384,0.04362755682733324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,512,12288,0.018976888722843595
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,512,65536,0.09934310780631171
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,512,65536,0.049215111467573375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,512,10240,0.008627555436558193
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,512,8192,0.008023999631404877
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,512,12288,0.04016622238688999
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,512,10240,0.017695110705163743
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,512,65536,0.023369777533743117
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,512,10240,0.03629422187805176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,512,8192,0.016705777910020616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,512,7168,0.007452444069915348
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,512,8192,0.033153778976864286
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,512,7168,0.016306666864289176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,512,6144,0.006979555719428592
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,512,7168,0.03381777803103129
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,512,6144,0.01568711135122511
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,512,5120,0.00721599989467197
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,512,6144,0.032115555471844144
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,512,5120,0.015207999282413058
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,512,4096,0.006542222367392645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,512,5120,0.030744890371958416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,512,4096,0.014898666077189975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,512,3584,0.006979555719428592
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,512,4096,0.029455112086402044
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,512,3584,0.01459022197458479
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,512,3072,0.006616000086069107
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,512,3584,0.029038220643997192
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,512,3072,0.014069333672523499
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,512,2560,0.006080889039569431
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,512,3072,0.02806666824552748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,512,2560,0.013861333330472311
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,512,2048,0.005407111098368962
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,512,2560,0.027772444817754958
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,512,2048,0.013365333278973898
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,512,1536,0.004751111070315043
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,512,2048,0.026476444469557867
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,512,1536,0.01313333296113544
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,512,1024,0.004042666819360522
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,512,1536,0.025255999631351892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,512,768,0.012321777641773224
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,512,1024,0.012675555215941535
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,512,768,0.003680888977315691
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,512,1024,0.024534222152498033
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,512,512,0.0033777778347333274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,512,768,0.023796444137891132
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,512,256,0.003031999907559819
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,512,512,0.012103111379676394
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,512,512,0.022766222556432087
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,512,256,0.011770666473441653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,512,128,0.002833777831660377
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,512,256,0.02233422299226125
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,512,128,0.011439111497667102
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,512,64,0.0027093332674768236
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,512,128,0.02276266614596049
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,512,32,0.0026382222357723448
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,512,64,0.011250666446155973
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,512,32,0.011876444849703046
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,256,16384,0.008164444731341468
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,256,65536,0.049103998475604586
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,256,16384,0.02146311104297638
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,256,65536,0.09843466679255168
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,256,12288,0.006996444529957241
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,256,16384,0.04168444540765551
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,256,12288,0.018671999375025432
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,256,65536,0.0147706667582194
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,256,10240,0.007060444189442529
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,256,12288,0.03813066747453477
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,256,10240,0.017640888690948486
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,256,8192,0.006664000037643645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,256,10240,0.03564711080657111
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,256,8192,0.016582222448454965
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,256,7168,0.006684444430801604
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,256,8192,0.03299377693070306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,256,7168,0.01613333324591319
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,256,6144,0.006488889041874144
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,256,7168,0.03330399923854404
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,256,6144,0.015575110912322998
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,256,5120,0.006863111009200414
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,256,6144,0.032467554012934365
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,256,5120,0.015090665883488126
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,256,4096,0.006454222318198945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,256,5120,0.030920889642503526
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,256,4096,0.014680888917711047
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,256,3584,0.0069146665434042616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,256,4096,0.029284444120195176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,256,3584,0.014498665928840637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,256,3072,0.027907556957668726
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,256,3584,0.02905244297451443
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,256,3072,0.006598222172922558
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,256,3072,0.014019555515713163
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,256,2560,0.00602311102880372
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,256,2560,0.013589333328935834
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,256,2048,0.005344888816277186
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,256,2560,0.027043556173642475
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,256,2048,0.013389333254761167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,256,1536,0.0046951112647851305
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,256,2048,0.026552889082166884
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,256,1536,0.012881778180599213
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,256,1024,0.004035555654101902
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,256,1536,0.02516355613867442
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,256,1024,0.012500444220172035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,256,768,0.003622222277853224
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,256,1024,0.023745778534147475
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,256,768,0.012339555554919772
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,256,512,0.003359111025929451
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,256,768,0.023228444986873206
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,256,512,0.01187111106183794
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,256,256,0.002962666667169995
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,256,512,0.022687999738587275
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,256,256,0.011540444360838996
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,256,128,0.002756444530354606
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,256,256,0.02299022177855174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,256,128,0.011414222419261932
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,256,64,0.002599999929467837
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,256,128,0.02251200046804216
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,256,32,0.002607999990383784
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,256,64,0.011161777708265515
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,256,32,0.01182933317290412
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,128,65536,0.011545778148704104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,128,65536,0.047968000173568726
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,128,16384,0.006695999867386288
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,128,65536,0.0975395573510064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,128,16384,0.020437333318922255
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,128,12288,0.006568888823191325
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,128,16384,0.04125688804520501
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,128,12288,0.018319111731317308
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,128,10240,0.006627555522653792
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,128,12288,0.038346668084462486
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,128,10240,0.017328888177871704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,128,8192,0.006551110910044775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,128,10240,0.03558399942186143
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,128,8192,0.01645866698688931
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,128,7168,0.0063386667105886675
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,128,8192,0.03280444608794318
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,128,7168,0.015862221519152325
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,128,6144,0.006194666855865055
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,128,7168,0.03309866786003113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,128,5120,0.006584888945023219
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,128,5120,0.030689779255125258
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,128,6144,0.015490666031837463
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,128,5120,0.014714666538768344
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,128,6144,0.03235200047492981
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,128,4096,0.006144000010357962
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,128,4096,0.014523555835088095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,128,3584,0.006578666468461354
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,128,4096,0.029623111089070637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,128,3584,0.014254222313563028
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,128,3072,0.006305777778228124
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,128,3584,0.02898933490117391
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,128,3072,0.013816888961527081
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,128,2560,0.005778666585683823
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,128,3072,0.02785244584083557
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,128,2560,0.013590222431553734
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,128,2048,0.005165333135260476
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,128,2560,0.026766222384240892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,128,2048,0.013152889079517789
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,128,1536,0.004531555705600315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,128,2048,0.02565777798493703
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,128,1536,0.012847111456924014
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,128,1024,0.003911111089918348
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,128,1536,0.02512888941499922
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,128,1024,0.012616000241703458
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,128,768,0.0036675557494163513
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,128,1024,0.02346755564212799
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,128,768,0.012168000141779581
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,128,512,0.0032906666811969546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,128,768,0.023392889234754775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,128,512,0.011892444557613797
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,128,256,0.0029111111329661477
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,128,512,0.022516444325447083
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,128,256,0.011603555745548673
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,128,128,0.002704888789190186
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,128,128,0.011316444310877057
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,128,256,0.022357333037588332
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,128,64,0.002581333327624533
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,16,128,128,0.02217688825395372
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,128,64,0.011317333413494958
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,128,32,0.00258577780591117
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,128,32,0.011607999602953592
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,64,65536,0.009525333013799455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,64,16384,0.0064239998658498125
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,64,65536,0.047467556264665395
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,64,16384,0.020364445116784837
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,64,12288,0.006324444380071428
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,64,7168,0.006222222414281633
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,64,12288,0.01809599995613098
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,64,10240,0.0064080001579390625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,64,8192,0.0063591111037466265
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,64,10240,0.01705511079894172
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,64,8192,0.01625333395269182
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,64,6144,0.006144888699054718
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,64,7168,0.015802666544914246
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,64,5120,0.006440889090299606
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,64,6144,0.01514400045077006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,64,5120,0.01464355488618215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,64,4096,0.006062222023804982
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,64,4096,0.01424711114830441
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,64,3584,0.006543111056089401
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,64,3072,0.006164444403515921
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,64,3584,0.014059555199411182
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,64,2560,0.005671111245950063
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,64,3072,0.013600889179441663
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,64,2048,0.005034666508436203
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,64,2560,0.013370667066839008
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,64,1536,0.004418666578001446
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,64,2048,0.013101333545313941
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,64,1024,0.0038000001675552795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,64,1536,0.01274577776590983
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,64,768,0.003463111196955045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,64,1024,0.01236088905069563
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,64,512,0.0031786666562159858
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,64,256,0.011494222614500256
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,64,768,0.01202666676706738
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,64,512,0.011892444557613797
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,64,256,0.0028533333291610083
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,64,128,0.0026933333526055017
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,64,64,0.002550222186578645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,64,128,0.011330666641394297
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,64,64,0.011149333582984077
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,64,32,0.0025635556214385563
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,64,32,0.011801777614487542
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,32,65536,0.00886044485701455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,32,16384,0.006534222099516127
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,32,65536,0.046712001164754234
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,32,16384,0.02035911050107744
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,32,12288,0.0064497776329517365
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,32,10240,0.0064035554726918536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,32,12288,0.018345778187115986
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,32,8192,0.006274666637182236
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,32,10240,0.017117333081033494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,32,7168,0.006159111029571957
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,32,8192,0.016303999556435477
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,32,7168,0.015816888875431485
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,32,6144,0.006020444548792309
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,32,5120,0.0063582224150498705
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,32,6144,0.01535288823975457
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,32,4096,0.006064888917737537
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,32,5120,0.014767999450365702
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,32,3584,0.00637511122557852
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,32,4096,0.014333332578341166
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,32,3072,0.006122666514582104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,32,3584,0.014077333940400017
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,32,2560,0.00563466673096021
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,32,3072,0.01366222235891554
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,32,2048,0.005052444421582752
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,32,2560,0.013491555220550962
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,32,2048,0.013155555559529198
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,32,1536,0.004448889030350579
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,32,1024,0.0037911112109820047
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,32,1536,0.012610666453838348
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,32,768,0.0034835553831524322
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,32,1024,0.012260444462299347
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,32,512,0.003173333282272021
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,32,768,0.011948444777064852
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,32,256,0.0028479999552170434
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,32,512,0.011852444046073489
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,32,256,0.011529777612951068
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,32,128,0.0027217778066794076
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,32,128,0.011232000258233813
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,32,64,0.002537777854336633
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,32,64,0.01125600023402108
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,16,32,32,0.002526222210791376
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,16,32,32,0.011400000088744693
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,65536,10240,0.16839467154608834
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,65536,12288,0.19868000348409018
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,65536,12288,0.23912533124287924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,65536,16384,0.29616978433397084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,65536,16384,0.25628532303704155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,65536,10240,0.32776978280809194
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,65536,10240,0.20620621575249565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,65536,12288,0.3915226724412706
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,65536,7168,0.13484266069200304
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,65536,8192,0.26293423440721303
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,65536,8192,0.1396577755610148
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,65536,8192,0.16947289307912192
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,65536,16384,0.5228595733642578
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,65536,7168,0.1527671151691013
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,65536,6144,0.10826755232281154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,65536,6144,0.2116302251815796
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,65536,6144,0.1348008844587538
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,65536,5120,0.09324177768495347
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,65536,7168,0.229967991511027
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,65536,4096,0.07813244395785861
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,65536,5120,0.11911643875969781
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,65536,4096,0.13397688335842556
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,65536,3584,0.07147555881076388
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,65536,4096,0.0994044476085239
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,65536,5120,0.16616800096299914
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,65536,3584,0.1164346668455336
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,65536,3072,0.06814577844407824
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,65536,3584,0.09203110800849067
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,65536,3072,0.1148151159286499
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,65536,2560,0.05947555436028374
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,65536,3072,0.08339199754926895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,65536,2560,0.07269689109590319
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,65536,2048,0.048672888014051646
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,65536,2560,0.08425244357850815
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,65536,1536,0.04224977890650431
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,65536,1536,0.054661333560943604
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,65536,1024,0.036510222487979464
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,65536,1024,0.036165333456463285
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,65536,2048,0.06280444728003608
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,65536,1024,0.04639644424120585
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,65536,768,0.03374933203061422
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,65536,1536,0.052128000391854175
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,65536,2048,0.06880178054173787
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,65536,768,0.04045155644416809
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,65536,256,0.026686223016844854
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,65536,512,0.030432889858881634
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,65536,128,0.00738666703303655
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,65536,256,0.0315688881609175
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,65536,512,0.03533777925703261
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,65536,512,0.01999022232161628
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,65536,128,0.02500444485081567
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,65536,256,0.011235555840863122
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,65536,64,0.005664000080691443
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,65536,32,0.021789333886570398
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,65536,128,0.028612444798151653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,65536,32,0.004313777718279096
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,65536,64,0.023187556200557288
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,65536,768,0.028163555595609877
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,16384,12288,0.06177600224812826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,16384,16384,0.07764444748560588
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,16384,16384,0.09885866774453057
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,16384,12288,0.0945991079012553
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,16384,10240,0.054456889629364014
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,16384,16384,0.12463377581702338
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,16384,12288,0.08110844426684909
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,16384,10240,0.07643555270300971
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,16384,8192,0.05083733465936449
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,16384,10240,0.079894224802653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,16384,8192,0.064301331837972
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,16384,8192,0.06457955307430692
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,16384,7168,0.04327377676963806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,16384,7168,0.056883553663889565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,16384,6144,0.03920355439186096
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,16384,7168,0.06004889143837822
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,16384,6144,0.055083556307686694
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,16384,5120,0.035338666703965925
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,16384,5120,0.041915555795033775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,16384,6144,0.049505776829189725
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,16384,65536,0.2616257667541504
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,16384,4096,0.031410667631361223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,16384,5120,0.0525546669960022
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,16384,4096,0.034408890538745456
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,16384,3584,0.030047112041049536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,16384,4096,0.04632000128428141
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,16384,3584,0.04347733325428433
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,16384,3072,0.02818044357829624
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,16384,2560,0.03868622250027127
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,16384,3584,0.030349334081013996
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,16384,3072,0.04141511188613044
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,16384,2560,0.025785777303907607
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,16384,3072,0.026579555537965562
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,16384,2560,0.023015111684799194
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,16384,1536,0.02124622298611535
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,16384,2048,0.0232595553000768
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,16384,2048,0.03507377703984579
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,16384,1536,0.03204088740878635
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,16384,1536,0.015121777852376303
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,16384,1024,0.01903199983967675
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,16384,2048,0.019077334139082167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,16384,768,0.00904622260067198
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,16384,1024,0.030376000536812678
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,16384,768,0.017826666434605915
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,16384,512,0.007086222370465596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,16384,512,0.016247999336984422
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,16384,768,0.029256890217463177
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,16384,256,0.005021333280536863
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,16384,1024,0.010991999672518836
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,16384,512,0.027979556057188246
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,16384,128,0.0034479999707804788
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,16384,256,0.013405332962671915
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,16384,256,0.023892444041040208
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,16384,128,0.013279111021094851
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,16384,64,0.00319377767542998
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,16384,128,0.02290311124589708
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,16384,32,0.0031973332580592898
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,16384,64,0.01402933398882548
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,16384,65536,0.30960000885857475
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,16384,32,0.012826667063766055
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,12288,16384,0.06347378094991048
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,12288,12288,0.07953333192401461
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,12288,16384,0.0989217758178711
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,12288,16384,0.10511110888587104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,12288,12288,0.050959110260009766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,12288,12288,0.07943466636869642
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,12288,10240,0.045440001620186694
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,12288,10240,0.06719644202126397
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,12288,10240,0.0724675522910224
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,16384,65536,0.5468106799655491
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,12288,8192,0.040017777019076876
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,12288,8192,0.05371377865473429
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,12288,65536,0.20599999692704943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,12288,8192,0.06352266338136461
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,12288,7168,0.036665777365366616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,12288,6144,0.03370577759212918
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,12288,6144,0.055171555942959256
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,12288,7168,0.05949422385957506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,12288,7168,0.04787288771735298
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,12288,6144,0.041555556986067034
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,12288,5120,0.03091199861632453
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,12288,4096,0.028767112228605483
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,12288,5120,0.051847110191980995
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,12288,5120,0.038464887274636164
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,12288,4096,0.045491556326548256
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,12288,3584,0.04282666577233208
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,12288,3584,0.02757866680622101
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,12288,4096,0.02849777870708042
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,12288,3584,0.025433777107132807
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,12288,3072,0.040559109714296125
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,12288,3072,0.024663110574086506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,12288,2560,0.022816000713242426
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,12288,3072,0.022289777795473736
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,12288,2048,0.02124533388349745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,12288,2560,0.03829333186149597
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,12288,2048,0.034492443005243935
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,12288,2560,0.019135110908084445
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,12288,1536,0.020019556085268658
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,12288,1024,0.009307555854320526
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,12288,1024,0.018287110659811232
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,12288,1536,0.03216444452603658
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,12288,2048,0.015950222810109455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,12288,768,0.007807110746701558
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,12288,1024,0.029069334268569946
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,12288,1536,0.012661333713266583
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,12288,768,0.017318222257826064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,12288,512,0.006400888992680444
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,12288,768,0.028586665789286297
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,12288,512,0.014211555322011312
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,12288,256,0.0037671112351947357
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,12288,512,0.026506667335828144
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,12288,128,0.0033271111961868075
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,12288,256,0.023522666758961145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,12288,256,0.013140444126394061
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,12288,128,0.013032000097963544
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,12288,64,0.003002666764789157
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,12288,64,0.013854222165213691
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,12288,32,0.003047111133734385
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,12288,128,0.02276977731121911
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,12288,32,0.012816888590653738
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,12288,65536,0.2942284478081597
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,10240,16384,0.056445333692762584
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,10240,16384,0.07989688714345296
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,10240,12288,0.04688088761435615
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,10240,12288,0.06958399878607856
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,10240,16384,0.09089244736565484
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,12288,65536,0.4013066556718614
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,10240,12288,0.07010577784644233
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,10240,10240,0.04194577866130405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,10240,10240,0.0649964412053426
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,10240,8192,0.03663288884692722
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,10240,10240,0.05951999955707126
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,10240,8192,0.048551999860339694
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,10240,7168,0.034090667963027954
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,10240,8192,0.05654222435421414
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,10240,7168,0.03885066509246826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,10240,7168,0.05238311158286201
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,10240,6144,0.03183199961980184
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,10240,65536,0.17920266257392037
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,10240,6144,0.03384000062942505
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,10240,5120,0.02909244431389703
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,10240,6144,0.047705776161617704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,10240,5120,0.044680890109803945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,10240,4096,0.027730666928821143
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,10240,5120,0.028864890336990356
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,10240,4096,0.023618666662110224
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,10240,3584,0.024910221497217815
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,10240,4096,0.040814220905303955
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,10240,3072,0.024335111180941265
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,10240,3072,0.036350223753187395
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,10240,3584,0.03875999980502658
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,10240,2560,0.02199199961291419
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,10240,3584,0.020985777179400127
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,10240,2560,0.03471822208828396
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,10240,2048,0.02071911096572876
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,10240,3072,0.018538667096032035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,10240,2560,0.015985776980717976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,10240,2048,0.03197777933544583
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,10240,1536,0.010392000277837118
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,10240,65536,0.23581600189208984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,10240,1536,0.019324445062213473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,10240,2048,0.013208000196350945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,10240,1024,0.008194666769769456
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,10240,1024,0.01796000036928389
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,10240,1536,0.030456887351142034
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,10240,768,0.00700800038046307
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,10240,1024,0.028524445162879095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,10240,768,0.01534400052494473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,10240,512,0.005705777969625261
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,10240,512,0.014031999640994601
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,10240,768,0.02827022141880459
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,10240,256,0.003636444608370463
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,10240,512,0.025157334076033697
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,10240,256,0.0131217779384719
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,10240,128,0.003169777699642711
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,10240,256,0.02347466680738661
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,10240,128,0.012818666795889536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,10240,32,0.0029253332565228143
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,10240,64,0.0029928889125585556
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,10240,128,0.023131555981106226
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,10240,64,0.013478221992651621
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,10240,32,0.01258666647805108
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,8192,16384,0.04680000080002678
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,8192,16384,0.07296178076002333
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,8192,12288,0.03877244393030802
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,8192,16384,0.06734577814737956
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,8192,12288,0.05236266718970405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,8192,10240,0.03513688842455546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,8192,12288,0.06278666522767809
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,8192,10240,0.0417537788550059
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,8192,10240,0.05736266904407077
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,8192,8192,0.03131644593344794
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,8192,8192,0.03408177693684896
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,8192,65536,0.1405644416809082
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,8192,8192,0.05041244294908312
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,8192,7168,0.029073778125974867
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,8192,7168,0.04758400056097242
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,8192,7168,0.03020711077584161
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,8192,6144,0.02685333291689555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,8192,6144,0.04535911149448819
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,8192,6144,0.026530666483773127
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,8192,5120,0.02479644450876448
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,8192,5120,0.04313155677583483
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,8192,65536,0.21266578303443062
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,8192,4096,0.02152266601721446
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,8192,5120,0.02292888859907786
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,8192,3584,0.0206666671567493
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,8192,3584,0.036590221855375506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,8192,3072,0.019647111495335896
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,8192,4096,0.037485334608289934
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,10240,65536,0.36981601185268825
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,8192,3072,0.03461511267556085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,8192,2560,0.018425777554512024
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,8192,3584,0.017166222135225933
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,8192,3072,0.015200000670221118
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,8192,4096,0.019539554913838703
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,8192,2048,0.011186666786670685
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,8192,2048,0.017326222525702584
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,8192,2560,0.03404888841840956
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,8192,2560,0.01311644415060679
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,8192,1536,0.009175111022260454
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,8192,1536,0.016543999314308167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,8192,2048,0.03160266743765937
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,8192,768,0.006272888845867581
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,8192,1024,0.007198221981525421
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,8192,1024,0.028343998723559912
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,8192,1024,0.014893333117167154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,8192,1536,0.03021866745418972
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,8192,768,0.012795555922720166
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,8192,512,0.005145777844720417
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,8192,512,0.012352888782819113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,8192,768,0.026874666412671406
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,8192,256,0.0036640001667870414
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,8192,512,0.024132443798912898
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,8192,256,0.011774222056070963
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,8192,128,0.003236444460021125
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,8192,256,0.023188443647490606
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,8192,128,0.011720889144473605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,8192,64,0.0029884444342719186
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,8192,128,0.022544888986481562
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,8192,32,0.0029724445194005966
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,8192,64,0.012450666891203986
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,8192,32,0.011378666592968835
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,7168,16384,0.04313866628540886
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,7168,16384,0.07105866405698988
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,7168,12288,0.04910133282343546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,8192,65536,0.26467466354370117
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,7168,12288,0.036121779017978244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,7168,16384,0.06094666984346178
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,7168,12288,0.06006399790445963
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,7168,10240,0.034854223330815635
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,7168,10240,0.05642133288913303
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,7168,10240,0.0417119993103875
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,7168,8192,0.02937155630853441
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,7168,8192,0.03353600038422479
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,7168,8192,0.0502782232231564
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,7168,7168,0.02754666739039951
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,7168,7168,0.04684622089068095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,7168,6144,0.02565866708755493
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,7168,7168,0.03033511175049676
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,7168,65536,0.12710932890574136
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,7168,6144,0.026709333062171936
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,7168,6144,0.04353866643375821
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,7168,5120,0.041929778125551015
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,7168,5120,0.0236826671494378
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,7168,4096,0.020717332760492962
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,7168,5120,0.02349600030316247
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,7168,4096,0.037273777855767146
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,7168,3584,0.01980977753798167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,7168,4096,0.020091555184788175
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,7168,3584,0.03621244430541992
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,7168,3072,0.016230222251680162
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,7168,3072,0.01869244376818339
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,7168,3584,0.018086221483018663
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,7168,3072,0.03478844298256768
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,7168,2560,0.01789422167672051
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,7168,2048,0.010444444086816577
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,7168,2560,0.032383110788133405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,7168,2048,0.016923555069499545
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,7168,1536,0.00869688888390859
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,7168,2560,0.012106666962305704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,7168,2048,0.03071911136309306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,7168,1536,0.015899555550681222
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,7168,1024,0.006788444601827198
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,7168,1536,0.029105779197480943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,7168,1024,0.01458755632241567
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,7168,768,0.005917333480384614
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,7168,1024,0.027215111586782668
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,7168,768,0.012645333177513547
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,7168,512,0.00444088876247406
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,7168,768,0.0243857784403695
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,7168,512,0.012255999777052136
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,7168,256,0.0033084443873829315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,7168,512,0.024375999967257183
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,7168,256,0.011697777443461947
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,7168,128,0.003061333257291052
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,7168,256,0.022848889231681824
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,7168,128,0.011519999967681037
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,7168,64,0.002879111096262932
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,7168,128,0.02253866692384084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,7168,64,0.012034667034943899
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,7168,32,0.002854222224818336
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,7168,32,0.011208888557222156
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,7168,65536,0.19353777832455107
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,6144,16384,0.040720889965693154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,6144,16384,0.06917777988645765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,6144,16384,0.0517537792523702
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,6144,12288,0.03429333368937174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,6144,12288,0.0395706660217709
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,6144,10240,0.031114667654037476
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,6144,10240,0.05654578076468574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,6144,12288,0.06112800041834513
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,6144,10240,0.033943110042148165
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,6144,8192,0.027927110592524212
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,6144,8192,0.027981334262424048
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,6144,65536,0.11894133355882432
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,6144,8192,0.04978933268123203
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,6144,7168,0.024873778223991394
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,6144,7168,0.026106667187478807
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,6144,7168,0.046486222081714206
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,6144,6144,0.02414488957987891
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,7168,65536,0.23188800281948516
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,6144,5120,0.021879111727078755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,6144,6144,0.021853332718213398
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,6144,6144,0.04336533281538221
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,6144,5120,0.040447112586763166
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,6144,65536,0.19301511181725395
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,6144,4096,0.020117332537968952
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,6144,5120,0.018936889039145578
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,6144,3584,0.0192266669538286
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,6144,4096,0.036307556761635676
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,6144,3584,0.03550666570663452
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,6144,3072,0.01829955478509267
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,6144,2560,0.010998222563001843
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,6144,3072,0.03486844566133287
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,6144,2560,0.017467555072572496
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,6144,3584,0.01549600064754486
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,6144,2048,0.009693333672152625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,6144,3072,0.01259022206068039
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,6144,4096,0.016232000456915963
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,6144,2560,0.032588445478015475
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,6144,2048,0.016493333710564505
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,6144,1024,0.0063946665161185795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,6144,1536,0.008037333687146505
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,6144,1024,0.01365066650840971
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,6144,2048,0.030373334884643555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,6144,1536,0.01569333341386583
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,6144,1536,0.02910311023394267
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,6144,768,0.005583999885453119
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,6144,1024,0.027416000763575237
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,6144,768,0.012523555921183692
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,6144,512,0.004104888687531154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,6144,768,0.02474044428931342
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,6144,512,0.012230222423871359
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,6144,256,0.0033235556135574975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,6144,512,0.02362755603260464
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,6144,256,0.011703111231327057
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,6144,128,0.0030577776746617425
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,6144,256,0.02296177711751726
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,6144,128,0.01165244479974111
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,6144,64,0.002801777794957161
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,6144,32,0.002893333426780171
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,6144,64,0.012016889121797351
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,6144,128,0.022648000054889258
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,6144,32,0.011165333290894827
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,5120,16384,0.03660710983806186
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,5120,16384,0.06627377536561754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,6144,65536,0.21387999587588838
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,5120,16384,0.04888888862397936
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,5120,12288,0.030799110730489094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,5120,12288,0.03784977727466159
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,5120,12288,0.06109244293636746
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,5120,10240,0.028284443749321833
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,5120,10240,0.03246577911906772
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,5120,8192,0.02605955633852217
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,5120,10240,0.05368711219893562
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,5120,8192,0.04645511176851061
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,5120,8192,0.027100443840026855
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,5120,65536,0.1013573341899448
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,5120,7168,0.023864888482623633
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,5120,6144,0.022163555026054382
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,5120,7168,0.024632889363500807
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,5120,6144,0.04220799936188591
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,5120,5120,0.02056711084312863
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,5120,7168,0.04641955428653293
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,5120,6144,0.018195556269751657
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,5120,5120,0.03968000080850389
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,5120,4096,0.019018666611777413
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,5120,5120,0.01610666679011451
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,5120,3584,0.018567111757066514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,5120,4096,0.036624001132117376
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,5120,3072,0.010900444454616971
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,5120,3584,0.0347688893477122
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,5120,4096,0.01369599997997284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,5120,3072,0.017525333497259352
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,5120,3584,0.011976000335481433
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,5120,2560,0.009673777553770278
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,5120,3072,0.03454400102297465
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,5120,2560,0.01685688893000285
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,5120,2048,0.008453333543406593
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,5120,2560,0.03219555483924018
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,5120,2048,0.016013332539134555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,5120,1536,0.007256888680987888
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,5120,1536,0.0136835558546914
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,5120,2048,0.030237333642111883
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,5120,1024,0.005924444645643234
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,5120,1536,0.02886399957868788
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,5120,1024,0.012543111211723752
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,5120,768,0.0047191112405723995
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,5120,1024,0.02609155575434367
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,5120,768,0.012328888807031842
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,5120,512,0.0038906666967603895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,5120,768,0.02426133387618595
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,5120,512,0.012103111379676394
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,5120,256,0.0032302221904198327
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,5120,65536,0.17295289039611816
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,5120,512,0.023440000083711412
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,5120,256,0.01162577751610014
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,5120,128,0.002991111121243901
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,5120,256,0.02304977840847439
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,5120,128,0.011345778074529437
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,5120,64,0.002756444530354606
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,5120,32,0.002805333377586471
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,5120,128,0.02310488952530755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,5120,64,0.011961778004964193
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,5120,32,0.010967999696731567
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,5120,65536,0.18569600582122803
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,4096,16384,0.032079110542933144
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,4096,16384,0.06441333558824328
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,4096,12288,0.03071199854214986
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,4096,16384,0.0394720004664527
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,4096,12288,0.029102222787009344
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,4096,10240,0.025166221790843542
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,4096,12288,0.05486844314469231
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,4096,10240,0.026847110854254827
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,4096,10240,0.05137422349717882
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,4096,8192,0.02306844459639655
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,4096,7168,0.020993777447276648
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,4096,8192,0.022421333524915908
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,4096,8192,0.04700266652637058
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,4096,6144,0.01994044416480594
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,4096,7168,0.04425244530042013
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,4096,7168,0.020414221617910597
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,4096,65536,0.08230044444402058
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,4096,5120,0.018831110662884183
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,4096,6144,0.041414221127827965
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,4096,5120,0.0391244457827674
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,4096,6144,0.01832711034350925
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,4096,4096,0.01778133379088508
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,4096,4096,0.035932444863849215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,4096,3584,0.012445333103338877
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,4096,5120,0.015576889117558798
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,4096,3584,0.017255999975734286
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,4096,4096,0.013524444566832649
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,4096,3072,0.011233777635627322
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,4096,3584,0.034789333740870156
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,4096,3072,0.016697777642144095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,4096,2560,0.008488000267081791
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,4096,3072,0.03294577863481309
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,4096,2560,0.016142222616407607
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,4096,2048,0.007419555551475949
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,4096,2560,0.032418668270111084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,4096,2048,0.015152888165579902
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,4096,1536,0.0064035554726918536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,4096,2048,0.030008888906902734
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,4096,1024,0.005361777626805835
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,4096,1536,0.01291822228166792
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,4096,1536,0.027095110880004034
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,4096,768,0.0038702223036024305
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,4096,1024,0.012668444050682915
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,4096,1024,0.025255110528733995
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,4096,768,0.01216533366176817
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,4096,512,0.003546666767862108
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,4096,768,0.024142222272025213
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,4096,512,0.012128888732857175
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,4096,256,0.0031884445084465873
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,4096,512,0.023896000451511804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,4096,65536,0.17380889256795248
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,4096,256,0.011505777637163797
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,4096,128,0.0028657778683635923
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,4096,256,0.023119111855824787
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,4096,128,0.011353777514563667
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,4096,64,0.0027422221998373666
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,4096,128,0.023120888405376013
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,4096,32,0.0027297778675953546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,4096,64,0.011794666449228922
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,4096,32,0.010934222075674268
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3584,16384,0.0299253331290351
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,4096,65536,0.1331768830617269
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,3584,16384,0.06411911381615533
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3584,16384,0.033597333563698664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3584,12288,0.027494221925735474
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,3584,12288,0.05420710974269443
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3584,12288,0.0266248881816864
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3584,10240,0.024005333582560223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3584,10240,0.023090667194790308
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,3584,10240,0.05010044574737549
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3584,8192,0.021446221404605444
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,3584,8192,0.0455297761493259
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3584,7168,0.020288000504175823
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,3584,7168,0.043021334542168506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3584,6144,0.019306666321224637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3584,7168,0.01794488893614875
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,3584,6144,0.03993155558904012
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3584,8192,0.01977511081430647
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3584,6144,0.016006223029560514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3584,5120,0.018312000566058688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3584,4096,0.012737777498033313
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,3584,5120,0.03736977775891622
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3584,4096,0.017679111825095285
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3584,5120,0.01442311041884952
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,3584,4096,0.035038222869237266
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3584,3584,0.011896889242861005
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3584,3584,0.01702933344576094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3584,65536,0.07424977752897474
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,3584,3584,0.033575998412238225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3584,3072,0.010834666589895884
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3584,3072,0.016401777664820354
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,3584,65536,0.16400444507598877
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3584,2560,0.007968000239796108
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3584,2560,0.015999999311235216
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,3584,3072,0.03333688775698344
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3584,2048,0.007110222346252865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,3584,2560,0.03122488988770379
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3584,2048,0.014460444450378418
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3584,1536,0.006286222073766921
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3584,1536,0.01294311136007309
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,3584,2048,0.029608001311620075
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3584,1024,0.004928888960017098
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,3584,1536,0.02649333410792881
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3584,1024,0.012507555385430654
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3584,768,0.004039111236731212
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,3584,1024,0.02458222210407257
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3584,768,0.012160888976520963
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3584,512,0.003488000068399641
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,3584,768,0.024496000674035814
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3584,512,0.011981333295504252
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3584,256,0.003156444471743372
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,3584,512,0.02316444449954563
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3584,256,0.011581333147154914
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3584,128,0.0028719999310043124
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,3584,128,0.0229111115137736
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,3584,256,0.023319111929999456
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3584,128,0.011477332976129321
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3584,64,0.002692444456948174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3584,32,0.0027137777457634607
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3584,64,0.011951111257076263
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3584,32,0.010817777779367236
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3584,65536,0.1211822165383233
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3072,16384,0.027971555789311726
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,3072,16384,0.06004977888531155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3072,12288,0.025136000580257837
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3072,12288,0.024111110303137038
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3072,10240,0.021905778182877436
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3072,16384,0.030425777037938435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,3072,10240,0.04955910974078708
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,3072,12288,0.05443111062049866
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3072,8192,0.020346666375796
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,3072,8192,0.04565688967704773
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3072,7168,0.019493333167499967
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3072,10240,0.021224000387721594
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3072,8192,0.018210666047202218
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3072,6144,0.014820444915029736
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,3072,7168,0.04178933302561442
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3072,6144,0.018375999397701688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3072,7168,0.016525333126386006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3072,5120,0.013258666627936892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,3072,6144,0.03995733459790548
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3072,5120,0.017815111411942374
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3072,4096,0.011957333319716983
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,3072,5120,0.03810311026043362
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3072,4096,0.01698755555682712
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3072,3584,0.011000889043013254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,3072,4096,0.03492355677816603
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3072,3584,0.016671111186345417
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3072,3072,0.010267555713653564
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,3072,3584,0.03347911106215583
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3072,3072,0.016047999262809753
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3072,2560,0.007708444363541073
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,3072,3072,0.03293422195646498
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3072,2560,0.015311110350820752
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3072,2048,0.00684799998998642
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3072,65536,0.06782666842142741
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,3072,2560,0.031136890252431233
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3072,1536,0.006136888845099344
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3072,2048,0.01408088869518704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,3072,65536,0.1498088836669922
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3072,1024,0.004411555412742826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3072,1536,0.013049778011110095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,3072,2048,0.02663199934694502
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,3072,1536,0.025804445147514343
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3072,1024,0.01256444470749961
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3072,768,0.003959111041492886
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,3072,1024,0.02475999957985348
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3072,768,0.01218844453493754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3072,256,0.00314666661951277
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3072,512,0.0034755555291970572
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,3072,768,0.02409422232045068
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3072,512,0.011913778053389655
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,3072,512,0.023232888844278123
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3072,256,0.011500444677140979
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3072,128,0.0028435554769304064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,3072,256,0.023345778385798138
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3072,128,0.011147555377748279
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3072,64,0.0027111110587914786
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,3072,128,0.0221733334991667
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3072,64,0.011782222323947482
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3072,32,0.0027822222974565295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,3072,32,0.011272889044549731
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2560,16384,0.026906667484177485
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,2560,16384,0.05685510900285509
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2560,12288,0.0214711113108529
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2560,16384,0.027113777067926195
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,3072,65536,0.10448622041278416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,2560,12288,0.05231644378768074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2560,10240,0.021249777740902368
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2560,12288,0.022807111342748005
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2560,8192,0.01980177727010515
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2560,65536,0.06475910875532362
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,2560,10240,0.04960977700021532
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,2560,8192,0.0454311105940077
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2560,10240,0.01869066721863217
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2560,8192,0.01605599953068627
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2560,7168,0.019089778264363606
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2560,6144,0.0129120002190272
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,2560,7168,0.0411208901140425
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2560,6144,0.01789688898457421
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2560,5120,0.011853333148691388
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2560,5120,0.01755466726091173
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,2560,6144,0.039854221873813205
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2560,4096,0.010421333213647207
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,2560,5120,0.038155555725097656
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2560,7168,0.014574221438831754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,2560,65536,0.14220000637902155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2560,4096,0.01685333251953125
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2560,3584,0.009918221996890174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2560,3072,0.009328888522254096
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2560,3584,0.01608800060219235
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2560,3072,0.014632888966136508
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2560,2560,0.007453333172533248
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,2560,3584,0.034252444903055825
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,2560,3072,0.03246577911906772
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,2560,4096,0.03354488809903463
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2560,2560,0.013884444203641681
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2560,2048,0.006647111227114995
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,2560,2560,0.030815998713175457
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2560,2048,0.013400000002649097
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2560,1536,0.005714666512277391
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,2560,2048,0.02683111031850179
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2560,1536,0.013005333642164866
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2560,1024,0.0041831110914548235
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,2560,1536,0.02586311101913452
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2560,768,0.0038133333954546186
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2560,1024,0.012283555335468717
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,2560,1024,0.024753777517212763
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2560,768,0.012281777958075205
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2560,256,0.0030986666679382324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2560,512,0.0034888887570963967
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,2560,768,0.02348088887002733
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2560,65536,0.09494844410154556
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,2560,256,0.023170666562186346
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2560,512,0.011918221910794577
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,2560,512,0.023035556077957153
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2560,256,0.011422221859296164
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2560,128,0.0028364445186323593
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2560,128,0.011482666763994427
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2560,64,0.002672000063790215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,2560,128,0.02200977835390303
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2560,64,0.011935110721323224
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2560,32,0.00267999991774559
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2560,32,0.011176000038782755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,2048,16384,0.05893422497643364
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2048,12288,0.018343110879262287
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2048,16384,0.025401777691311304
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2048,16384,0.022295999858114455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2048,12288,0.022239999638663396
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,2048,12288,0.051790220869912036
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2048,10240,0.0203804439968533
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,2048,10240,0.048325333330366343
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2048,8192,0.01943288909064399
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2048,65536,0.05973333120346069
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2048,10240,0.01609244445959727
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,2048,8192,0.044010665681627065
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2048,8192,0.01386844449573093
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2048,7168,0.012780444489585029
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2048,7168,0.018499554859267343
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2048,6144,0.011894221935007306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,2048,7168,0.041675554381476514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2048,6144,0.017875555488798354
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,2048,65536,0.1596497827106052
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2048,5120,0.010880000061459012
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,2048,6144,0.03993955585691664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2048,5120,0.017126222451527912
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2048,4096,0.009760888914267222
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,2048,5120,0.03722666700681051
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2048,4096,0.016184889607959323
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2048,3584,0.008991999758614434
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,2048,4096,0.03450489044189453
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2048,3584,0.01534933348496755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2048,3072,0.008647999829716152
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,2048,3584,0.03360533383157518
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2048,3072,0.014129777749379476
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,2048,2560,0.02888266576661004
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2048,2048,0.013199999928474426
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2048,2560,0.007179555793603261
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,2048,3072,0.02968533171547784
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2048,2048,0.00645955569214291
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2048,2560,0.014044443766276041
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2048,65536,0.07304710812038846
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,2048,2048,0.02688088847531213
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2048,1536,0.005434666656785541
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2048,1536,0.01291822228166792
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2048,1024,0.004038222134113312
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2048,1024,0.0123857781291008
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,2048,1536,0.026293334033754136
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2048,768,0.0037377778854635027
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,2048,1024,0.024357333779335022
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2048,768,0.012179555164443122
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2048,512,0.0033902221669753394
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,2048,768,0.02392533255947961
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2048,512,0.01183199965291553
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2048,256,0.00304888892504904
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,2048,512,0.023715555667877197
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2048,256,0.01146755533085929
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2048,128,0.0027964444210131965
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,2048,256,0.022914666268560622
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2048,128,0.011329777538776398
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2048,64,0.002671111168132888
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,2048,128,0.022309333086013794
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,2048,32,0.0026897777699761917
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2048,64,0.01181777815024058
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,2048,32,0.010936888555685679
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1536,16384,0.02355466617478265
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1536,12288,0.020742222666740417
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,1536,16384,0.057550224992964
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1536,12288,0.01459822224246131
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1536,10240,0.013092444174819522
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1536,16384,0.018008000320858426
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,1536,12288,0.050959110260009766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1536,10240,0.019906666543748643
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1536,8192,0.011741333537631564
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,1536,10240,0.04857155680656433
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1536,8192,0.018796443939208984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1536,7168,0.010904889139864178
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,1536,8192,0.0444106658299764
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1536,7168,0.017789334058761597
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1536,6144,0.010031999813185798
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,1536,7168,0.04183822207980686
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1536,6144,0.01739555597305298
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1536,5120,0.009445333646403419
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1536,5120,0.016487111647923786
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1536,65536,0.05592088566886055
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,1536,6144,0.03895466526349386
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1536,4096,0.008738666772842407
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,1536,5120,0.03688888748486837
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1536,4096,0.014653333359294467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1536,3584,0.008553778131802877
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,1536,4096,0.03315111001332601
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1536,3584,0.014496889379289416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1536,3072,0.008103999826643202
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,1536,65536,0.13647733794318304
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,1536,3584,0.0321084459622701
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1536,3072,0.013760888742076026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1536,2560,0.006984000404675801
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,1536,3072,0.029600001043743555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1536,2560,0.013573333621025085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1536,2048,0.013066666821638743
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,1536,2560,0.02834222382969326
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1536,2048,0.005475555443101459
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1536,1536,0.004915555732117759
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,1536,2048,0.026314665873845417
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1536,1536,0.012803555362754397
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1536,65536,0.05442400111092461
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1536,1024,0.004039999925427967
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,1536,1536,0.025592888394991558
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1536,1024,0.012367111113336353
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1536,768,0.003645333151022593
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,1536,1024,0.02420088814364539
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1536,768,0.012071111136012606
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1536,512,0.0033644443998734155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,1536,768,0.02384800049993727
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1536,512,0.011815999945004782
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1536,256,0.003035555490189128
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1536,256,0.011446221835083432
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,1536,512,0.023572444915771484
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1536,128,0.002789333255754577
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,1536,256,0.02273155583275689
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1536,128,0.011237333218256632
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1536,64,0.0026364444444576898
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,1536,128,0.022427555587556627
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1536,32,0.0027111110587914786
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1536,64,0.011602666642930774
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1536,32,0.010760889285140567
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1024,16384,0.022052443689770166
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1024,12288,0.0120328888297081
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,1024,16384,0.04812444580925835
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1024,12288,0.020257777637905545
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1024,16384,0.01444533301724328
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1024,10240,0.011140444212489657
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,1024,12288,0.04502311017778185
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1024,10240,0.019090667366981506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1024,8192,0.010017777482668558
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,1024,10240,0.04358488983578152
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1024,8192,0.01798311041461097
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1024,65536,0.04865422182612949
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1024,7168,0.009285333255926767
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,1024,8192,0.04088177614741855
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1024,6144,0.008362666600280339
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1024,7168,0.017369776964187622
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1024,6144,0.016075554821226332
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,1024,7168,0.037929776642057635
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1024,5120,0.008450667063395182
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,1024,65536,0.12010755803849961
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,1024,6144,0.03546844588385688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1024,5120,0.01516711049609714
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1024,4096,0.007736889024575551
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1024,3584,0.0075333333677715724
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,1024,5120,0.03322044346067641
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,1024,4096,0.03119200136926439
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1024,65536,0.03949955436918471
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1024,4096,0.014862222803963555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1024,3584,0.014537778165605334
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1024,3072,0.0070364442136552595
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1024,3072,0.014107555150985718
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1024,2560,0.006494222415818109
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,1024,3584,0.02977422210905287
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,1024,3072,0.028916445043351915
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1024,2048,0.005947555518812603
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1024,2560,0.013762666947311826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1024,1536,0.013188444077968597
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,1024,2560,0.027201778358883325
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1024,2048,0.013411555853154926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1024,1536,0.004729777574539185
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,1024,2048,0.025797333982255723
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1024,1024,0.004056888735956616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,1024,1536,0.024851555625597637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1024,768,0.003748444633351432
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1024,1024,0.012464889221721225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,1024,1024,0.02362488872475094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1024,256,0.0030497778207063675
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1024,768,0.012335999972290464
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1024,512,0.0033599999215867785
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,1024,768,0.023802666200531855
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1024,512,0.011848889291286469
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,1024,512,0.022861333356963262
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1024,64,0.0026577777332729767
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1024,256,0.01147555559873581
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1024,128,0.0028168888141711554
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,1024,256,0.022656889425383672
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1024,128,0.011364444262451597
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,1024,128,0.021905778182877436
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,1024,32,0.002696000039577484
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1024,64,0.011732444167137146
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,1024,32,0.01108266661564509
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,768,16384,0.011634666886594562
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,768,16384,0.02165066699186961
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,768,12288,0.009949333137936061
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,768,16384,0.045612444480260216
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,768,12288,0.019856888386938307
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,768,10240,0.009718221922715506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,768,10240,0.019118222925398085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,768,8192,0.00902133352226681
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,768,12288,0.04349955585267809
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,768,8192,0.016546666622161865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,768,65536,0.047172443734274976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,768,10240,0.04094400008519491
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,768,7168,0.0083155557513237
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,768,8192,0.03859555390146043
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,768,7168,0.016343999240133498
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,768,65536,0.10477066702312893
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,768,6144,0.008007110820876228
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,768,7168,0.033392889632119074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,768,6144,0.015615999698638916
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,768,5120,0.007948444121413762
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,768,65536,0.03050222330623203
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,768,6144,0.03289777702755398
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,768,4096,0.01462933255566491
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,768,4096,0.007110222346252865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,768,5120,0.03197688857714335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,768,5120,0.014967999524540372
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,768,3584,0.007129777636792924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,768,4096,0.028006222512986925
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,768,3584,0.01445777714252472
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,768,3072,0.0069919998447100324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,768,3584,0.027719111906157598
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,768,3072,0.014031110538376702
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,768,2560,0.008389333056079017
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,768,3072,0.027080888549486797
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,768,2560,0.013665777941544851
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,768,2048,0.0074746666683091064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,768,2048,0.013362666798962487
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,768,2560,0.026707554856936138
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,768,1536,0.006039111150635614
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,768,1536,0.012777778009573618
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,768,1024,0.0046311111913787
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,768,1536,0.02440888848569658
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,768,2048,0.02514666650030348
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,768,1024,0.012545777691735161
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,768,768,0.004110222061475118
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,768,768,0.012101333174440594
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,768,512,0.011873777541849347
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,768,512,0.00369777778784434
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,768,1024,0.023519999451107446
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,768,768,0.023434667123688593
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,768,256,0.0031884445084465873
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,768,256,0.011438222395049201
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,768,512,0.022687999738587275
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,768,128,0.002903999967707528
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,768,128,0.011241777903503843
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,768,64,0.0026231110095977783
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,768,256,0.022385777698622808
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,768,64,0.011811555259757571
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,768,128,0.02233155568440755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,768,32,0.0026480000880029467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,768,32,0.010815111299355825
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,512,16384,0.009720889230569204
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,512,12288,0.008595555192894405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,512,16384,0.021586666504542034
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,512,16384,0.04282844397756788
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,512,65536,0.046720888879564076
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,512,12288,0.018405333161354065
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,512,10240,0.008488888541857401
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,512,12288,0.040580444865756564
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,512,10240,0.03676088982158237
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,512,10240,0.01777333352300856
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,512,8192,0.007948444121413762
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,512,8192,0.016491555505328707
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,512,7168,0.0075937774446275495
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,512,7168,0.016051555673281353
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,512,6144,0.006806222101052602
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,512,65536,0.09351999892128839
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,512,65536,0.022542221678627863
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,512,7168,0.032074666685528226
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,512,8192,0.0336426662074195
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,512,6144,0.015574221809705099
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,512,5120,0.007465778125656976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,512,6144,0.03106844425201416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,512,5120,0.015032000011867948
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,512,4096,0.0064071110553211635
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,512,4096,0.014395554860432943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,512,5120,0.029863112502627905
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,512,3584,0.0069724445541699724
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,512,4096,0.027775999572541978
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,512,3584,0.014360000689824423
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,512,3072,0.006568888823191325
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,512,3584,0.027419555518362258
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,512,3072,0.0140337778462304
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,512,2560,0.006182222316662471
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,512,3072,0.02697244452105628
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,512,2560,0.013637333280510373
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,512,2048,0.005508444375462002
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,512,2560,0.02586666742960612
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,512,2048,0.013364444176355997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,512,1536,0.004864888886610667
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,512,2048,0.024915556112925213
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,512,1536,0.012919999659061432
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,512,1024,0.0041413332025210066
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,512,1536,0.02438755498992072
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,512,1024,0.012551111479600271
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,512,768,0.0038213332494099936
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,512,1024,0.023174222972657945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,512,768,0.012082666986518435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,512,512,0.0034835553831524322
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,512,256,0.01145066652033064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,512,768,0.023149333066410486
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,512,512,0.011957333319716983
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,512,256,0.0031306667046414483
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,512,512,0.022896000080638464
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,512,256,0.02233244478702545
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,512,128,0.00286666676402092
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,512,128,0.011231111155615913
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,512,64,0.002708444371819496
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,512,128,0.02199733257293701
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,512,32,0.0027537778433826235
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,512,64,0.011691555380821228
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,512,32,0.010709332923094431
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,256,65536,0.04551911022928026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,256,16384,0.007867555651399825
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,256,16384,0.020403555697864957
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,256,16384,0.040639999839994646
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,256,65536,0.09245333406660293
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,256,12288,0.006950221955776215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,256,12288,0.018380444910791185
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,256,10240,0.007125332951545715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,256,65536,0.014504000544548035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,256,12288,0.03814311159981622
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,256,10240,0.01742755538887448
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,256,8192,0.006672888994216919
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,256,10240,0.03560977843072679
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,256,8192,0.01651822196112739
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,256,7168,0.00658044425977601
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,256,8192,0.03307377629809909
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,256,7168,0.01591822173860338
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,256,6144,0.030717333157857258
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,256,6144,0.0063386667105886675
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,256,7168,0.03225244416130914
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,256,6144,0.015440889530711703
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,256,5120,0.006879999819729064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,256,5120,0.0148053334818946
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,256,4096,0.0064399999876817065
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,256,5120,0.029538667864269678
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,256,4096,0.014494222071435718
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,256,3584,0.006823110911581252
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,256,4096,0.027585777971479628
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,256,3584,0.014399111270904541
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,256,3072,0.006476444502671559
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,256,3072,0.026455110973782007
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,256,3584,0.027427555786238775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,256,3072,0.013843555417325763
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,256,2560,0.006070222291681502
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,256,2048,0.0053680001033677
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,256,2560,0.013434666726324292
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,256,2560,0.02597066594494714
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,256,2048,0.013144888811641268
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,256,1536,0.0046880000995265115
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,256,2048,0.02462311089038849
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,256,1536,0.012644444074895648
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,256,1024,0.003995555556482739
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,256,1536,0.024509333901935156
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,256,1024,0.012446222205956778
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,256,768,0.003628444340493944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,256,1024,0.023758222659428913
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,256,768,0.011923555698659686
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,256,512,0.0033253334048721525
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,256,768,0.022838221655951604
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,256,512,0.011780444118711682
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,256,256,0.00295733329322603
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,256,512,0.023048889305856492
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,256,256,0.011369778050316704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,256,128,0.0027742222365405825
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,256,256,0.022847111026446026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,256,128,0.011189333266682096
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,256,64,0.0026453334010309647
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,256,128,0.021811554829279583
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,256,32,0.0026417778184016547
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,256,64,0.011505777637163797
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,256,32,0.010683555569913654
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,128,65536,0.011197333534558615
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,128,65536,0.04518489042917887
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,128,16384,0.04038755430115594
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,128,16384,0.006392000036107169
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,128,10240,0.006454222318198945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,128,12288,0.03737688726849026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,128,65536,0.09231999847624038
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,128,16384,0.01980888843536377
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,128,12288,0.006280888699822956
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,128,12288,0.01814933286772834
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,128,10240,0.017120889491505094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,128,8192,0.006337777607970768
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,128,10240,0.035397334231270686
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,128,8192,0.016013332539134555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,128,7168,0.006222222414281633
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,128,8192,0.033364444971084595
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,128,6144,0.006109333286682765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,128,7168,0.01568266749382019
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,128,6144,0.01533333294921451
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,128,7168,0.03227644496493869
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,128,5120,0.0064586665895250105
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,128,6144,0.03053155541419983
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,128,5120,0.01475822263293796
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,128,4096,0.006070222291681502
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,128,5120,0.029426667425367568
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,128,3072,0.006142222219043308
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,128,4096,0.014118222726715935
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,128,3584,0.006492444624503453
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,128,4096,0.028331554598278467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,128,3584,0.014002665877342224
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,128,3584,0.027660444378852844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,128,3072,0.0136808885468377
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,128,2560,0.006059555543793573
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,128,3072,0.02647999922434489
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,128,2560,0.013227555486891003
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,128,2048,0.005373333477311664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,128,2560,0.026351999905374315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,128,2048,0.012915555801656512
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,128,1536,0.0046657779150538975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,128,1536,0.023912888434198167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,128,2048,0.02498666610982683
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,128,1536,0.012466666599114736
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,128,1024,0.003972444269392225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,128,1024,0.012256888879670037
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,128,768,0.0036071112586392295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,128,1024,0.023772444989946153
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,128,768,0.012002666791280111
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,128,512,0.00328533330725299
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,128,768,0.023379556006855432
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,128,512,0.011746666497654386
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,128,256,0.0029404444826973807
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,128,512,0.02292622294690874
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,128,256,0.011318222516112857
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,128,128,0.0026906666656335196
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,128,256,0.022309333086013794
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,128,128,0.0111191107167138
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,128,64,0.0025439999169773525
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,128,64,0.01167022188504537
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,128,32,0.002556444456179937
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,128,32,0.010640888578361936
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,64,65536,0.009125333693292405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,64,65536,0.044024000565210976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,64,16384,0.006270222365856171
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,64,12288,0.006167999986145232
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,64,16384,0.019700444406933255
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,64,10240,0.0063884444534778595
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,64,12288,0.017723555366198223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,64,8192,0.015968888998031616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,64,10240,0.01682488951418135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,64,8192,0.006236444330877728
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,64,7168,0.006125333408514659
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,64,7168,0.01554666625128852
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,64,6144,0.005961777849329843
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,64,5120,0.0063768890168931745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,64,6144,0.015063999427689446
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,64,4096,0.005981333139869902
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,64,5120,0.014467555615637036
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,64,4096,0.0140346669488483
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,64,3584,0.0063697778516345555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,64,3072,0.006147555592987273
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,64,3584,0.013801777528391944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,64,3072,0.013428444663683573
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,64,2560,0.005649777750174205
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,64,2048,0.005027555757098728
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,64,2560,0.013134222063753339
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,64,2048,0.012708444562223224
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,64,1536,0.004422222160630756
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,64,1536,0.012329777909649743
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,64,1024,0.0037866665257347953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,64,768,0.003499555504984326
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,64,1024,0.012284444438086616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,64,512,0.003154666680428717
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,64,768,0.012019555601808759
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,64,512,0.011735999749766456
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,64,256,0.0028524444335036804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,64,128,0.002700444517864121
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,64,256,0.011228444675604502
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,64,64,0.002499555548032125
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,32,65536,0.00832622249921163
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,64,128,0.011194667054547204
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,64,32,0.0025253333151340485
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,64,64,0.011776000261306763
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,64,32,0.010499555203649731
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,32,16384,0.006239999913507038
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,32,65536,0.043466667334238686
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,32,12288,0.006171555568774541
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,32,16384,0.019848888119061787
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,32,12288,0.01792711185084449
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,32,10240,0.006287999865081575
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,32,8192,0.006189333481921091
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,32,10240,0.016823111308945548
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,32,7168,0.006058666855096817
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,32,8192,0.016001777516471017
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,32,6144,0.005969777703285217
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,32,7168,0.015662221444977652
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,32,6144,0.01518133282661438
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,32,5120,0.006270222365856171
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,32,5120,0.01457333399189843
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,32,4096,0.005927111125654644
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,32,4096,0.014152000347773233
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,32,3584,0.0064177778032090925
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,32,3072,0.006099555641412735
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,32,3584,0.01386666629049513
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,32,2560,0.00592533333433999
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,32,3072,0.01348000019788742
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,32,2048,0.0052764444715446895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,32,2560,0.01314933349688848
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,32,2048,0.012923555241690742
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,32,1536,0.004576000074545543
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,32,1024,0.0037662221325768363
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,32,1536,0.012431999875439538
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,32,1024,0.012095111111799875
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,32,768,0.003488000068399641
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,32,768,0.011959111524952782
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,32,512,0.0032035555276605818
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,32,256,0.002806222273243798
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,32,512,0.011582222249772815
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,32,128,0.0026373333401150177
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,32,256,0.011192888849311404
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,32,64,0.002493333278430833
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,32,128,0.01129244433508979
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,32,64,0.011420444481902652
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,8,32,32,0.002537777854336633
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,8,32,32,0.010418666733635796
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,65536,12288,0.19564089510175917
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,65536,12288,0.24227733082241484
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,65536,16384,0.2546302212609185
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,65536,10240,0.3265102174546984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,65536,10240,0.16728711128234863
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,65536,16384,0.3062586784362793
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,65536,10240,0.20906044377221
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,65536,8192,0.15324444240993923
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,65536,12288,0.3906995455423991
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,8,128,128,0.0222044438123703
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,65536,16384,0.5217857890658909
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,65536,8192,0.27472978168063694
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,65536,8192,0.16963999801211885
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,65536,7168,0.12482489479912652
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,65536,6144,0.10737955570220947
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,65536,5120,0.09242222044203018
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,65536,7168,0.15236622757381862
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,65536,6144,0.133459554778205
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,65536,4096,0.07761422130796644
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,65536,5120,0.16526933511098227
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,65536,5120,0.11896444691552056
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,65536,7168,0.22961600621541342
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,65536,4096,0.09805778000089858
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,65536,4096,0.13362577226426867
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,65536,3584,0.07017244233025445
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,65536,6144,0.19747022787729898
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,65536,3584,0.13210133711496988
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,65536,3584,0.09091999795701768
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,65536,3072,0.06776889165242513
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,65536,3072,0.08173333273993598
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,65536,2560,0.055584887663523354
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,65536,3072,0.10082044204076131
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,65536,2048,0.048862222168180675
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,65536,2560,0.07354666789372762
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,65536,1536,0.041759110159344144
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,65536,2048,0.06797333558400472
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,65536,2560,0.08417599731021458
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,65536,1536,0.052071998516718544
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,65536,2048,0.06212977568308512
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,65536,1536,0.05432088838683235
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,65536,768,0.033923556407292686
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,65536,1024,0.03571555680698819
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,65536,1024,0.043640888399548
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,65536,512,0.02979466650221083
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,65536,1024,0.03662577933735318
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,65536,768,0.027803556786643133
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,65536,256,0.026529777381155226
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,65536,128,0.0074577778577804565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,65536,256,0.031325333648257785
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,65536,768,0.04113066527578566
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,65536,128,0.025714667307005987
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,65536,64,0.005605333381228977
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,65536,512,0.03526844580968221
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,65536,128,0.029041777054468792
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,65536,256,0.011281777587201862
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,65536,64,0.02169422143035465
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,65536,32,0.004303110970391167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,65536,32,0.020853334003024634
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,65536,512,0.01975555552376641
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,16384,12288,0.061372445689307324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,16384,16384,0.07658400138219197
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,16384,16384,0.10824888944625854
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,16384,12288,0.09469688600964016
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,16384,10240,0.05415644579463535
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,16384,12288,0.08589066399468316
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,16384,10240,0.08016977707544963
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,16384,10240,0.07424266470803155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,16384,16384,0.12482755713992649
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,16384,8192,0.04680622286266751
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,16384,8192,0.06445955567889743
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,16384,8192,0.0642497804429796
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,16384,7168,0.04580000042915344
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,16384,7168,0.056906667020585805
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,16384,6144,0.038669334517584905
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,16384,7168,0.059097780121697314
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,16384,6144,0.04916888806555006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,16384,6144,0.054083555936813354
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,16384,5120,0.03487911158137851
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,16384,5120,0.041780445310804576
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,16384,65536,0.2577928966946072
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,16384,5120,0.05045955710940891
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,16384,4096,0.0311155551009708
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,16384,4096,0.04474133253097534
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,16384,3584,0.031077331966824
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,16384,4096,0.034527109728919134
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,16384,3584,0.030194666650560167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,16384,3072,0.027646222048335608
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,16384,3584,0.043622222211625844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,16384,3072,0.04085244403945075
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,16384,3072,0.026552889082166884
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,16384,2048,0.023175999522209167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,16384,2560,0.022830223043759663
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,16384,2560,0.02561155623859829
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,16384,2048,0.035064889325035944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,16384,1536,0.02077155477470822
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,16384,2560,0.037853333685133196
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,16384,2048,0.018831110662884183
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,16384,1536,0.03222577770551046
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,16384,1024,0.018371555540296767
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,16384,768,0.00904799997806549
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,16384,1024,0.029520889123280842
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,16384,1024,0.011053333679835001
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,16384,1536,0.01516533394654592
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,16384,768,0.017646221650971305
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,16384,512,0.007084444165229797
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,16384,512,0.01626844373014238
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,16384,768,0.029242667886945937
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,16384,256,0.005186666631036335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,16384,512,0.02740977704524994
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,16384,256,0.013255111045307584
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,16384,128,0.0034337778472238113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,16384,256,0.023396443989541795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,16384,128,0.01388177772363027
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,16384,64,0.003063999944263034
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,16384,128,0.023545778459972803
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,16384,32,0.003136000078585413
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,16384,64,0.012852444416946836
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,16384,65536,0.30786310301886666
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,16384,32,0.012182222472296821
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,12288,16384,0.06272444460127088
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,12288,12288,0.08069777488708496
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,12288,16384,0.10299022330178155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,12288,16384,0.10452889071570502
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,12288,12288,0.05099466774198744
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,12288,12288,0.0837822225358751
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,12288,10240,0.04548089040650261
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,12288,10240,0.06711555851830377
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,12288,8192,0.039510223600599505
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,12288,10240,0.0737431115574307
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,12288,8192,0.05425777700212267
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,16384,65536,0.542468441857232
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,12288,8192,0.06418310933642916
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,12288,7168,0.03628177775277032
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,12288,65536,0.22255200809902617
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,12288,7168,0.04762044548988342
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,12288,7168,0.06122577852673001
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,12288,6144,0.0355679988861084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,12288,6144,0.04116800096299913
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,12288,5120,0.030576000610987347
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,12288,6144,0.05587377813127306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,12288,5120,0.03826399975352817
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,12288,5120,0.051720887422561646
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,12288,4096,0.029065777858098347
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,12288,4096,0.02824444572130839
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,12288,4096,0.04534844557444254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,12288,3584,0.027059555053710938
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,12288,65536,0.30212799708048504
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,12288,3584,0.04324266645643446
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,12288,2560,0.023385778069496155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,12288,3584,0.025249777568711176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,12288,3072,0.024587555064095393
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,12288,2560,0.03750755720668369
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,12288,3072,0.0402817792362637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,12288,2048,0.021122665868865118
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,12288,1536,0.012837332983811697
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,12288,2048,0.033923556407292686
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,12288,2560,0.01904622217019399
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,12288,1536,0.019810666640599568
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,12288,1024,0.009494221872753566
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,12288,3072,0.022058667408095464
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,12288,2048,0.01591377788119846
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,12288,1024,0.017860444055663217
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,12288,1536,0.032413333654403687
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,12288,1024,0.028888887829250757
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,12288,768,0.007798222204049428
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,12288,768,0.016573333077960543
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,12288,512,0.00620266670982043
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,12288,768,0.02853155467245314
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,12288,512,0.015062222878138224
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,12288,256,0.003834666477309333
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,12288,512,0.026668444275856018
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,12288,256,0.013026666310098437
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,12288,128,0.003313777761326896
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,12288,256,0.023200889428456623
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,12288,128,0.013705777625242868
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,12288,64,0.0030293334275484085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,12288,32,0.011867555479208628
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,12288,128,0.023216888308525085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,12288,32,0.0030008889734745026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,12288,64,0.012464889221721225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,10240,16384,0.05563377671771579
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,10240,16384,0.0838551123936971
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,10240,12288,0.06890844636493258
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,10240,16384,0.09059022532569037
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,10240,12288,0.046375112401114575
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,10240,10240,0.04176355401674906
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,10240,12288,0.0706897775332133
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,10240,10240,0.05860977702670627
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,10240,10240,0.0611706640985277
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,10240,8192,0.0365591115421719
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,10240,8192,0.047982222504085965
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,10240,65536,0.17526043785942924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,10240,8192,0.05466577741834852
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,10240,7168,0.03842755489879184
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,10240,7168,0.03377244538731045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,10240,7168,0.051474667257732816
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,10240,6144,0.031139555904600356
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,10240,6144,0.036871112055248685
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,10240,5120,0.02880711025661892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,10240,65536,0.23856798807779947
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,10240,6144,0.046930667426851064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,10240,5120,0.04474222328927782
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,12288,65536,0.4000115659501817
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,10240,4096,0.026035555534892615
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,10240,4096,0.023579556081030104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,10240,3584,0.024523556232452393
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,10240,5120,0.028677334388097126
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,10240,4096,0.040207111173205905
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,10240,3072,0.022892443670166865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,10240,3584,0.020917332834667627
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,10240,3584,0.04107200105985006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,10240,2560,0.02183377742767334
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,10240,3072,0.03499555587768555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,10240,2560,0.033913777934180364
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,10240,2048,0.020456888609462313
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,10240,1536,0.010751999914646149
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,10240,2048,0.03172355559137132
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,10240,1536,0.018744000130229525
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,10240,3072,0.018395556343926322
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,10240,1024,0.008170666793982187
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,10240,1536,0.030231999026404485
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,10240,1024,0.017502221796247695
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,10240,2048,0.013199110825856527
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,10240,2560,0.016057777735922072
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,10240,512,0.005659555395444234
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,10240,768,0.0069448889957533936
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,10240,1024,0.028919110695521038
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,10240,768,0.01532799998919169
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,10240,768,0.028424888849258423
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,10240,512,0.013819555441538492
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,10240,256,0.0034719999465677473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,10240,256,0.013005333642164866
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,10240,512,0.02423644396993849
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,10240,128,0.0032391111469931076
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,10240,256,0.02381422287887997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,10240,128,0.013624000052611033
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,10240,64,0.0029315555261241067
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,10240,128,0.02421955598725213
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,10240,64,0.012455110748608908
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,10240,32,0.002934222213096089
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,10240,32,0.011671110987663269
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,8192,16384,0.04633333285649618
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,8192,16384,0.07821422153049044
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,8192,12288,0.052002665069368154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,8192,12288,0.038431111309263445
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,8192,16384,0.06715466578801473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,10240,65536,0.3648844559987386
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,8192,10240,0.03468088971243964
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,8192,12288,0.06385778056250678
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,8192,10240,0.057927111784617104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,8192,10240,0.042103999190860324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,8192,8192,0.031559109687805176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,8192,8192,0.034204443295796715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,8192,7168,0.028663999504513208
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,8192,8192,0.049666666322284274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,8192,7168,0.04751111070315043
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,8192,7168,0.030245333909988403
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,8192,6144,0.02659911082850562
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,8192,6144,0.026686223016844854
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,8192,6144,0.04483200112978617
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,8192,5120,0.02429955535464817
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,8192,5120,0.04290666513972812
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,8192,5120,0.023054222265879314
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,8192,4096,0.019162666466501024
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,8192,3584,0.02027555637889438
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,8192,4096,0.02125599980354309
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,8192,4096,0.03750222259097629
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,8192,3584,0.03821688890457153
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,8192,3072,0.019156444403860304
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,8192,3072,0.03421155611673991
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,8192,65536,0.2145528925789727
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,8192,65536,0.13853332731458876
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,8192,3584,0.01703644461101956
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,8192,2560,0.013184888495339287
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,8192,2560,0.018443556295500863
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,8192,3072,0.01512977812025282
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,8192,2560,0.032864888509114586
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,8192,2048,0.017210667332013447
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,8192,1536,0.0092275556590822
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,8192,2048,0.030979557169808283
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,8192,1536,0.016149333781666227
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,8192,1024,0.007180444068378872
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,8192,2048,0.011191999746693505
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,8192,1024,0.015023110641373528
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,8192,1536,0.03031022349993388
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,8192,768,0.0063742221229606206
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,8192,1024,0.027535999814669292
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,8192,768,0.02645066711637709
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,8192,768,0.013435555828942193
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,8192,512,0.00508266646001074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,8192,512,0.012246222131782107
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,8192,256,0.011528888510333167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,8192,256,0.0035617777870761026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,8192,512,0.02379111117786831
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,8192,128,0.003120888852410846
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,8192,256,0.022808889547983806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,8192,128,0.011918221910794577
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,8192,64,0.0029048888633648553
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,8192,128,0.022829333941141765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,8192,32,0.00300799993177255
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,8192,64,0.01127644462717904
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,8192,32,0.01036444471942054
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,7168,16384,0.04303288790914747
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,7168,16384,0.07452444235483806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,8192,65536,0.2637271086374919
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,7168,12288,0.036405334870020546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,7168,12288,0.05926666657129923
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,7168,16384,0.06039466460545858
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,7168,12288,0.04981688989533318
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,7168,10240,0.032687111033333674
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,7168,8192,0.02904799911710951
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,7168,10240,0.05562577644983927
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,7168,10240,0.044749332798851855
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,7168,8192,0.033906665113237165
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,7168,7168,0.027129777603679236
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,7168,65536,0.12637333075205484
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,7168,8192,0.049703111251195274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,7168,7168,0.04636889033847385
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,7168,7168,0.0316826668050554
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,7168,6144,0.02509333358870612
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,7168,5120,0.022670222653283015
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,7168,6144,0.04424266682730781
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,7168,4096,0.02033066584004296
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,7168,5120,0.041611555549833514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,7168,4096,0.03566666775279575
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,7168,5120,0.023266666465335425
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,7168,6144,0.026548445224761963
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,7168,3584,0.019574221637513902
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,7168,65536,0.19478310479058158
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,7168,4096,0.019934222102165222
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,7168,3584,0.035030222601360746
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,7168,3072,0.01885066595342424
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,7168,2560,0.017672888106769986
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,7168,3072,0.0328400002585517
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,7168,2048,0.010401777923107147
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,7168,3584,0.01795733306143019
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,7168,2560,0.0318213336997562
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,7168,2048,0.016752888758977253
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,7168,2560,0.012084444363911947
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,7168,1536,0.00869511150651508
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,7168,1536,0.015781333049138386
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,7168,3072,0.016153777639071148
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,7168,2048,0.030460443761613634
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,7168,1024,0.006794666664467917
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,7168,1536,0.02960977620548672
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,7168,1024,0.013667555318938361
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,7168,768,0.00600622221827507
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,7168,1024,0.02787733409139845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,7168,768,0.012578667038016848
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,7168,512,0.004187555362780889
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,7168,768,0.02442933287885454
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,7168,256,0.0033484444850020935
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,7168,128,0.0030533334033356774
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,7168,512,0.012100444071822695
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,7168,512,0.023396443989541795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,7168,256,0.011669333610269757
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,7168,256,0.022690667046440974
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,7168,128,0.011970666547616323
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,7168,64,0.0028008888992998335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,7168,128,0.023856888214747112
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,7168,32,0.0028604444944196274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,7168,64,0.011025778121418424
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,7168,32,0.010656000011497073
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,6144,16384,0.03993955585691664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,6144,16384,0.07377688752280341
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,7168,65536,0.23206400871276855
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,6144,12288,0.03348799877696567
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,6144,16384,0.05557777484258016
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,6144,12288,0.03918933206134372
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,6144,12288,0.059860441419813365
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,6144,10240,0.030751999881532457
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,6144,10240,0.05496711201137967
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,6144,8192,0.02773777809407976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,6144,10240,0.03737066520584954
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,6144,8192,0.047648890150917896
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,6144,8192,0.02774222195148468
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,6144,7168,0.026205332742797002
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,6144,7168,0.04548355605867174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,6144,7168,0.024432889289326135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,6144,6144,0.04262755645645989
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,6144,5120,0.02142133315404256
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,6144,65536,0.11736267142825657
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,6144,6144,0.024243555135197107
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,6144,5120,0.03977244430118137
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,6144,6144,0.023009777069091797
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,6144,4096,0.019874667127927143
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,6144,5120,0.018664000762833487
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,6144,4096,0.03501333461867438
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,6144,3584,0.019114666514926486
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,6144,3072,0.012518222133318583
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,6144,4096,0.01644444465637207
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,6144,3072,0.01809599995613098
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,6144,3584,0.036489778094821505
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,6144,3584,0.01496444476975335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,6144,3072,0.03335466649797227
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,6144,2560,0.011696000066068439
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,6144,2560,0.017344888713624742
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,6144,2048,0.009970666633711921
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,6144,2560,0.0317884451813168
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,6144,2048,0.01647200021478865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,6144,65536,0.19250578350490996
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,6144,2048,0.0307644440068139
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,6144,1536,0.008099555969238281
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,6144,1536,0.015489778584904142
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,6144,1024,0.006392000036107169
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,6144,768,0.005568000177542369
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,6144,1536,0.02991466720898946
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,6144,1024,0.012835555606418185
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,6144,1024,0.02717422280046675
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,6144,768,0.01236355553070704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,6144,512,0.01200355589389801
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,6144,512,0.003839111162556542
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,6144,768,0.024199111594094172
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,6144,256,0.0033271111961868075
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,6144,512,0.02369155486424764
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,6144,256,0.011431111229790581
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,6144,128,0.00304888892504904
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,6144,256,0.023024888502226934
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,6144,128,0.023039110832744177
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,6144,128,0.01184177812602785
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,6144,64,0.0028106667515304354
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,6144,64,0.010891555911964841
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,6144,32,0.0028400001012616684
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,6144,32,0.010280888941552904
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,5120,16384,0.036109334892696805
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,5120,16384,0.07167466481526692
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,5120,16384,0.04859377940495809
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,5120,12288,0.03760622276200189
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,5120,12288,0.030755556292004053
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,5120,12288,0.05908089213901096
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,5120,10240,0.02840977907180786
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,5120,10240,0.03218222326702542
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,5120,10240,0.05272533496220907
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,5120,8192,0.02693066663212246
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,5120,8192,0.025647999511824712
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,5120,65536,0.10049333175023396
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,5120,8192,0.04562133219507006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,5120,7168,0.024051555328898962
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,5120,6144,0.02125511070092519
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,5120,7168,0.044646223386128746
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,6144,65536,0.19697954919603136
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,5120,6144,0.04088266690572103
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,5120,5120,0.020299555526839364
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,5120,6144,0.018248889181349013
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,5120,5120,0.03951999876234267
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,5120,7168,0.024175999893082514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,5120,4096,0.018651554981867473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,5120,4096,0.03413777881198459
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,5120,5120,0.015751111838552687
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,5120,65536,0.17531111505296496
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,5120,3584,0.01814133259985182
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,5120,4096,0.013539555172125498
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,5120,3584,0.034833778937657676
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,5120,3072,0.010794666906197866
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,5120,3072,0.017267554998397827
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,5120,3072,0.03254133462905884
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,5120,2560,0.009646221995353699
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,5120,2560,0.016732444365819294
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,5120,3584,0.012177777787049612
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,5120,2048,0.008416888614495596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,5120,2560,0.031756444109810725
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,5120,2048,0.015839111473825242
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,5120,1536,0.007347555624114142
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,5120,2048,0.029913776450686987
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,5120,1536,0.014348444011476306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,5120,1024,0.0058871110280354815
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,5120,1536,0.02929688824547662
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,5120,1024,0.012706666356987424
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,5120,768,0.023994666006830003
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,5120,768,0.005111111121045219
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,5120,1024,0.024326221810446844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,5120,768,0.01223288890388277
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,5120,512,0.0036888888312710654
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,5120,512,0.011895111037625207
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,5120,256,0.0032222221295038858
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,5120,512,0.02401422295305464
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,5120,256,0.01149511088927587
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,5120,128,0.002975111206372579
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,5120,128,0.023078221413824294
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,5120,256,0.02242311173015171
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,5120,128,0.011991110940774282
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,5120,64,0.0027111110587914786
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,5120,64,0.011429333024554782
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,5120,32,0.00282222218811512
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,5120,32,0.010126222338941362
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,5120,65536,0.18748622470431855
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,4096,16384,0.03189155459403992
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,4096,16384,0.07450577947828504
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,4096,16384,0.03977155685424805
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,4096,12288,0.027235555979940627
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,4096,12288,0.05791911151674059
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,4096,12288,0.03072533342573378
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,4096,10240,0.027395556370417278
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,4096,10240,0.025277333127127752
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,4096,10240,0.049525333775414355
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,4096,65536,0.08003377914428711
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,4096,8192,0.023221333821614582
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,4096,8192,0.0222080002228419
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,4096,7168,0.020863110820452373
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,4096,7168,0.04432266619470385
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,4096,6144,0.020120888948440552
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,4096,8192,0.04535466763708326
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,4096,6144,0.0391937792301178
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,4096,7168,0.020153777466879953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,4096,5120,0.018460444278187223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,4096,6144,0.017490666773584154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,4096,65536,0.18001511361863878
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,4096,5120,0.03772799836264716
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,4096,4096,0.017794667018784415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,4096,5120,0.015557333827018738
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,4096,3584,0.012453333371215396
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,4096,4096,0.013607111242082385
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,4096,4096,0.034599112139807806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,4096,3584,0.01707555519209968
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,4096,3072,0.011405333048767514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,4096,3584,0.03417422374089559
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,4096,3072,0.016544888416926067
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,4096,2560,0.008450667063395182
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,4096,3072,0.032518221272362605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,4096,2560,0.0158915552828047
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,4096,2048,0.007463110817803278
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,4096,2048,0.015322667029168872
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,4096,2560,0.030841777722040813
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,4096,1536,0.006514666808976068
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,4096,1536,0.013502221968438892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,4096,2048,0.028875556257035997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,4096,1024,0.00544888898730278
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,4096,1536,0.02792000108295017
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,4096,1024,0.012495111260149213
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,4096,768,0.004328000048796336
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,4096,1024,0.024420445164044697
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,4096,768,0.012159110771285163
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,4096,512,0.003522666792074839
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,4096,512,0.023961777488390606
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,4096,768,0.023822221491071913
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,4096,512,0.011872000164455838
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,4096,256,0.0031662223239739737
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,4096,256,0.011352889239788055
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,4096,128,0.002875555513633622
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,4096,256,0.022664000590642292
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,4096,128,0.012030222349696688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,4096,64,0.0027057776848475137
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,4096,128,0.02313599983851115
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,4096,64,0.010687111152542962
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,4096,32,0.002752888947725296
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,4096,32,0.010101333260536194
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3584,16384,0.029529776838090684
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,3584,16384,0.06865511337916057
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3584,12288,0.026383999321195815
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3584,16384,0.03344977895418803
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3584,12288,0.025814221964942083
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,4096,65536,0.13293777571784124
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3584,10240,0.023706666297382776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,3584,12288,0.05528533458709717
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3584,8192,0.02139288849300808
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,3584,8192,0.043904887305365674
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,3584,10240,0.04914311236805386
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3584,10240,0.023032888770103455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3584,7168,0.020512888828913372
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3584,8192,0.019495111372735765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,3584,7168,0.04184088773197598
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3584,7168,0.017857778403494094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3584,6144,0.019152000546455383
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3584,6144,0.015848888291252982
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3584,5120,0.018039999736679923
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,3584,6144,0.04083911246723599
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,3584,5120,0.03753955496682061
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3584,4096,0.012607110871209038
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3584,4096,0.017472889688279893
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3584,5120,0.014518222875065274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3584,3584,0.011753777662913004
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3584,65536,0.07195022371080187
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,3584,65536,0.16292711098988852
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3584,3072,0.010855110983053843
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,3584,4096,0.03317244516478644
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3584,3584,0.016698666744761996
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,3584,3584,0.03294577863481309
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3584,3072,0.016235555211702984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3584,2560,0.008071111308203803
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,3584,3072,0.03239377670817905
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3584,2560,0.015544889701737298
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3584,2048,0.007061333292060428
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,3584,2560,0.030581331915325586
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3584,2048,0.015112888481881885
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3584,1536,0.006245333287451003
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,3584,2048,0.028520888752407495
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3584,1536,0.012905778156386482
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3584,1024,0.005351999981535806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,3584,1536,0.02548711167441474
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3584,1024,0.012364444633324942
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3584,768,0.0042462220622433555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,3584,1024,0.02425066630045573
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3584,768,0.012149333126015134
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3584,512,0.003495111233658261
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,3584,768,0.023752000596788194
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3584,512,0.011714666253990598
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3584,256,0.0031439999325407874
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,3584,512,0.02325155503220028
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3584,256,0.011419555379284753
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3584,128,0.0028417776856157514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,3584,256,0.02257333364751604
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3584,128,0.011564444336626263
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3584,64,0.0027075554761621687
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,3584,128,0.023659555448426142
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3584,64,0.011182222101423474
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3584,32,0.002752888947725296
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3584,32,0.01001422190003925
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3584,65536,0.12542043791876897
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3072,16384,0.02809422214825948
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,3072,16384,0.06727821959389581
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3072,12288,0.023732443650563557
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3072,16384,0.031494223409228854
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3072,12288,0.024816888901922438
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,3072,12288,0.05527555611398485
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3072,10240,0.022244445151752893
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3072,10240,0.020970667401949566
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3072,8192,0.020424000091022916
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,3072,10240,0.049478222926457725
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,3072,8192,0.042945779032177396
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3072,7168,0.019711999429596793
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,3072,7168,0.04182933436499702
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3072,65536,0.06675910949707031
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3072,8192,0.01811288959450192
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3072,6144,0.018754666050275166
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3072,5120,0.013085333009560903
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3072,7168,0.016315556234783597
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,3072,6144,0.038386666112475924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3072,5120,0.01755466726091173
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,3072,5120,0.03672088848219977
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3072,4096,0.012039110892348819
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3072,6144,0.015223110715548197
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3072,4096,0.01721955504682329
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,3072,4096,0.03281689021322463
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3072,3584,0.011063999599880643
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3072,3584,0.016195555528004963
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3072,3072,0.010227555202113258
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,3072,3584,0.033913777934180364
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,3072,65536,0.15175378322601318
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3072,3072,0.016040000650617812
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3072,2560,0.007707555260923173
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,3072,3072,0.03163199954562717
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3072,2048,0.006768888897365994
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3072,2560,0.01532444523440467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,3072,2560,0.0306311117278205
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3072,2048,0.014316444595654806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3072,1536,0.006096000058783426
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,3072,2048,0.028575999869240656
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3072,1536,0.012903999951150684
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3072,1024,0.004791111167934206
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,3072,1536,0.026147555973794725
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3072,1024,0.012497777740160624
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3072,768,0.0038177776667806837
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,3072,1024,0.02421955598725213
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3072,768,0.012131555212868584
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3072,512,0.0035253332720862497
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,3072,768,0.023694222172101338
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3072,512,0.011783999701340994
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3072,256,0.003145777723855443
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,3072,512,0.023782221807373896
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3072,256,0.011490667031870948
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3072,128,0.0028577778074476454
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,3072,128,0.022798221972253587
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,3072,256,0.02313333253065745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3072,128,0.011680888632933298
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3072,64,0.0027191111197074256
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3072,32,0.002671111168132888
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3072,64,0.011187555889288584
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,3072,32,0.010008888940016428
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,3072,65536,0.10399644242392646
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2560,12288,0.02369511127471924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2560,16384,0.02749600013097127
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,2560,12288,0.05307733350329929
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2560,12288,0.021154666940371197
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,2560,16384,0.06398666567272611
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2560,16384,0.026745777991082933
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2560,10240,0.021480000681347314
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,2560,10240,0.04768888817893135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2560,8192,0.01938311093383365
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2560,10240,0.01841600073708428
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,2560,8192,0.044362667534086436
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2560,7168,0.018387556076049805
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2560,6144,0.013024888932704926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2560,8192,0.015980444020695157
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,2560,7168,0.04283644424544441
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2560,7168,0.014615999327765571
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2560,6144,0.018417777286635507
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2560,65536,0.06456266509162055
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2560,5120,0.011676444775528379
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,2560,6144,0.03869688842031691
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2560,5120,0.017887110511461895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,2560,5120,0.03715199894375271
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2560,4096,0.010577778021494547
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2560,4096,0.01683466633160909
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2560,3584,0.00996977753109402
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,2560,4096,0.03313422203063965
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2560,3584,0.01592266725169288
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2560,3072,0.009323555562231276
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2560,3072,0.015360888507631091
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,2560,3584,0.03349688980314467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2560,2560,0.007297777467303806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,2560,3072,0.03219111098183526
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2560,2560,0.014153778553009033
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2560,2048,0.006654222392373615
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,2560,2560,0.02997244397799174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2560,2048,0.013363555901580386
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2560,1536,0.005640888793600931
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,2560,2048,0.02682400080892775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2560,1536,0.012930666406949362
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2560,1024,0.004104888687531154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,2560,65536,0.14412711726294622
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,2560,1536,0.02509244448608822
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2560,1024,0.012431110772821637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2560,768,0.0037768888804647657
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,2560,1024,0.024660444921917383
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2560,768,0.012016889121797351
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2560,512,0.0034551111360390982
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,2560,768,0.0236488895283805
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2560,512,0.011756444142924415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2560,256,0.003101333354910215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,2560,512,0.02366311185889774
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2560,256,0.01128000020980835
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2560,65536,0.09454933140012954
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2560,128,0.0028613333900769553
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,2560,256,0.022441777918073866
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2560,128,0.011715555356608497
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2560,64,0.002667555585503578
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,2560,128,0.023552889625231426
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2560,32,0.0026764445420768526
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2560,64,0.010496888723638324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2560,32,0.009942221972677443
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2048,12288,0.022337777747048274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,2048,16384,0.06342755423651801
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2048,16384,0.025664889150195654
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,2048,12288,0.05303644471698337
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2048,12288,0.018179555733998615
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2048,16384,0.023009777069091797
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2048,10240,0.02085688875781165
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,2048,10240,0.047945777575174965
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2048,10240,0.016525333126386006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2048,8192,0.019401777121755812
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,2048,8192,0.04287822379006279
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2048,7168,0.012724444270133972
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2048,8192,0.013727111121018728
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2048,7168,0.018831999765502084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2048,6144,0.011744000017642975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,2048,7168,0.04116800096299913
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2048,6144,0.018029333816634286
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2048,5120,0.010834666589895884
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2048,5120,0.017260443833139207
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2048,65536,0.05956800116433037
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,2048,6144,0.03801777627733018
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2048,4096,0.00978311151266098
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,2048,5120,0.037265777587890625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2048,4096,0.01616533266173469
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,2048,4096,0.033424887392256
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2048,3584,0.00903911143541336
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2048,3584,0.015417777829700045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2048,3072,0.008592888712882996
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,2048,65536,0.15194043848249647
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,2048,3584,0.03404533200793796
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2048,3072,0.014179555906189812
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2048,2560,0.007184888753626082
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,2048,3072,0.03007733490731981
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2048,2560,0.013254221942689685
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2048,2048,0.006426666759782367
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,2048,2560,0.02835822105407715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2048,2048,0.01332622187005149
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,2048,2048,0.02652799917591943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2048,1536,0.0047528888616296975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2048,1536,0.012715555727481842
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2048,1024,0.0041617775956789655
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,2048,1536,0.02613777750068241
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2048,1024,0.0123831108212471
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2048,768,0.00370399985048506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,2048,1024,0.02422488894727495
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2048,768,0.011969777444998423
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2048,512,0.003354666754603386
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,2048,768,0.024522667129834492
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2048,512,0.011606222225560082
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2048,256,0.003031999907559819
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,2048,512,0.02369155486424764
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2048,256,0.011230222052998014
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2048,128,0.002790222151411904
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,2048,256,0.02254044512907664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2048,128,0.011936888926559024
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2048,64,0.002647111192345619
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,2048,128,0.022593778040673997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2048,32,0.002647111192345619
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2048,64,0.010619555910428366
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,2048,32,0.009941333697901832
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,2048,65536,0.06979288657506307
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1536,16384,0.02401155564520094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,1536,16384,0.06227911180920071
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1536,12288,0.021653332644038733
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1536,10240,0.013415999710559845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,1536,12288,0.04926844437917074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1536,16384,0.01849244369400872
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1536,12288,0.015154666370815702
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1536,10240,0.02029511166943444
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1536,8192,0.011847111086050669
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,1536,10240,0.04607999987072415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1536,8192,0.019077334139082167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,1536,8192,0.04292622208595276
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1536,7168,0.01094133324093289
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1536,7168,0.017847999930381775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1536,6144,0.010135999984211391
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1536,65536,0.05606666538450453
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,1536,7168,0.04046222236421373
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1536,6144,0.01755466726091173
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1536,5120,0.009500444763236577
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,1536,6144,0.037061333656311035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1536,5120,0.01676266723208957
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1536,4096,0.008632889224423302
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,1536,5120,0.03562666641341315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,1536,65536,0.14777689509921604
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1536,4096,0.015224888920783997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1536,3584,0.014303111367755465
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1536,3584,0.008488000267081791
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,1536,4096,0.032949331733915545
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1536,3072,0.00817599975400501
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,1536,3584,0.031870222753948636
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1536,3072,0.013918222652541267
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1536,2560,0.008754666480753157
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,1536,3072,0.028431998358832464
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1536,2560,0.013561777770519257
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1536,2048,0.008292444050312042
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1536,1536,0.006018666757477655
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,1536,2560,0.027003554834259882
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1536,2048,0.013112000293201871
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,1536,2048,0.025240888198216755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1536,1536,0.012541333006487953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1536,1024,0.005148444324731827
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,1536,1536,0.025775111383861963
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1536,1024,0.01219288922018475
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1536,768,0.004389333228270213
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,1536,768,0.023131555981106226
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,1536,1024,0.023835556374655828
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1536,768,0.01182933317290412
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1536,512,0.0038968887594011095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1536,512,0.011632888681358762
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1536,256,0.0032879999942249725
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1536,65536,0.05451644460360209
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,1536,512,0.022642667094866436
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1536,256,0.011253332926167382
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1536,128,0.002985777747299936
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,1536,256,0.02254044512907664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1536,128,0.011587555209795633
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1536,64,0.0027911110470692315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,1536,128,0.022780444886949327
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1536,32,0.00282577777074443
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1536,64,0.010767111347781287
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1536,32,0.009913778139485253
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1024,16384,0.01442577772670322
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1024,16384,0.022135110365019903
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1024,12288,0.01203733351495531
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1024,12288,0.02016177773475647
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,1024,16384,0.05319644345177544
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1024,10240,0.011226666470368704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,1024,12288,0.04660444458325704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1024,10240,0.019168888529141743
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1024,8192,0.009908444351620143
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,1024,10240,0.042796442906061806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1024,8192,0.018100443813535903
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1024,7168,0.00925422211488088
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1024,65536,0.047113776206970215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1024,7168,0.017245334055688646
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,1024,8192,0.040476444694730974
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1024,6144,0.008415111237102086
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,1024,7168,0.038711110750834145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1024,6144,0.01609688831700219
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,1024,6144,0.03542666633923849
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1024,5120,0.00850311087237464
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,1024,65536,0.11287200450897217
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1024,5120,0.014927110738224454
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,1024,4096,0.02870755394299825
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1024,4096,0.007696888513035244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,1024,5120,0.033857779370413885
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1024,4096,0.01443288889196184
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1024,3584,0.007628444168302748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1024,3584,0.01420177850458357
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1024,2560,0.01421688828203413
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1024,3072,0.007338666253619724
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1024,2560,0.01329955541425281
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1024,65536,0.0396088891559177
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,1024,3584,0.028735111157099407
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1024,3072,0.01386222243309021
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,1024,3072,0.02755733331044515
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,1024,2560,0.026289777623282537
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1024,2048,0.008368000388145447
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1024,2048,0.013056889176368713
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1024,1024,0.007751999629868402
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1024,1536,0.0069413334131240845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,1024,2048,0.024570667081409033
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1024,1536,0.012676444318559436
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,1024,1536,0.024988444315062627
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1024,1024,0.012279999752839407
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1024,768,0.005592000153329637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,1024,1024,0.023583110835817125
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1024,768,0.011943110989199745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,1024,768,0.023021333747439917
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1024,512,0.00408266650305854
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1024,512,0.011666666302416058
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,1024,512,0.022846221923828125
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1024,256,0.003995555556482739
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1024,256,0.011197333534558615
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1024,128,0.003400000019205941
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,1024,256,0.022423999177085027
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1024,128,0.011643555429246692
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1024,64,0.00304888892504904
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,1024,128,0.02293777796957228
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,1024,32,0.003080000066094928
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1024,64,0.010790222220950656
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,1024,32,0.010088000032636855
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,768,16384,0.01183555523554484
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,768,12288,0.010092444717884064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,768,16384,0.021615111165576514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,768,12288,0.019934222102165222
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,768,16384,0.04811999864048428
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,768,10240,0.009533333281675974
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,768,12288,0.04665689004792107
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,768,10240,0.01815111107296414
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,768,8192,0.009197333620654212
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,768,65536,0.04565422071350945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,768,10240,0.04025600022739834
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,768,8192,0.01741422216097514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,768,7168,0.008405333591832055
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,768,7168,0.016388444436921012
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,768,8192,0.03740444448259141
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,768,65536,0.10623111327489217
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,768,6144,0.007872888611422645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,768,7168,0.035327110025617815
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,768,6144,0.015207999282413058
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,768,5120,0.008072888685597314
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,768,6144,0.0314648879898919
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,768,4096,0.007074666519959767
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,768,65536,0.02978044417169359
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,768,4096,0.027815110153622095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,768,3584,0.014107555150985718
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,768,5120,0.01482399966981676
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,768,5120,0.030581331915325586
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,768,4096,0.014301333162519666
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,768,3584,0.007309333317809635
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,768,3072,0.006710222197903528
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,768,3584,0.028914666838116113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,768,3072,0.013734222286277346
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,768,2560,0.007884444461928474
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,768,3072,0.026706667410002813
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,768,2560,0.013506666653686099
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,768,2048,0.007650666766696506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,768,2560,0.025546666648652818
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,768,2048,0.013186666700575085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,768,1536,0.006305777778228124
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,768,2048,0.024439111351966858
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,768,1536,0.012626666989591388
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,768,1024,0.004920888692140579
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,768,1536,0.023801777097913954
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,768,1024,0.012331555287043253
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,768,768,0.004211555752489301
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,768,1024,0.023196443915367126
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,768,768,0.011936888926559024
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,768,512,0.00370399985048506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,768,768,0.022795556320084467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,768,512,0.01166488892502255
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,768,256,0.0032426667296224167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,768,512,0.022685334086418152
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,768,256,0.011253332926167382
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,768,128,0.002965333354141977
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,768,256,0.022222222553359136
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,768,128,0.011601777540312873
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,768,64,0.0028328889360030494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,768,128,0.022696000006463792
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,768,64,0.010786666638321347
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,768,32,0.002777777819169892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,768,32,0.009908444351620143
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,512,16384,0.021440888444582622
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,512,16384,0.00979288915793101
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,512,12288,0.008664888640244802
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,512,16384,0.044024000565210976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,512,12288,0.019592000378502738
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,512,65536,0.04445777667893303
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,512,10240,0.008436444732877944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,512,12288,0.04148088892300924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,512,10240,0.017512000269360013
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,512,8192,0.008083555433485243
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,512,65536,0.09343377749125163
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,512,10240,0.0354968872335222
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,512,65536,0.022081777453422546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,512,8192,0.0162764439980189
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,512,7168,0.007703999678293864
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,512,8192,0.03197600113021003
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,512,7168,0.015950222810109455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,512,6144,0.007359111474619971
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,512,7168,0.030618667602539062
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,512,6144,0.015268445014953613
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,512,4096,0.006512889017661412
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,512,5120,0.007188444336255391
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,512,6144,0.029834667841593426
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,512,5120,0.014892444014549255
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,512,5120,0.02832533253563775
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,512,4096,0.014316444595654806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,512,3584,0.006913777854707506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,512,4096,0.02696177860101064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,512,3584,0.014055111342006259
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,512,3072,0.006524444454246097
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,512,3584,0.027071999178992376
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,512,3072,0.013901333014170328
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,512,2560,0.006215111249023014
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,512,3072,0.025773333178626165
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,512,2560,0.013304000099500021
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,512,2048,0.005508444375462002
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,512,2560,0.02496355606449975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,512,2048,0.013088888592190213
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,512,1536,0.004834666848182678
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,512,2048,0.024175999893082514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,512,1536,0.012770666844315
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,512,1024,0.0041964443193541635
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,512,1536,0.02404800057411194
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,512,1024,0.012246222131782107
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,512,768,0.0038275557259718576
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,512,1024,0.023152889476882085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,512,768,0.01200177768866221
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,512,512,0.00350577798154619
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,512,512,0.022557333111763
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,512,768,0.02290311124589708
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,512,512,0.011529777612951068
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,512,256,0.0031591111587153543
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,512,256,0.011277333729796939
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,512,128,0.0028880000528362063
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,512,256,0.02218222286966112
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,512,64,0.0026853332916895547
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,512,128,0.011632888681358762
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,512,128,0.02258577777279748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,512,32,0.0027377777215507296
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,512,64,0.010873777998818291
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,512,32,0.01001422190003925
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,256,65536,0.04386933313475715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,256,16384,0.007736889024575551
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,256,65536,0.014666666587193808
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,256,16384,0.02021511064635383
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,256,12288,0.006780444333950679
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,256,65536,0.0909644431538052
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,256,16384,0.041814221276177295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,256,12288,0.01809155609872606
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,256,10240,0.0069751110341813825
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,256,12288,0.03626399901178148
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,256,10240,0.017112000121010672
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,256,8192,0.006671111202902264
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,256,10240,0.03419111172358195
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,256,8192,0.016343999240133498
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,256,7168,0.006544888847404056
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,256,8192,0.031679110394583807
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,256,7168,0.015760888655980427
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,256,6144,0.0063822223908371395
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,256,7168,0.030238221089045208
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,256,6144,0.015013333823945789
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,256,5120,0.00674222202764617
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,256,6144,0.029570665624406602
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,256,5120,0.014661333627170987
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,256,4096,0.006213333457708359
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,256,5120,0.02847111225128174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,256,4096,0.014262222581439547
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,256,3584,0.006720888945791457
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,256,4096,0.02643111182583703
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,256,3584,0.014118222726715935
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,256,3072,0.0063546668324205614
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,256,3584,0.026092444856961567
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,256,3072,0.013746666411558786
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,256,2560,0.006065777606434292
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,256,3072,0.02549955579969618
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,256,2560,0.013117333253224691
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,256,2048,0.005365333209435145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,256,2560,0.024718221690919664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,256,2048,0.013001778059535556
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,256,1536,0.0046986668474144405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,256,2048,0.024290665984153748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,256,1536,0.012618666721714867
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,256,1024,0.004028444488843282
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,256,1536,0.023880889018376667
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,256,1024,0.012177777787049612
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,256,768,0.0036319999231232535
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,256,1024,0.022960889670583937
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,256,768,0.011894221935007306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,256,512,0.00332266671790017
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,256,768,0.02296888828277588
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,256,512,0.011504000259770287
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,256,256,0.0029484445436133277
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,256,512,0.02261688808600108
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,256,256,0.01126933346192042
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,256,128,0.0027822222974565295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,256,256,0.022242666946517095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,256,128,0.011559111376603445
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,256,64,0.002712888850106133
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,256,128,0.022974222898483276
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,256,32,0.002622222320901023
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,256,64,0.010599111517270407
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,256,32,0.009857777920034196
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,128,65536,0.04295288854175144
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,128,65536,0.010947555303573608
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,128,16384,0.006555555595291986
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,128,16384,0.01982133256064521
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,128,12288,0.0063546668324205614
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,128,65536,0.09098311265309651
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,128,16384,0.04094400008519491
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,128,12288,0.018024888303544786
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,128,10240,0.006565333240562015
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,128,12288,0.03630400035116408
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,128,10240,0.016919111212094624
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,128,8192,0.006440889090299606
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,128,10240,0.03419822123315599
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,128,8192,0.01609422266483307
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,128,7168,0.006177777631415262
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,128,8192,0.03149866726663377
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,128,7168,0.015600000818570455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,128,6144,0.006076444354322221
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,128,6144,0.015070221490330167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,128,7168,0.030419554975297716
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,128,4096,0.006036444256703059
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,128,5120,0.0064080001579390625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,128,6144,0.029431111282772485
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,128,5120,0.014672888649834527
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,128,5120,0.028078221612506445
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,128,4096,0.013921777407328287
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,128,3584,0.0064088888466358185
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,128,4096,0.026608000199000042
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,128,3584,0.01365066650840971
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,128,3584,0.025920000341203477
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,128,3072,0.00620977787507905
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,128,3072,0.013395555317401886
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,128,2560,0.005716444303592046
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,128,3072,0.02530488868554433
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,128,2560,0.013056889176368713
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,128,1536,0.004494222087992562
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,128,2048,0.005105777747101254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,128,2560,0.02497422198454539
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,128,2048,0.0127368892232577
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,128,2048,0.023963555693626404
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,128,1536,0.012434666355450949
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,128,1024,0.003810666501522064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,128,1024,0.012015110916561551
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,128,1536,0.023783110909991797
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,128,768,0.0034871109657817413
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,128,1024,0.023197333017985027
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,128,768,0.011805333197116852
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,128,512,0.0033039999090962936
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,128,768,0.022670222653283015
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,128,512,0.011524444652928246
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,128,256,0.002955555501911375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,128,512,0.02259644369284312
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,128,256,0.011309333145618439
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,128,128,0.0027706666539112725
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,128,256,0.02234399980968899
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,128,128,0.011499555574523078
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,128,64,0.0025964443468385274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,4,128,128,0.022691556149058874
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,128,32,0.002628444383541743
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,128,64,0.010732444624106089
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,128,32,0.009740444521109263
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,64,65536,0.008975110948085785
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,64,16384,0.006242666807439592
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,64,65536,0.04262755645645989
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,64,12288,0.006134222365087933
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,64,16384,0.01961599952644772
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,64,10240,0.0063493334584765965
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,64,12288,0.01789155602455139
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,64,8192,0.006151999864313338
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,64,10240,0.016812443733215332
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,64,8192,0.015815110670195687
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,64,7168,0.0061039999127388
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,64,7168,0.015423999892340766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,64,6144,0.006008000009589725
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,64,6144,0.014730667074521383
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,64,5120,0.006393777827421824
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,64,3072,0.006088888893524806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,64,5120,0.014241778188281588
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,64,4096,0.005934222290913264
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,64,4096,0.01366666704416275
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,64,3584,0.006425777657164468
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,64,3584,0.01352533366945055
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,64,2560,0.005580444302823808
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,64,3072,0.01309333327743742
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,64,2048,0.004951111144489712
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,64,2560,0.01312711089849472
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,64,1536,0.004367111043797599
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,64,2048,0.012698666916953193
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,64,1024,0.0037902221083641052
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,64,1536,0.012125333150227865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,64,768,0.0034799998005231223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,64,1024,0.011968889170222811
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,64,768,0.011767110890812345
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,64,512,0.0031582222630580268
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,64,256,0.0029306666304667792
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,64,512,0.011468444433477191
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,64,128,0.0027742222365405825
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,64,256,0.011063111325105032
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,64,64,0.002584888910253843
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,64,128,0.011525332927703857
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,64,64,0.010573333336247338
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,32,65536,0.041866666740841336
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,64,32,0.0026062221990691293
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,64,32,0.009767110976907942
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,32,65536,0.00795999997191959
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,32,16384,0.006160000132189856
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,32,12288,0.00618577789929178
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,32,16384,0.019648000597953796
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,32,12288,0.01777333352300856
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,32,10240,0.006296000132958095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,32,10240,0.01699466672208574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,32,8192,0.006191111273235745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,32,7168,0.006026666611433029
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,32,8192,0.01593155496650272
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,32,7168,0.015426667200194465
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,32,6144,0.005956444475385878
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,32,6144,0.015009777413474189
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,32,5120,0.006335111127959357
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,32,4096,0.0058711109062035876
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,32,5120,0.014340443743599786
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,32,3584,0.006439111298984951
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,32,4096,0.013854222165213691
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,32,3072,0.006089777582221561
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,32,3584,0.013520888984203339
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,32,2560,0.005616000129116907
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,32,3072,0.013351110948456658
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,32,2048,0.005000000198682149
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,32,2560,0.01294311136007309
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,32,1536,0.004403555558787452
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,32,2048,0.012739555703269111
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,32,1024,0.0037342223028341928
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,32,1536,0.012098666694429187
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,32,768,0.0034684443639384378
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,32,1024,0.01201333353916804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,32,512,0.003168888803985384
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,32,768,0.011633777783976661
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,32,512,0.011471110913488599
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,32,256,0.0027786667148272195
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,32,128,0.0026400000270869997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,32,256,0.011271110839313932
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,32,64,0.00250133333934678
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,32,128,0.01143200033240848
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,32,64,0.010642666783597736
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,4,32,32,0.0025048889219760895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,4,32,32,0.009762666291660732
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,65536,10240,0.16682933436499703
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,65536,12288,0.24018756548563638
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,65536,12288,0.1947022279103597
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,65536,16384,0.2535315619574653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,65536,16384,0.31060979101392955
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,65536,10240,0.3264666663275825
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,65536,10240,0.20868799421522352
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,65536,8192,0.13757067256503636
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,65536,8192,0.16992178228166369
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,65536,7168,0.1209448840883043
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,65536,12288,0.39038043551974827
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,65536,8192,0.261531564924452
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,65536,16384,0.520850658416748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,65536,7168,0.15498844782511392
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,65536,6144,0.1071466671095954
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,65536,7168,0.22949777709113228
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,65536,6144,0.14070755905575222
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,65536,4096,0.07699910799662273
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,65536,5120,0.09153155485788982
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,65536,5120,0.11749866273668076
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,65536,4096,0.13257243898179796
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,65536,3584,0.070177780257331
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,65536,4096,0.10081244177288479
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,65536,6144,0.1965822246339586
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,65536,3584,0.09090755383173625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,65536,3584,0.11587822437286377
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,65536,3072,0.06758399804433186
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,65536,5120,0.16493777434031168
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,65536,3072,0.08185778061548869
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,65536,3072,0.10117333465152317
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,65536,2560,0.05895555681652493
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,65536,2560,0.07384355862935384
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,65536,2048,0.047930667797724404
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,65536,1536,0.042944888273874916
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,65536,2560,0.08382133642832439
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,65536,1024,0.035953776703940496
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,65536,1024,0.03566222058402167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,65536,1536,0.05421955717934502
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,65536,2048,0.06399289104673597
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,65536,1536,0.051967998345692955
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,65536,2048,0.06784266895718045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,65536,1024,0.0449813339445326
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,65536,768,0.03402400016784668
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,65536,768,0.027620444695154827
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,65536,256,0.02676800058947669
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,65536,768,0.04154133465554979
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,65536,512,0.029438220792346533
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,65536,128,0.007445333732499017
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,65536,256,0.011286222272449069
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,65536,128,0.024336889386177063
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,65536,256,0.03352533446417914
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,65536,64,0.00553688903649648
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,65536,128,0.02756888833310869
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,65536,512,0.037075555986828275
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,65536,512,0.01974311139848497
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,65536,32,0.004290666845109728
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,65536,64,0.02105955613984002
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,65536,32,0.020875554945733812
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,16384,16384,0.07651644282870822
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,16384,16384,0.10663111342324151
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,16384,12288,0.06119733386569553
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,16384,12288,0.09452355570263332
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,16384,10240,0.05392355720202128
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,16384,12288,0.08171200089984469
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,16384,10240,0.07526310947206286
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,16384,16384,0.12479644351535374
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,16384,10240,0.07974933253394233
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,16384,8192,0.04615199897024366
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,16384,8192,0.06422311067581177
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,16384,7168,0.04567733407020569
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,16384,8192,0.06459999746746488
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,16384,7168,0.05921510855356852
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,16384,6144,0.038359999656677246
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,16384,6144,0.05439022183418274
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,16384,7168,0.056663109196556934
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,16384,6144,0.04905955659018623
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,16384,5120,0.041738665766186185
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,16384,5120,0.03456799851523505
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,16384,5120,0.04990666773584154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,16384,4096,0.030743112166722614
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,16384,4096,0.04510844416088528
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,16384,3584,0.029143999020258587
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,16384,3584,0.03009422289000617
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,16384,4096,0.039044443104002215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,16384,3584,0.04419111212094625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,16384,3072,0.027152000202072993
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,16384,3072,0.02642311155796051
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,16384,3072,0.041469332244661115
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,16384,2560,0.025080889463424683
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,16384,2048,0.022304000125990972
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,16384,2560,0.039149334033330284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,16384,2560,0.022727999422285292
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,16384,2048,0.03524444500605265
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,16384,1536,0.020232000284724765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,16384,2048,0.01869155632125007
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,16384,1536,0.03347466720475091
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,16384,1024,0.018375999397701688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,16384,768,0.009104000197516548
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,16384,1024,0.0310204459561242
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,16384,1536,0.015009777413474189
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,16384,768,0.01793333391348521
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,16384,1024,0.011016888750924004
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,16384,512,0.007019555403126611
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,16384,512,0.016028443972269695
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,16384,768,0.030637333790461223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,16384,256,0.00499199993080563
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,16384,128,0.0034417777011791864
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,16384,65536,0.3103875584072537
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,16384,65536,0.2578195465935601
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,16384,256,0.013737777868906656
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,16384,512,0.028735111157099407
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,16384,256,0.025083555115593806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,16384,128,0.01275644451379776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,16384,128,0.02180355621708764
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,16384,64,0.0031386667655573953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,16384,32,0.003108444520168834
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,16384,64,0.012348444097571902
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,16384,32,0.011765333513418833
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,12288,16384,0.06259822183185153
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,12288,12288,0.07863910992940266
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,12288,16384,0.1078631083170573
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,12288,12288,0.05042666527960035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,12288,16384,0.10399022367265488
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,12288,10240,0.04542577928966946
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,12288,12288,0.08555110957887437
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,12288,8192,0.04007644454638163
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,12288,10240,0.07847022347980075
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,12288,10240,0.06872533427344428
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,12288,8192,0.05463022324774
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,16384,65536,0.5378746456570095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,12288,8192,0.06495110856162177
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,12288,7168,0.03634399837917752
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,12288,65536,0.22057067023383248
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,12288,7168,0.04721777637799581
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,12288,7168,0.062331557273864746
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,12288,6144,0.03389777739842733
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,12288,6144,0.04196000099182129
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,12288,6144,0.05548710955513848
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,12288,5120,0.030192000998391047
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,12288,5120,0.03534133235613505
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,12288,5120,0.05134044422043694
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,12288,4096,0.028769777880774602
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,12288,4096,0.028297778632905748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,12288,3584,0.02588711182276408
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,12288,4096,0.04636977778540718
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,12288,3584,0.025010666913456384
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,12288,3072,0.024381332927280005
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,12288,3584,0.043776889642079674
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,12288,2560,0.021920889616012573
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,12288,3072,0.02199199961291419
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,12288,2560,0.03834133346875509
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,12288,2048,0.020440888073709276
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,12288,3072,0.0410480002562205
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,12288,2560,0.018834667073355783
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,12288,2048,0.035079999102486506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,12288,2048,0.01573422220018175
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,12288,1536,0.01927199959754944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,12288,1024,0.009341333475377824
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,12288,1024,0.01776000029510922
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,12288,1536,0.032433778047561646
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,12288,1536,0.012718222207493253
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,12288,768,0.007816889219813878
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,12288,1024,0.03051111102104187
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,12288,768,0.017274666163656447
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,12288,512,0.006298666612969504
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,12288,768,0.03004266818364461
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,12288,256,0.0038008888562520346
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,12288,256,0.013340444200568728
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,12288,512,0.013979555832015144
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,12288,512,0.026763555076387193
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,12288,128,0.003346666693687439
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,12288,256,0.025084444218211707
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,12288,128,0.012516444755925072
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,12288,64,0.003040888864133093
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,12288,128,0.021627555290857952
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,12288,32,0.003031111011902491
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,12288,64,0.011917333636018964
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,12288,32,0.011555555793974133
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,12288,65536,0.3233475685119629
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,10240,16384,0.055663108825683594
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,10240,16384,0.08371822039286296
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,10240,12288,0.046104888121287026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,10240,16384,0.08959022495481704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,10240,12288,0.06860355536142985
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,12288,65536,0.397177775700887
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,10240,12288,0.06953599717881945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,10240,10240,0.04100266761249966
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,10240,8192,0.03574311070972019
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,10240,10240,0.06825244426727295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,10240,10240,0.05821422073576185
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,10240,8192,0.04752711123890347
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,10240,8192,0.05476088987456428
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,10240,7168,0.0335822237862481
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,10240,7168,0.05170399943987528
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,10240,7168,0.03814488980505202
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,10240,6144,0.03273066547181871
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,10240,65536,0.1747359964582655
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,10240,6144,0.037112000915739275
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,10240,5120,0.028456889920764502
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,10240,6144,0.0461751123269399
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,10240,5120,0.027238221632109746
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,10240,4096,0.025941333836979334
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,10240,4096,0.023795555035273235
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,10240,5120,0.04353777898682488
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,10240,4096,0.04020444552103678
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,10240,3072,0.0225600004196167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,10240,3584,0.024145777026812237
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,10240,3072,0.036791112687852644
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,10240,2560,0.02126844392882453
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,10240,3584,0.039466665850745306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,10240,3072,0.0185217774576611
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,10240,3584,0.020787555310461257
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,10240,65536,0.23575732443067762
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,10240,2048,0.01976355579164293
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,10240,2560,0.03651288814014859
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,10240,2560,0.01587199999226464
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,10240,1536,0.010738666686746808
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,10240,2048,0.03327111072010464
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,10240,1536,0.019015999303923715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,10240,1024,0.008231999973456064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,10240,2048,0.012891555825869242
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,10240,1536,0.031210667557186548
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,10240,1024,0.017450667089886136
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,10240,768,0.006931555353932911
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,10240,1024,0.03016622198952569
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,10240,768,0.01627377834584978
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,10240,512,0.00564444437623024
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,10240,768,0.029626667499542236
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,10240,512,0.01368622233470281
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,10240,256,0.003669333126809862
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,10240,512,0.025919111238585577
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,10240,256,0.013325333595275879
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,10240,128,0.0032346666687064697
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,10240,256,0.025046222739749487
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,10240,128,0.012445333103338877
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,10240,64,0.0029528888149393927
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,10240,128,0.0217875556813346
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,10240,32,0.002895111011134254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,10240,64,0.011557333171367645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,10240,32,0.011690666278203329
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,8192,16384,0.04602755440606011
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,8192,16384,0.08124177985721164
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,8192,16384,0.06694222158855863
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,8192,12288,0.037789334853490196
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,8192,12288,0.0640524427096049
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,8192,10240,0.033887111478381686
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,8192,10240,0.04187466700871786
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,8192,10240,0.059848891364203564
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,8192,12288,0.05169333351982964
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,8192,8192,0.03160888950030009
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,8192,65536,0.1373786661359999
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,8192,8192,0.05012177758746677
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,8192,8192,0.03420177764362759
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,8192,7168,0.028258668051825628
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,8192,7168,0.030277334981494482
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,8192,7168,0.04932622114817301
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,8192,6144,0.026280888252788123
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,8192,6144,0.026570666167471144
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,8192,65536,0.20925511254204643
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,8192,5120,0.024008888337347243
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,8192,6144,0.04478844337993198
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,8192,5120,0.04199377695719401
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,8192,3584,0.02013688948419359
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,8192,4096,0.022515555222829182
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,8192,5120,0.02288888891537984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,8192,4096,0.019306666321224637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,8192,4096,0.03727644350793626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,8192,3584,0.03819200065400865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,10240,65536,0.3753884368472629
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,8192,3584,0.017024000485738117
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,8192,3072,0.01891644464598762
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,8192,3072,0.03514488869243198
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,8192,2560,0.018055111169815063
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,8192,2048,0.01686311099264357
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,8192,2560,0.03412977854410807
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,8192,1536,0.009182222187519073
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,8192,3072,0.01512888901763492
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,8192,2048,0.031621333625581526
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,8192,2560,0.013088888592190213
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,8192,1536,0.015980444020695157
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,8192,2048,0.011124444504578909
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,8192,1024,0.00720711135201984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,8192,1536,0.030360887447992962
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,8192,1024,0.0144195556640625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,8192,768,0.006241777704821692
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,8192,1024,0.02908088763554891
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,8192,768,0.013960000541475085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,8192,512,0.004998222407367495
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,8192,768,0.028869334194395278
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,8192,512,0.012146666646003723
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,8192,256,0.003354666754603386
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,8192,512,0.0247013337082333
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,8192,256,0.01204533295498954
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,8192,64,0.0029528888149393927
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,8192,128,0.0032008888406885993
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,8192,256,0.024691555235120986
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,8192,128,0.010926222635640038
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,8192,128,0.021576889687114294
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,8192,32,0.002983999955985281
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,8192,64,0.010372444159454769
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,8192,32,0.010100444157918295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,7168,16384,0.042308443122439914
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,7168,16384,0.07288089063432482
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,8192,65536,0.26175999641418457
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,7168,12288,0.047127111090554126
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,7168,12288,0.0354053344991472
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,7168,12288,0.0624471108118693
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,7168,10240,0.03203377789921231
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,7168,16384,0.05981333388222588
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,7168,10240,0.056996444861094155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,7168,10240,0.04483644498719109
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,7168,8192,0.028451555305057104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,7168,8192,0.033161779244740806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,7168,8192,0.0495742228296068
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,7168,7168,0.026723555392689172
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,7168,7168,0.04787110951211718
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,7168,7168,0.03019555409749349
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,7168,6144,0.024823110964563157
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,7168,6144,0.04305422306060791
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,7168,5120,0.022652443912294176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,7168,6144,0.026200888885392085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,7168,5120,0.02295911146534814
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,7168,4096,0.01995733380317688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,7168,4096,0.036293334431118436
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,7168,3584,0.01941066649225023
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,7168,3584,0.03674044542842441
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,7168,5120,0.04115555683771769
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,7168,4096,0.0200311111079322
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,7168,3072,0.018387556076049805
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,7168,3584,0.017824888229370117
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,7168,2560,0.012413333687517377
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,7168,3072,0.03428444597456191
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,7168,2560,0.01756533318095737
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,7168,3072,0.016014221641752455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,7168,65536,0.2039182186126709
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,7168,65536,0.12485688262515598
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,7168,2048,0.010464888479974536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,7168,2560,0.032891554964913264
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,7168,2048,0.01657511128319634
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,7168,1536,0.00866844422287411
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,7168,2048,0.031176888280444678
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,7168,1536,0.01614755557643043
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,7168,1024,0.0068408888247278
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,7168,1536,0.030018667380015056
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,7168,1024,0.014051554931534661
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,7168,1024,0.027043556173642475
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,7168,768,0.006010666489601135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,7168,768,0.012923555241690742
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,7168,512,0.004642666627963384
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,7168,768,0.028367112080256145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,7168,512,0.024717332588301763
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,7168,512,0.011897777517636618
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,7168,256,0.003415999934077263
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,7168,256,0.011957333319716983
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,7168,128,0.01108888867828581
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,7168,128,0.003121777748068174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,7168,256,0.024689777029885188
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,7168,64,0.0028479999552170434
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,7168,128,0.02123911182085673
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,7168,64,0.01035377797153261
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,7168,32,0.002850666642189026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,7168,32,0.010173333187898
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,6144,16384,0.03983111182848612
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,6144,16384,0.07163466347588433
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,6144,16384,0.05789955457051595
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,7168,65536,0.23188800281948516
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,6144,12288,0.03436533278889126
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,6144,12288,0.0433884428607093
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,6144,12288,0.06160800324545967
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,6144,10240,0.03195377853181627
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,6144,8192,0.027341332700517442
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,6144,10240,0.036488887336519026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,6144,10240,0.056419557995266385
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,6144,8192,0.029747555653254192
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,6144,7168,0.025759999950726826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,6144,8192,0.04925866590605842
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,6144,65536,0.11486932966444229
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,6144,7168,0.024153777294688757
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,6144,5120,0.02160444524553087
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,6144,7168,0.04578488734033373
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,6144,6144,0.02309155629740821
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,6144,6144,0.023574221465322707
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,6144,6144,0.04239822096294827
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,6144,5120,0.03911911116706
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,6144,4096,0.01921066641807556
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,6144,5120,0.020083554916911654
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,6144,4096,0.03534311056137085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,6144,3584,0.018976888722843595
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,6144,3584,0.037167112032572426
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,6144,3072,0.017865777015686035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,6144,3072,0.03418933351834615
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,6144,65536,0.19753777980804443
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,6144,2560,0.011772444678677453
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,6144,4096,0.016748444901572335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,6144,3584,0.015031110909250049
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,6144,2560,0.017163554827372234
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,6144,3072,0.01348444405529234
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,6144,2048,0.009929777847396003
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,6144,1536,0.008231110870838165
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,6144,2560,0.03265955713060167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,6144,2048,0.016189333465364244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,6144,2048,0.0310479998588562
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,6144,1536,0.015172445111804538
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,6144,1024,0.006395555618736479
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,6144,1536,0.03131555517514547
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,6144,1024,0.012747555143303342
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,6144,768,0.005588444570700328
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,6144,1024,0.02679644525051117
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,6144,768,0.012802666260136498
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,6144,512,0.003999111139112049
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,6144,512,0.011863110793961419
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,6144,768,0.02646133303642273
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,6144,256,0.0033297776761982176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,6144,512,0.024696888195143804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,6144,256,0.011894221935007306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,6144,128,0.0030453333424197305
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,6144,256,0.024646222591400146
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,6144,128,0.010969777901967367
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,6144,64,0.0028106667515304354
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,6144,128,0.02218933403491974
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,6144,64,0.010261333651012844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,6144,32,0.0028328889360030494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,6144,32,0.010068444742096795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,5120,16384,0.036095112562179565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,5120,16384,0.07004533211390178
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,5120,12288,0.03717155588997735
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,5120,16384,0.04808533191680908
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,5120,12288,0.030846221579445735
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,5120,12288,0.058783113956451416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,5120,10240,0.027914666467242773
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,5120,10240,0.05419199996524387
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,5120,10240,0.03161866797341241
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,5120,8192,0.025337777203983728
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,5120,8192,0.026749332745869953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,5120,8192,0.04737066560321384
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,5120,65536,0.10007822513580322
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,5120,7168,0.04506133331192864
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,5120,7168,0.024102222588327196
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,5120,7168,0.02318222158484989
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,5120,6144,0.021983999345037673
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,6144,65536,0.19635466734568277
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,5120,6144,0.042500442928738065
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,5120,5120,0.020264888803164165
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,5120,6144,0.018473777506086562
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,5120,4096,0.01885422236389584
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,5120,5120,0.039346665143966675
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,5120,5120,0.016120889120631747
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,5120,4096,0.03538844320509169
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,5120,4096,0.01387288918097814
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,5120,3584,0.01868266695075565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,5120,65536,0.17935289276970756
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,5120,3072,0.010971555279360877
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,5120,3584,0.03556088937653436
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,5120,3584,0.012383999923865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,5120,3072,0.017082666357358296
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,5120,2560,0.00959911114639706
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,5120,3072,0.03364177876048618
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,5120,2560,0.03219822380277846
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,5120,2048,0.030677331818474665
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,5120,2048,0.008428444465001425
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,5120,2560,0.01644355555375417
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,5120,2048,0.015795555379655626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,5120,1536,0.007184888753626082
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,5120,1536,0.014846222268210517
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,5120,1024,0.005927111125654644
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,5120,1536,0.030377778742048476
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,5120,1024,0.012460444536474017
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,5120,768,0.005060444275538127
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,5120,1024,0.026312889324294195
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,5120,768,0.012644444074895648
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,5120,512,0.003914666672547658
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,5120,768,0.025945777694384258
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,5120,512,0.011752888560295105
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,5120,256,0.0032053333189752367
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,5120,512,0.024661334024535284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,5120,256,0.011744889120260874
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,5120,128,0.0030133333057165146
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,5120,256,0.024711110525661047
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,5120,128,0.010849778023031024
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,5120,64,0.0027626665929953256
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,5120,128,0.021526222427686054
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,5120,32,0.010151111417346531
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,5120,64,0.010117332968446944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,5120,32,0.0028133332315418455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,4096,16384,0.031067556805080835
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,4096,16384,0.06976622343063354
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,4096,12288,0.030790223015679255
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,4096,12288,0.026787555880016748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,4096,16384,0.03929333223236932
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,4096,12288,0.056418663925594754
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,4096,10240,0.02481599979930454
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,4096,10240,0.026595556073718604
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,4096,10240,0.052616000175476074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,5120,65536,0.18441243966420492
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,4096,8192,0.021852445271280076
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,4096,8192,0.021937777598698933
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,4096,7168,0.021942221456103857
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,4096,8192,0.04541244440608554
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,4096,7168,0.045151111152437
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,4096,6144,0.01943288909064399
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,4096,7168,0.020125332805845473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,4096,5120,0.018671999375025432
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,4096,6144,0.03917955689960056
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,4096,5120,0.03863288958867391
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,4096,6144,0.017264889346228707
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,4096,4096,0.01761511133776771
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,4096,5120,0.015440000428093804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,4096,3584,0.012402666939629449
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,4096,4096,0.03585955500602722
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,4096,3584,0.01703288820054796
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,4096,3584,0.0344177782535553
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,4096,3072,0.011218667030334473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,4096,4096,0.013463999662134381
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,4096,3072,0.016529776983790927
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,4096,65536,0.07849778069390191
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,4096,2560,0.008392000363932716
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,4096,3072,0.03276888860596551
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,4096,65536,0.17932977941301134
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,4096,2560,0.015818667080667283
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,4096,1536,0.0064239998658498125
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,4096,2048,0.007495111061467065
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,4096,2560,0.03351022137535943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,4096,1536,0.0127697777416971
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,4096,2048,0.015027556154463025
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,4096,2048,0.03042311138576931
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,4096,1024,0.00565244464410676
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,4096,1536,0.028802666399214003
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,4096,1024,0.012275555895434486
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,4096,768,0.003965333518054751
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,4096,1024,0.02587466604179806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,4096,768,0.012486221889654795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,4096,512,0.003592888928121991
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,4096,768,0.02595377796226078
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,4096,512,0.011744000017642975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,4096,256,0.0032115555885765287
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,4096,512,0.024477332830429077
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,4096,256,0.011895111037625207
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,4096,128,0.0029360000044107437
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,4096,256,0.02455111179086897
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,4096,128,0.01090755561987559
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,4096,64,0.0027466666781240036
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,4096,128,0.021377777059872944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,4096,32,0.0027226667023367355
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,4096,64,0.01014311114947001
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,4096,32,0.009889778163697984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,4096,65536,0.1334177785449558
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3584,16384,0.029515554507573444
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,3584,16384,0.06338044669893053
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3584,12288,0.02569955587387085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3584,16384,0.034830222527186074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,3584,12288,0.055592887931399874
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3584,12288,0.026005332668622334
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3584,8192,0.021144888467258875
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3584,10240,0.023268444670571223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3584,10240,0.02279022170437707
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,3584,10240,0.051592889759275645
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,3584,8192,0.044446223311954074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3584,65536,0.07085511419508192
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3584,7168,0.01980533368057675
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3584,8192,0.01941866676012675
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,3584,7168,0.04353511002328661
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3584,6144,0.018651554981867473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,3584,6144,0.03952711158328586
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3584,5120,0.01811911165714264
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,3584,5120,0.03777155611250136
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3584,4096,0.012780444489585029
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3584,6144,0.015847111741701763
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3584,4096,0.01719999975628323
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3584,7168,0.018263111511866253
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3584,5120,0.014362666341993542
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3584,3584,0.01181777815024058
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3584,3584,0.016398222910033334
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,3584,4096,0.033857779370413885
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3584,3072,0.010858666565683154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,3584,65536,0.16698933972252739
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,3584,3584,0.033783111307356097
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3584,2560,0.008157333566082848
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3584,3072,0.01649599936273363
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,3584,3072,0.03254400028122796
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3584,2560,0.015384000208642749
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3584,2048,0.007080000307824876
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,3584,2560,0.03295022249221802
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3584,2048,0.014487110906177096
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3584,1536,0.006315555423498154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,3584,2048,0.029600001043743555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3584,1024,0.0124177775449223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3584,1536,0.013941332697868347
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3584,1024,0.004865777575307422
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,3584,1536,0.026502221822738647
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3584,768,0.0038648889296584656
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,3584,1024,0.02556266718440586
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3584,768,0.012488000094890594
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3584,512,0.003674666500753827
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,3584,768,0.02569777766863505
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3584,512,0.011714666253990598
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3584,256,0.003154666680428717
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,3584,512,0.024495111571417913
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3584,256,0.011712888876597086
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3584,128,0.0028497777465316984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,3584,256,0.024082667297787134
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3584,128,0.010767999622556897
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3584,64,0.0026897777699761917
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,3584,128,0.021271111236678228
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3584,64,0.010187555518415239
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3584,32,0.0026702222724755607
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3584,32,0.009977777798970541
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3072,16384,0.027843554814656574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,3072,16384,0.06804000006781684
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3072,12288,0.025240888198216755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3072,16384,0.029418667157491047
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3072,12288,0.02455644475089179
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3072,10240,0.022410665949185688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3584,65536,0.12352266576555039
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,3072,12288,0.05481422278616163
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,3072,10240,0.050254222419526845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3072,8192,0.020948444803555805
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,3072,8192,0.044475555419921875
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3072,10240,0.021894221504529316
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3072,7168,0.018980445133315194
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3072,8192,0.018067555295096505
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,3072,7168,0.04146755403942532
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3072,6144,0.015415999624464246
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3072,6144,0.018329777651362948
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3072,7168,0.016181333197487723
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3072,5120,0.013809777796268463
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,3072,6144,0.03939555419815911
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3072,5120,0.017449777987268236
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3072,4096,0.01186044431395001
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3072,4096,0.01681066718366411
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,3072,5120,0.03683288892110189
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3072,3584,0.011223999990357293
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,3072,4096,0.03326488865746392
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3072,3584,0.01646133263905843
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3072,3072,0.010123555858929953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,3072,3584,0.033618665403789945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3072,3072,0.015531554818153381
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3072,2560,0.007701333198282454
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3072,65536,0.06629866361618042
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,3072,3072,0.032285332679748535
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3072,2560,0.015180443723996481
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,3072,65536,0.15787466367085776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,3072,2560,0.03135910961363051
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3072,2048,0.006813333266311222
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3072,2048,0.01330044451687071
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3072,1536,0.006032889087994893
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,3072,2048,0.02776177724202474
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3072,1536,0.0127368892232577
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3072,1024,0.004425777743260066
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,3072,1536,0.026595556073718604
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3072,1024,0.012285333540704517
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3072,768,0.004012444367011388
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,3072,1024,0.025380444195535447
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3072,768,0.012464889221721225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3072,512,0.003528000166018804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,3072,768,0.0255395554833942
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3072,512,0.011546666423479715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3072,256,0.003120888852410846
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,3072,512,0.024215110474162634
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3072,256,0.011771555576059552
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3072,128,0.0028373334142896864
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,3072,256,0.024412444896168176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3072,128,0.010831111007266574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3072,64,0.0026737778551048706
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,3072,128,0.020962667134073045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3072,32,0.0026835555003749
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3072,64,0.009903999666372934
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,3072,65536,0.103584885597229
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,3072,32,0.009850666754775578
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2560,16384,0.027121777335802715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,2560,16384,0.05918311410480075
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2560,12288,0.02348444362481435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2560,12288,0.02141155633661482
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2560,16384,0.026733333865801494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2560,10240,0.02124888863828447
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,2560,10240,0.049822221199671425
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,2560,12288,0.04979999860127767
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2560,8192,0.019890666007995605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,2560,8192,0.04244088795449999
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2560,7168,0.018887110882335238
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2560,8192,0.015662221444977652
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,2560,7168,0.04291022155019972
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2560,6144,0.012947555217478009
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2560,10240,0.01854044364558326
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2560,7168,0.014567999376191033
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2560,6144,0.018427555759747822
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2560,5120,0.011663999822404651
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,2560,6144,0.039600888888041176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2560,5120,0.017364444004164804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2560,4096,0.0105253333846728
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2560,4096,0.016552888684802584
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,2560,5120,0.0349333319399092
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2560,3584,0.010024000373151567
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,2560,4096,0.03344710999064975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2560,3584,0.016023111012246873
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2560,3072,0.0092622223827574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,2560,3584,0.03423466616206699
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2560,3072,0.015258666541841296
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2560,2560,0.007343110938866933
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,2560,3072,0.03214755654335022
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2560,65536,0.06484355529149373
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2560,2560,0.013840888937314352
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2560,2048,0.006695999867386288
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,2560,65536,0.14845421579149035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2560,2048,0.013254221942689685
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2560,1536,0.005702222386995952
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,2560,2560,0.029397332006030615
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,2560,2048,0.026857778429985046
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2560,1536,0.01253600004646513
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2560,1024,0.004101333518822988
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,2560,1536,0.026927999324268762
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2560,1024,0.012152888708644442
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2560,768,0.003740444365474913
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,2560,1024,0.025761778155962627
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2560,768,0.012448888685968188
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2560,512,0.0034346667428811393
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,2560,768,0.025032000409232244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2560,512,0.011577777564525604
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2560,256,0.003080888961752256
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,2560,512,0.02421155571937561
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2560,256,0.011834666960769229
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2560,128,0.0028284444577164124
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,2560,256,0.024282667371961806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2560,128,0.010465777582592435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2560,64,0.002663111107216941
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2560,32,0.002728888971938027
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,2560,128,0.02126311096880171
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2560,64,0.01000533335738712
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2560,32,0.010178666975763109
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2560,65536,0.09700355264875625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2048,12288,0.022319111559126113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,2048,16384,0.06166399849785698
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2048,16384,0.02568711174858941
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,2048,12288,0.05230577786763509
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2048,12288,0.018055111169815063
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2048,10240,0.020875554945733812
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2048,16384,0.02172800070709652
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2048,8192,0.019259555472267997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2048,10240,0.016594666573736403
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,2048,10240,0.05216622352600098
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2048,7168,0.012805333567990197
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2048,8192,0.01367555558681488
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,2048,8192,0.04250133368704054
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2048,7168,0.01829777823554145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,2048,7168,0.045751998821894325
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2048,6144,0.011757333245542316
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2048,65536,0.05270933442645603
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2048,6144,0.017704889178276062
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2048,5120,0.010873777998818291
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2048,5120,0.01683200067943997
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,2048,6144,0.041604446040259466
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2048,4096,0.00963288876745436
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,2048,5120,0.03587111168437534
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2048,4096,0.016188444362746347
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2048,3584,0.008995555341243744
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,2048,4096,0.034527109728919134
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2048,3584,0.015463999576038785
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,2048,65536,0.1663395563761393
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2048,3072,0.008616888688670265
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,2048,3584,0.03321689036157396
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2048,3072,0.014382221632533602
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,2048,3072,0.03005511230892605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2048,2560,0.007134222322040134
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2048,2560,0.013344888885815939
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2048,2048,0.013119999733236102
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2048,2048,0.0064328888224230866
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,2048,2560,0.02842666705449422
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2048,1536,0.005053333524200651
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,2048,2048,0.027127999398443434
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2048,1536,0.012538666526476542
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2048,1024,0.004039999925427967
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,2048,1536,0.026288888520664636
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2048,1024,0.012049777640236748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2048,768,0.0036640001667870414
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2048,65536,0.06946400139066908
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,2048,1024,0.02553066611289978
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2048,768,0.01237066669596566
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2048,512,0.0033564445459180405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,2048,768,0.02540622154871623
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2048,512,0.011543110840850405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2048,256,0.003030222323205736
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,2048,512,0.02418488926357693
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,2048,128,0.020942222740915086
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2048,128,0.0027866665687825945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2048,256,0.011721777419249216
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,2048,256,0.024551999237802293
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2048,128,0.010592888626787396
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2048,64,0.0026524443593290118
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,2048,32,0.0026604444202449587
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2048,64,0.010246222217877706
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,2048,32,0.009658666948477427
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1536,16384,0.023966223001480103
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,1536,16384,0.058323555522494845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1536,12288,0.020251555575264823
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1536,10240,0.01367200000418557
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,1536,12288,0.051029334465662636
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1536,12288,0.014379555980364481
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1536,10240,0.019268444842762418
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1536,16384,0.018754666050275166
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,1536,10240,0.04820800158712599
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1536,8192,0.011505777637163797
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1536,8192,0.018619555566045973
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1536,7168,0.010943111446168689
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,1536,8192,0.04252622193760342
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1536,7168,0.017834666702482436
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1536,6144,0.009928888744778102
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1536,6144,0.017433777451515198
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,1536,7168,0.040236443281173706
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1536,5120,0.009365333451165093
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,1536,6144,0.038669334517584905
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1536,5120,0.0162462227874332
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1536,4096,0.00867377801073922
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,1536,5120,0.03532000051604377
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1536,4096,0.014661333627170987
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1536,65536,0.05580800109439426
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1536,3584,0.008512000242869059
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1536,3072,0.008117333054542542
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,1536,4096,0.03195733494228787
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1536,3584,0.015000889698664347
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,1536,3072,0.029626667499542236
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,1536,65536,0.14453066719902888
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,1536,3584,0.031009776724709406
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1536,3072,0.013763555222087435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1536,2560,0.00927111092540953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1536,2560,0.013051555388503604
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,1536,2048,0.026664889521068994
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1536,2048,0.008007110820876228
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,1536,2560,0.028036445379257202
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1536,2048,0.012947555217478009
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1536,1536,0.006621333460013072
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1536,65536,0.05393777622116936
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1536,768,0.004584889031118817
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1536,1536,0.012473777764373355
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1536,1024,0.005223999834722943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,1536,1536,0.025722665919197932
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1536,1024,0.011952000359694162
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1536,512,0.011428444749779172
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,1536,1024,0.026231999198595684
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1536,768,0.01222400036123064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1536,512,0.003927999900446998
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,1536,768,0.0258879992696974
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1536,256,0.003257777748836411
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,1536,512,0.02420622275935279
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1536,256,0.01167022188504537
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1536,64,0.002793777734041214
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1536,128,0.003022222262289789
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,1536,256,0.024251555403073628
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1536,128,0.010583110981517367
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,1536,128,0.021204445097181533
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1536,32,0.0028159999185138275
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1536,64,0.009893333746327294
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1536,32,0.009515555368529426
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1024,16384,0.022279999322361414
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1024,12288,0.012063999970753988
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,1024,16384,0.0528675549560123
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1024,12288,0.019847111569510568
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1024,10240,0.011132444772455426
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1024,16384,0.014596444037225513
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,1024,12288,0.04522666666242811
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1024,10240,0.019034667147530448
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1024,8192,0.009923555784755284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1024,65536,0.04715289009941948
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,1024,10240,0.044716444280412465
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1024,8192,0.017637333936161466
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1024,7168,0.009209777745935652
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,1024,8192,0.03912533322970072
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1024,7168,0.01630577776167128
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1024,6144,0.008381333616044786
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,1024,7168,0.03970844546953837
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1024,6144,0.01550222271018558
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1024,5120,0.008344000412358178
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,1024,65536,0.12049867047203912
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,1024,6144,0.03535644544495477
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1024,5120,0.014793778459231058
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1024,4096,0.007739555504586961
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,1024,5120,0.03405066662364536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1024,4096,0.014179555906189812
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1024,65536,0.0395395557085673
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1024,3584,0.007858666280905405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,1024,4096,0.029519110918045044
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1024,3072,0.01350400017367469
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1024,3584,0.013789333403110504
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1024,3072,0.007080888582600488
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,1024,3584,0.029845333761639063
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,1024,3072,0.028431110911899145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1024,2560,0.014369777507252164
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1024,2560,0.013502221968438892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1024,2048,0.010457778142558204
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,1024,2560,0.026854222019513447
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1024,2048,0.013103110922707452
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1024,1024,0.007816000117195977
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1024,1536,0.008583111067612966
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,1024,2048,0.02650044361750285
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1024,1536,0.012471111284361945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,1024,1536,0.025361778007613286
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1024,1024,0.011980444192886353
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1024,768,0.0063688887490166565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,1024,1024,0.024745777249336243
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1024,512,0.004916444420814514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1024,768,0.012381333443853589
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,1024,768,0.025457777910762366
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1024,512,0.011511999699804517
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1024,256,0.003956444561481476
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,1024,512,0.02388711108101739
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1024,256,0.011571555501884885
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1024,128,0.0033448889023727844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,1024,256,0.024114666713608637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1024,128,0.011044444309340583
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1024,64,0.0030053332448005676
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,1024,128,0.020984888076782227
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,1024,32,0.003071111109521654
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1024,64,0.010001777774757808
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,1024,32,0.009690666364298927
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,768,16384,0.011720000041855706
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,768,16384,0.021385777327749465
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,768,12288,0.010296888649463654
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,768,12288,0.01947377787695991
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,768,10240,0.009631111390060848
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,768,16384,0.04855288730727302
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,768,12288,0.044144001271989614
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,768,8192,0.00887288898229599
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,768,10240,0.0184568895234002
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,768,8192,0.01664977769056956
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,768,10240,0.04217244519127739
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,768,7168,0.008355555435021719
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,768,8192,0.037335998482174344
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,768,65536,0.04418222109476725
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,768,7168,0.01574222246805827
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,768,6144,0.007840889195601145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,768,7168,0.03710222244262695
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,768,6144,0.015207999282413058
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,768,65536,0.11252000596788193
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,768,5120,0.007822222179836696
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,768,6144,0.03364533185958862
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,768,5120,0.01480977733929952
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,768,4096,0.007408888803588019
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,768,5120,0.030192000998391047
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,768,4096,0.014092443717850579
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,768,3584,0.007140444384680853
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,768,3584,0.028208000792397395
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,768,3072,0.013427555561065674
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,768,4096,0.029169778029123943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,768,3584,0.014008000493049622
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,768,3072,0.006777777853939269
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,768,65536,0.029993779129452173
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,768,2560,0.008712000317043727
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,768,3072,0.02720977697107527
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,768,2560,0.013375110924243927
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,768,2048,0.0069226668112807805
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,768,2560,0.026642666922675237
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,768,2048,0.012973333398501078
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,768,1536,0.005564444594913059
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,768,2048,0.026029333472251892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,768,1536,0.012477333347002665
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,768,1024,0.005014222115278244
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,768,1536,0.024920889072948035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,768,1024,0.011920000116030375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,768,768,0.004318222403526306
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,768,1024,0.02458400030930837
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,768,768,0.012355555262830524
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,768,512,0.0037724446091387006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,768,768,0.024849777420361836
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,768,512,0.01144088887506061
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,768,256,0.0032355555643637977
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,768,512,0.024661334024535284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,768,256,0.011741333537631564
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,768,128,0.002954666606254048
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,768,256,0.024069334069887798
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,768,128,0.01070844464831882
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,768,64,0.0027822222974565295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,768,128,0.02093688812520769
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,768,32,0.0028213332924577924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,768,64,0.010274666878912184
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,768,32,0.00978133330742518
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,512,16384,0.009906666974226633
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,512,65536,0.043599112166298755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,512,16384,0.021631111701329548
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,512,16384,0.0439991123146481
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,512,12288,0.008640888664457533
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,512,65536,0.09439377652274238
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,512,12288,0.018575999471876357
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,512,10240,0.008403555386596257
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,512,12288,0.040566222535239324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,512,10240,0.01720266706413693
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,512,8192,0.008012444608741337
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,512,10240,0.03512710995144314
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,512,8192,0.01605688863330417
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,512,7168,0.007640000018808577
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,512,8192,0.03236533204714457
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,512,7168,0.015660444895426433
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,512,6144,0.0069751110341813825
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,512,7168,0.030963556634055242
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,512,65536,0.021877333521842957
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,512,6144,0.015097777048746744
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,512,5120,0.007144889069928064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,512,6144,0.029655112160576716
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,512,4096,0.006461333483457565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,512,5120,0.014505777094099255
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,512,5120,0.029234667619069416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,512,4096,0.01422044469250573
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,512,3584,0.006767999794748094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,512,4096,0.027486221657858953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,512,3584,0.013982221484184265
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,512,3072,0.006657777975002925
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,512,3584,0.026802667313151892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,512,3072,0.013330666555298699
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,512,2560,0.006119110931952794
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,512,3072,0.027103111147880554
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,512,2560,0.013345777988433838
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,512,2048,0.005485333502292633
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,512,2560,0.025964443882306416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,512,2048,0.01294844432009591
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,512,1536,0.004862222406599256
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,512,2048,0.02528711160024007
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,512,1536,0.012623111406962076
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,512,1024,0.004134222037262387
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,512,1536,0.025643555654419795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,512,1024,0.01202933324707879
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,512,768,0.003765333443880081
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,512,1024,0.024452444579866197
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,512,768,0.01218933363755544
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,512,512,0.0034835553831524322
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,512,768,0.024816888901922438
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,512,512,0.011494222614500256
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,512,256,0.003126222226354811
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,512,512,0.024363555841975745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,512,256,0.011578666667143503
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,512,128,0.0028924445311228433
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,512,256,0.024079110887315538
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,512,128,0.010560000108347999
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,512,64,0.0026737778551048706
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,512,128,0.02088533341884613
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,512,32,0.0027173333283927706
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,512,64,0.010284444524182213
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,512,32,0.009908444351620143
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,256,16384,0.019842666056421068
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,256,65536,0.09160177575217353
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,256,65536,0.04290133383538988
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,256,16384,0.007822222179836696
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,256,12288,0.006816888848940532
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,256,65536,0.014514666464593677
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,256,16384,0.04132533404562209
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,256,12288,0.018035555879275005
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,256,10240,0.0069208890199661255
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,256,10240,0.0169840008020401
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,256,12288,0.036539554595947266
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,256,8192,0.006527110934257507
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,256,10240,0.03436977664629618
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,256,8192,0.01610577768749661
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,256,7168,0.0064204442832205035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,256,8192,0.031339555978775024
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,256,5120,0.006753777878151999
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,256,7168,0.015571556157535978
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,256,6144,0.0063164445261160536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,256,7168,0.030360000001059637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,256,6144,0.015119110544522604
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,256,6144,0.029968000120586816
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,256,5120,0.028407110108269587
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,256,4096,0.014195554786258273
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,256,5120,0.01461866663561927
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,256,4096,0.00629333323902554
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,256,3584,0.006766222417354584
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,256,3584,0.026605332891146343
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,256,4096,0.027287999788920086
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,256,3584,0.013845333622561561
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,256,3072,0.006487111250559489
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,256,3072,0.01349955548842748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,256,2560,0.006030222194062338
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,256,3072,0.02641333308484819
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,256,2560,0.013292444248994192
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,256,2048,0.00536088893810908
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,256,2560,0.02580533259444767
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,256,2048,0.01293955577744378
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,256,1536,0.0046897778908411665
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,256,2048,0.02511288887924618
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,256,1536,0.012444444828563266
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,256,1024,0.003995555556482739
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,256,1024,0.011985777980751462
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,256,1536,0.024658666716681585
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,256,768,0.0036515556275844574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,256,1024,0.024636444118287828
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,256,768,0.012283555335468717
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,256,512,0.003334222154484855
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,256,768,0.02482044365670946
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,256,512,0.011415999796655444
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,256,256,0.002987555538614591
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,256,64,0.0026373333401150177
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,256,512,0.023732443650563557
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,256,32,0.0026186667382717133
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,256,128,0.0027431110954946945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,256,256,0.01184800018866857
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,256,32,0.009723555710580613
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,256,256,0.023338667220539514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,256,128,0.010964444114102257
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,256,128,0.020869332883093093
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,256,64,0.009891555541091496
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,128,65536,0.011300444602966309
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,128,16384,0.006272000157170826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,128,65536,0.042393777105543345
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,128,16384,0.019461333751678467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,128,65536,0.09163555834028457
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,128,12288,0.006148444281684027
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,128,16384,0.04103555613093906
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,128,12288,0.017631111873520743
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,128,10240,0.00637155564294921
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,128,12288,0.03653600149684482
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,128,10240,0.016864000095261466
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,128,8192,0.006298666612969504
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,128,10240,0.03401777810520596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,128,8192,0.015850666496488783
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,128,7168,0.006151999864313338
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,128,8192,0.03207644489076402
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,128,7168,0.015203555425008139
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,128,6144,0.029071112473805744
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,128,6144,0.00600355573826366
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,128,7168,0.03029511041111416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,128,6144,0.01498222185505761
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,128,5120,0.0063582224150498705
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,128,5120,0.01441333360142178
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,128,4096,0.006026666611433029
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,128,5120,0.029041777054468792
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,128,4096,0.01371022231049008
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,128,3584,0.006478222293986215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,128,4096,0.02718488872051239
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,128,3584,0.013696889082590738
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,128,3072,0.006091555373536215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,128,3584,0.02659822172588772
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,128,3072,0.013188444077968597
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,128,2560,0.025516443782382544
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,128,2560,0.005703111075692707
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,128,3072,0.026741334133678015
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,128,2560,0.012907555533779992
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,128,2048,0.005120000077618493
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,128,2048,0.012585777375433179
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,128,1536,0.004490666505363253
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,128,2048,0.025202666719754536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,128,1536,0.012156444291273752
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,128,1024,0.0038115556041399636
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,128,1536,0.02446577780776554
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,128,1024,0.011891555454995898
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,128,768,0.0034959999223550162
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,128,1024,0.024689777029885188
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,128,768,0.012243555651770698
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,128,512,0.0033191111352708605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,128,768,0.02464444438616435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,128,512,0.011376889215575324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,128,256,0.0029768887907266617
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,128,512,0.023385778069496155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,128,256,0.011739555332395764
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,128,128,0.002776888923512565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,128,256,0.02383822202682495
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,128,128,0.010622222390439777
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,128,64,0.0026568888376156488
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,2,128,128,0.021783111823929682
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,128,64,0.00979911122057173
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,128,32,0.002630222174856398
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,128,32,0.009783999787436591
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,64,65536,0.008903111020723978
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,64,16384,0.006152888966931238
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,64,65536,0.04124177826775445
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,64,12288,0.006059555543793573
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,64,16384,0.01927199959754944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,64,10240,0.006239111224810283
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,64,12288,0.017433777451515198
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,64,8192,0.006229333165619109
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,64,10240,0.01647911138004727
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,64,7168,0.006031999985376994
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,64,8192,0.015576889117558798
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,64,7168,0.015124445160230001
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,64,6144,0.005939555664857228
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,64,5120,0.0063680000603199005
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,64,6144,0.014575999644067554
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,64,4096,0.0058666666348775225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,64,5120,0.01419911119672987
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,64,4096,0.013623110949993134
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,64,3584,0.0064284445510970215
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,64,3584,0.013464888764752282
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,64,3072,0.006116444451941385
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,64,2560,0.005626666463083691
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,64,3072,0.013171555267439948
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,64,2560,0.012879999975363413
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,64,2048,0.004966222163703707
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,64,1536,0.004426666845877965
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,64,2048,0.012508444488048553
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,64,1536,0.012090666426552666
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,64,1024,0.0037804444630940757
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,64,1024,0.011770666473441653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,64,512,0.011432888607184092
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,64,768,0.003488000068399641
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,64,768,0.01202488856183158
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,64,512,0.003157333367400699
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,64,256,0.0029564443975687027
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,64,64,0.009847111172146266
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,64,256,0.011538666983445486
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,64,128,0.0027555556346972785
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,64,64,0.0025822222232818604
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,64,128,0.010909332997269101
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,64,32,0.0025946667624844448
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,32,65536,0.008198222352398766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,64,32,0.00979555563794242
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,32,16384,0.006252444452709622
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,32,65536,0.04118666715092129
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,32,10240,0.016606221596399944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,32,16384,0.01960088809331258
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,32,7168,0.006070222291681502
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,32,12288,0.006060444646411472
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,32,12288,0.017567111386193168
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,32,10240,0.006315555423498154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,32,8192,0.006102222121424145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,32,8192,0.015561777684423657
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,32,7168,0.01517599986659156
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,32,6144,0.005884444548024072
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,32,5120,0.006310222049554189
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,32,6144,0.014814222852389017
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,32,4096,0.005937777873542573
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,32,5120,0.014344000154071383
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,32,4096,0.013732444081041547
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,32,3584,0.006412444429265128
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,32,3072,0.005965333431959152
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,32,3584,0.013450666434235044
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,32,3072,0.01311911145846049
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,32,2560,0.005583111196756363
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,32,1024,0.004058666527271271
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,32,2560,0.012964444855848948
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,32,2048,0.005009777843952179
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,32,1536,0.004453333301676644
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,32,2048,0.012511110968059964
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,32,512,0.011415999796655444
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,32,1536,0.012081777883900536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,32,768,0.0034248888906505373
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,32,1024,0.011733333269755045
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,32,768,0.011902222202883827
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,32,512,0.003167111012670729
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,32,256,0.002785777880085839
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,32,128,0.0026577777332729767
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,32,256,0.011531555818186866
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,32,64,0.002551111082235972
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,32,128,0.01034311122364468
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,32,64,0.009898666706350114
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,2,32,32,0.0025457777082920074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,2,32,32,0.009753777749008602
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,65536,12288,0.19458933671315512
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,65536,10240,0.1662035518222385
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,65536,16384,0.25239109992980957
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,65536,10240,0.3260248767005073
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,65536,12288,0.3901582294040256
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,65536,12288,0.24116267098320854
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,65536,8192,0.1367688841289944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,65536,16384,0.29939643541971844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,65536,16384,0.5223315556844076
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,65536,7168,0.1207760042614407
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,65536,10240,0.20968711376190186
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,65536,8192,0.2615075641208225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,65536,6144,0.11737422148386638
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,65536,7168,0.22857599788241914
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,65536,8192,0.16959910922580293
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,65536,5120,0.10073866446812947
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,65536,6144,0.19640444384680855
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,65536,5120,0.16562310854593912
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,65536,4096,0.07614044348398845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,65536,4096,0.15127289295196533
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,65536,3584,0.07604800330268012
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,65536,3584,0.11537599563598633
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,65536,7168,0.15218755933973524
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,65536,6144,0.13586044311523438
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,65536,3072,0.06319910950130887
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,65536,3072,0.1008453302913242
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,65536,2560,0.055345777008268565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,65536,5120,0.11668533749050564
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,65536,2048,0.06743911239835951
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,65536,2048,0.04803555541568332
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,65536,2560,0.08358133501476711
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,65536,1536,0.04231999980078804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,65536,1536,0.05172266562779745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,65536,4096,0.10059911012649536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,65536,1024,0.035798221826553345
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,65536,1024,0.036481777826944985
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,65536,3584,0.09138577514224583
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,65536,768,0.03380089004834493
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,65536,768,0.027411555250485737
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,65536,512,0.019520888725916546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,65536,512,0.030187557140986126
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,65536,256,0.011230222052998014
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,65536,256,0.02588533361752828
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,65536,3072,0.08104533619350858
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,65536,128,0.00737955586777793
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,65536,2560,0.07223288880454169
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,65536,128,0.023549333214759827
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,65536,64,0.005495111147562663
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,65536,128,0.026199110680156287
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,65536,64,0.020808888806237113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,65536,32,0.004381333374314838
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,65536,32,0.020806221498383414
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,65536,2048,0.060919112629360624
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,65536,1536,0.05632266733381483
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,65536,1024,0.0440977778699663
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,65536,768,0.04077600108252631
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,65536,512,0.03582311007711623
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,65536,256,0.03016888764169481
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,16384,16384,0.07606666617923312
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,16384,16384,0.10147111283408271
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,16384,12288,0.06028978029886881
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,16384,12288,0.09488177961773342
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,16384,10240,0.053370667828453906
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,16384,16384,0.1259582175148858
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,16384,12288,0.08366400003433228
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,16384,65536,0.2587653266059028
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,16384,10240,0.0799200005001492
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,16384,10240,0.07364355855517916
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,16384,8192,0.050042665667004056
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,16384,8192,0.06446399953630236
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,16384,7168,0.04195644458134969
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,16384,8192,0.06203555398517185
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,16384,7168,0.05795555644565158
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,16384,7168,0.056735111607445605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,16384,6144,0.03808088766203986
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,16384,6144,0.053789334164725415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,16384,5120,0.034428444173600935
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,16384,5120,0.05051022105746799
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,16384,6144,0.04929511083496941
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,16384,4096,0.030413332912656996
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,16384,4096,0.03423022230466207
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,16384,5120,0.04201066825124952
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,16384,4096,0.043622222211625844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,16384,3584,0.029329776763916016
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,16384,3584,0.030485332012176514
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,16384,3584,0.042872889174355396
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,16384,3072,0.026761778526835974
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,16384,3072,0.039510223600599505
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,16384,3072,0.026367111338509455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,16384,2048,0.022105778257052105
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,16384,2560,0.026569777064853247
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,16384,2560,0.03828888800409105
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,16384,2560,0.0225600004196167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,16384,2048,0.033696889877319336
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,16384,1536,0.03308444552951389
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,16384,1536,0.014872888724009195
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,16384,2048,0.018574222922325134
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,16384,1536,0.020635555187861126
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,16384,1024,0.018191999859280057
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,16384,768,0.00907022257645925
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,16384,768,0.018033777674039204
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,16384,1024,0.030024889442655776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,16384,512,0.007123555574152205
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,16384,1024,0.010991999672518836
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,16384,768,0.02944444285498725
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,16384,512,0.016602666841612924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,16384,256,0.005076444397370021
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,16384,512,0.028633776638242934
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,16384,256,0.013094222380055321
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,16384,256,0.023212444451120164
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,16384,128,0.0034808889031410217
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,16384,128,0.011888888974984487
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,16384,64,0.0031395554542541504
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,16384,128,0.021773333350817364
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,16384,64,0.01187111106183794
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,16384,32,0.003119999956753519
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,16384,32,0.011706666813956367
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,12288,16384,0.061830222606658936
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,16384,65536,0.5549377865261501
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,16384,65536,0.3102373282114665
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,12288,16384,0.10291022062301636
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,12288,16384,0.10895111163457234
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,12288,12288,0.07794400056203206
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,12288,12288,0.05785777833726671
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,12288,10240,0.044603556394577026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,12288,10240,0.07185866435368855
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,12288,10240,0.06819555494520399
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,12288,12288,0.0828639997376336
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,12288,8192,0.052689777480231396
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,12288,65536,0.200619551870558
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,12288,8192,0.03903288973702325
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,12288,7168,0.03560622202025519
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,12288,8192,0.06145333581500583
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,12288,7168,0.062069336573282875
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,12288,7168,0.048400001393424145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,12288,6144,0.03300088975164626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,12288,6144,0.056107554170820445
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,12288,5120,0.03177244464556376
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,12288,6144,0.04075466593106588
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,12288,5120,0.034479111433029175
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,12288,4096,0.02737599942419264
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,12288,5120,0.05093422200944689
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,12288,65536,0.3242746724022759
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,12288,4096,0.04507022102673849
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,12288,3584,0.026276444395383198
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,12288,3584,0.04198844565285576
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,12288,2560,0.02216533323129018
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,12288,3584,0.024953777591387432
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,12288,4096,0.02867199977238973
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,12288,3072,0.0247022228108512
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,12288,3072,0.03913066784540812
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,12288,2560,0.037308444579442344
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,12288,3072,0.022190221481853064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,12288,2048,0.02100533412562476
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,12288,2560,0.018784888916545443
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,12288,2048,0.03369777732425266
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,12288,1536,0.019690665933820937
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,12288,1024,0.009479111267460717
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,12288,1024,0.017486222916179232
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,12288,1536,0.012856889102194043
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,12288,2048,0.01571377780702379
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,12288,1536,0.031883554326163396
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,12288,768,0.008441777692900764
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,12288,1024,0.028845333390765723
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,12288,512,0.01425066590309143
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,12288,512,0.026932444837358262
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,12288,768,0.017775999175177682
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,12288,512,0.006280888699822956
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,12288,768,0.02832977639304267
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,12288,256,0.012464889221721225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,12288,128,0.0033440000067154565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,12288,256,0.0038702223036024305
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,12288,128,0.012075555821259817
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,12288,256,0.023044443792766992
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,12288,64,0.003020444470975134
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,12288,128,0.02036977807680766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,12288,32,0.003042666655447748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,12288,64,0.011591999895042844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,12288,32,0.011614222493436603
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,10240,16384,0.05496088994873894
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,10240,16384,0.083405335744222
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,10240,12288,0.06853155295054118
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,10240,12288,0.04543733265664843
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,10240,16384,0.08948444657855564
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,10240,12288,0.06658933560053508
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,10240,10240,0.04052622119585673
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,10240,10240,0.057830221123165555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,10240,10240,0.06100710895326403
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,10240,8192,0.03558044301138984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,10240,8192,0.04718488785955641
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,10240,8192,0.053229331970214844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,10240,7168,0.033061332172817655
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,10240,7168,0.03644888930850559
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,10240,7168,0.05022133390108744
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,10240,65536,0.17355822192298043
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,10240,6144,0.030637333790461223
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,10240,6144,0.03313244382540385
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,10240,6144,0.04576799935764737
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,10240,5120,0.029934220843844946
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,10240,5120,0.042993777328067355
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,10240,5120,0.028255999088287354
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,10240,4096,0.025430222352345783
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,10240,4096,0.03859733210669623
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,12288,65536,0.398801777097914
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,10240,4096,0.023656888140572443
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,10240,3584,0.024051555328898962
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,10240,3072,0.022075555390781824
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,10240,65536,0.2370684411790636
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,10240,3584,0.020689777202076383
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,10240,3584,0.0378115541405148
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,10240,3072,0.034493333763546415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,10240,2560,0.022056889202859666
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,10240,2048,0.019606222709019978
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,10240,2560,0.03379111157523261
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,10240,2560,0.015953777564896476
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,10240,1536,0.010445333189434476
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,10240,2048,0.03141511148876614
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,10240,1536,0.018923555811246235
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,10240,3072,0.02014133334159851
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,10240,2048,0.013173333472675748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,10240,1024,0.008081778056091731
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,10240,1536,0.03049422303835551
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,10240,1024,0.01705511079894172
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,10240,768,0.0070186663005087115
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,10240,1024,0.028757333755493164
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,10240,768,0.01630577776167128
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,10240,512,0.0058248887459437055
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,10240,768,0.02797422144148085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,10240,512,0.014073777529928418
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,10240,256,0.0035964445107513
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,10240,512,0.025568000144428674
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,10240,256,0.012669333153300814
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,10240,128,0.0032542221662071017
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,10240,256,0.022279111875428095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,10240,128,0.011648889217111798
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,10240,64,0.002969777832428614
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,10240,128,0.02075111038155026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,10240,32,0.0029502221279674103
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,10240,64,0.011866666376590729
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,10240,32,0.011372444530328115
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,8192,16384,0.04530310961935255
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,8192,16384,0.07486310932371351
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,8192,12288,0.051743110020955406
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,8192,12288,0.037541333172056414
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,8192,16384,0.06663555569118924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,8192,12288,0.05818577607472738
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,8192,10240,0.033779554896884494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,8192,10240,0.05473777651786804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,8192,10240,0.04235733217663235
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,8192,8192,0.029868443806966145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,8192,8192,0.03444622291458978
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,8192,8192,0.048496001296573214
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,8192,7168,0.027832888894610937
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,8192,65536,0.1371119949552748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,8192,7168,0.030293332205878362
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,8192,7168,0.0480951103899214
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,8192,6144,0.02595377796226078
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,8192,6144,0.026869333452648584
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,8192,5120,0.023894222246276006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,8192,6144,0.04399466514587402
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,8192,5120,0.022784888744354248
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,8192,5120,0.039847112364239164
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,8192,4096,0.020791111720932853
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,8192,3584,0.02051199972629547
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,8192,4096,0.019114666514926486
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,8192,4096,0.03615999884075589
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,10240,65536,0.3727413283454047
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,8192,3584,0.03585244549645318
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,8192,3072,0.018838221828142803
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,8192,3072,0.032927112446890935
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,8192,2560,0.018415111634466384
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,8192,3584,0.0170168893204795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,8192,2560,0.03267822331852383
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,8192,3072,0.015134221977657743
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,8192,2048,0.01681422193845113
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,8192,1536,0.009259555902745988
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,8192,2048,0.030101332399580214
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,8192,1536,0.016280000408490498
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,8192,2560,0.013194666968451606
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,8192,2048,0.011115555961926779
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,8192,1024,0.007158222297827403
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,8192,1024,0.014399111270904541
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,8192,1536,0.030441777573691473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,8192,768,0.012990222209029727
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,8192,768,0.006265777680608962
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,8192,1024,0.02753244505988227
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,8192,512,0.005335111171007156
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,8192,768,0.02681866619322035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,8192,512,0.012619555824332766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,8192,256,0.0035413333939181436
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,8192,512,0.023864888482623633
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,8192,256,0.010986666712496014
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,8192,128,0.0031475555151700974
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,8192,256,0.02142933342191908
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,8192,128,0.010600888894663917
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,8192,64,0.0029484445436133277
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,8192,32,0.0029120000286234748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,8192,128,0.020554666717847187
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,8192,64,0.010276444256305695
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,8192,65536,0.20784000555674234
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,8192,32,0.010013333625263637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,7168,16384,0.04199999901983473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,7168,16384,0.06995999813079834
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,8192,65536,0.26207465595669216
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,7168,12288,0.03526488939921061
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,7168,16384,0.05953422519895765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,7168,12288,0.0504755543337928
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,7168,10240,0.031680001152886286
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,7168,12288,0.05900888972812229
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,7168,10240,0.04028088847796122
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,7168,10240,0.05345955491065979
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,7168,8192,0.028156446086035833
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,7168,8192,0.034965333011415266
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,7168,65536,0.12531467278798422
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,7168,8192,0.04745599958631727
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,7168,7168,0.02983377708329095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,7168,7168,0.02680177821053399
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,7168,7168,0.04647733436690437
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,7168,6144,0.024717332588301763
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,7168,6144,0.02739022175470988
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,7168,6144,0.04310933417744107
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,7168,5120,0.022501332892311945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,7168,5120,0.03783822059631348
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,7168,4096,0.020483555065260995
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,7168,5120,0.023102222217453852
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,7168,3584,0.019352000620630052
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,7168,65536,0.20870844523111978
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,7168,4096,0.036566221051745944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,7168,4096,0.019819556011093985
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,7168,3584,0.034767998589409724
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,7168,3072,0.015949333707491558
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,7168,3072,0.01828444500764211
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,7168,2560,0.017645332548353408
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,7168,3072,0.033333334657881
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,7168,2048,0.01148444414138794
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,7168,2048,0.01648355523745219
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,7168,2560,0.03218222326702542
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,7168,1536,0.008642666869693333
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,7168,3584,0.01756000022093455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,7168,1024,0.006790222393141852
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,7168,2048,0.02978666623433431
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,7168,1536,0.02915999955601162
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,7168,1536,0.016083555089102853
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,7168,2560,0.01236088905069563
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,7168,1024,0.013789333403110504
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,7168,768,0.00591822216908137
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,7168,1024,0.02765688962406582
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,7168,512,0.00452622233165635
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,7168,768,0.012747555143303342
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,7168,768,0.024446222517225478
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,7168,512,0.012469333079126147
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,7168,256,0.003403555601835251
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,7168,512,0.025047111842367385
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,7168,256,0.011375111010339526
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,7168,128,0.0031386667655573953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,7168,256,0.021838222940762837
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,7168,128,0.010087110930018956
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,7168,64,0.002834666727317704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,7168,128,0.021186666356192693
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,7168,64,0.010058666268984476
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,7168,32,0.00286311118139161
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,7168,32,0.009952888720565373
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,6144,16384,0.039057777987586126
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,7168,65536,0.2308195432027181
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,6144,12288,0.04333333174387614
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,6144,16384,0.07224710782368977
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,6144,12288,0.03361777795685662
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,6144,16384,0.05761866437064277
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,6144,10240,0.030588444736268785
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,6144,12288,0.057983113659752734
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,6144,10240,0.05462755428420173
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,6144,10240,0.03711466656790839
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,6144,8192,0.027172444595230952
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,6144,8192,0.04616711205906338
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,6144,7168,0.025136000580257837
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,6144,7168,0.026639110512203638
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,6144,8192,0.029882666137483384
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,6144,65536,0.11202400260501438
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,6144,6144,0.023158222436904907
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,6144,5120,0.02104533380932278
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,6144,7168,0.04496799906094869
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,6144,6144,0.02164799968401591
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,6144,5120,0.03862933317820231
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,6144,4096,0.01924711134698656
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,6144,6144,0.04425866736306084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,6144,4096,0.034697777695126004
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,6144,3584,0.019632889164818656
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,6144,3584,0.034019556310441755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,6144,4096,0.01664799948533376
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,6144,5120,0.019949333535300363
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,6144,3072,0.017679999272028606
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,6144,3584,0.01499288943078783
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,6144,2560,0.01109777804878023
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,6144,3072,0.012640888492266336
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,6144,3072,0.03276888860596551
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,6144,2560,0.017332444588343304
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,6144,2048,0.010035555395815108
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,6144,65536,0.1980631086561415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,6144,2048,0.01590222285853492
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,6144,1536,0.008057778080304464
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,6144,2560,0.032822221517562866
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,6144,1536,0.015666666958067153
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,6144,2048,0.030538668235143025
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,6144,1024,0.006342222293217977
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,6144,1024,0.012368888490729861
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,6144,1536,0.029091556866963703
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,6144,768,0.005631111148330901
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,6144,1024,0.02752799954679277
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,6144,768,0.012891555825869242
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,6144,512,0.003929777691761653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,6144,768,0.025820444027582806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,6144,512,0.012201777762836881
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,6144,256,0.003362666608558761
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,6144,512,0.02480977773666382
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,6144,256,0.011321778098742167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,6144,128,0.003034666594531801
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,6144,256,0.021548445026079815
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,6144,128,0.010131555298964182
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,6144,64,0.0028533333291610083
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,6144,128,0.021408889028761122
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,6144,32,0.009959999885823991
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,6144,32,0.0027599999060233435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,6144,64,0.010463111102581024
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,5120,16384,0.035327110025617815
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,5120,16384,0.06674577792485555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,5120,16384,0.04790044493145413
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,6144,65536,0.19611822234259713
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,5120,12288,0.030271109607484605
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,5120,12288,0.03688800003793504
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,5120,10240,0.029290666182835896
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,5120,12288,0.055664890342288546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,5120,8192,0.024812445044517517
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,5120,10240,0.050923556089401245
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,5120,10240,0.03164711263444688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,5120,8192,0.04456444581349691
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,5120,8192,0.026559111144807603
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,5120,7168,0.02345866627163357
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,5120,7168,0.04316799839337667
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,5120,6144,0.021300445000330608
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,5120,7168,0.023846222294701472
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,5120,6144,0.0399893323580424
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,5120,5120,0.019860444797409903
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,5120,5120,0.03730666637420654
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,5120,6144,0.018631110588709515
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,5120,4096,0.01865955524974399
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,5120,5120,0.015978667471143935
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,5120,4096,0.013844444519943662
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,5120,4096,0.03635377685228983
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,5120,3584,0.018152889278199937
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,5120,65536,0.09915733337402344
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,5120,3072,0.0107351111041175
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,5120,3584,0.033032000064849854
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,5120,3584,0.012307555311255984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,5120,3072,0.01715377800994449
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,5120,2560,0.009732444253232744
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,5120,3072,0.032151109642452665
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,5120,2560,0.016725333200560678
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,5120,2048,0.008477333519193862
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,5120,2560,0.03229066729545593
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,5120,1536,0.0074551113777690465
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,5120,2048,0.015923556354310777
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,5120,1536,0.014111111561457316
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,5120,2048,0.03013955553372701
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,5120,1024,0.006005333529578314
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,5120,768,0.005114666703674528
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,5120,1536,0.02901955445607503
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,5120,1024,0.012254222399658628
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,5120,1024,0.024801777468787298
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,5120,768,0.012639111114872826
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,5120,512,0.0037173334923055437
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,5120,768,0.024106666445732117
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,5120,512,0.011961778004964193
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,5120,256,0.0032400000426504347
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,5120,65536,0.17557422320048013
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,5120,512,0.024104000793562994
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,5120,256,0.011218667030334473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,5120,128,0.003030222323205736
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,5120,256,0.022642667094866436
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,5120,128,0.010036444498433007
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,5120,64,0.002784888984428512
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,5120,128,0.02032088902261522
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,5120,64,0.009950222240553962
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,5120,32,0.00281866660548581
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,5120,32,0.00981600003110038
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,4096,16384,0.030671109755833943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,5120,65536,0.18921155399746367
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,4096,12288,0.029976000388463337
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,4096,16384,0.0673182209332784
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,4096,16384,0.038592000802357994
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,4096,12288,0.026698667142126296
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,4096,12288,0.05470400055249532
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,4096,10240,0.05007466673851013
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,4096,10240,0.02460711035463545
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,4096,8192,0.04239733351601494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,4096,10240,0.026177777184380427
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,4096,8192,0.02188088827662998
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,4096,8192,0.022078222698635522
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,4096,7168,0.02068711154990726
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,4096,7168,0.04148533278041416
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,4096,6144,0.01964977714750502
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,4096,6144,0.039078222380744085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,4096,65536,0.07835555738872953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,4096,6144,0.017642666896184284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,4096,7168,0.020198222663667467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,4096,5120,0.0181688881582684
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,4096,5120,0.015495111544926962
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,4096,5120,0.0362097786532508
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,4096,4096,0.017232000827789307
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,4096,3584,0.012503999802801343
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,4096,3584,0.016886222693655226
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,4096,4096,0.03622577918900384
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,4096,4096,0.013375110924243927
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,4096,3072,0.011424000064531961
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,4096,3584,0.03329599897066752
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,4096,3072,0.016233777006467182
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,4096,2560,0.008469333251317343
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,4096,3072,0.03152088986502753
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,4096,2560,0.015968888998031616
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,4096,2048,0.007445333732499017
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,4096,2560,0.03163377775086297
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,4096,2048,0.014678221609857349
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,4096,1536,0.006538666784763336
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,4096,2048,0.029320889049106177
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,4096,1536,0.01421333352724711
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,4096,1024,0.005508444375462002
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,4096,1024,0.011830222275522021
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,4096,1536,0.02924444609218174
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,4096,768,0.004000888930426704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,4096,768,0.012516444755925072
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,4096,1024,0.024315555890401203
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,4096,768,0.02394666605525547
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,4096,512,0.0036915557252036203
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,4096,512,0.01201777739657296
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,4096,256,0.0032248888164758682
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,4096,512,0.02369155486424764
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,4096,65536,0.177511109246148
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,4096,256,0.010536889235178629
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,4096,128,0.0029191111938820947
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,4096,256,0.0225582222143809
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,4096,64,0.009747555686367882
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,4096,128,0.010086221827401055
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,4096,64,0.002708444371819496
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,4096,128,0.02031822171476152
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,4096,32,0.0026977778308921387
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,4096,32,0.010219555762079027
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,4096,65536,0.13196444511413574
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3584,16384,0.029049777322345312
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,3584,16384,0.06046399805280897
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3584,16384,0.033302221033308245
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3584,12288,0.026138666603300307
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3584,12288,0.026661333110597398
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,3584,12288,0.05169511172506544
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3584,8192,0.020791111720932853
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,3584,10240,0.049025777313444346
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3584,10240,0.022669333550665114
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,3584,8192,0.04142755601141188
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3584,7168,0.01757777730623881
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3584,7168,0.019671110643280875
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3584,8192,0.019620445039537217
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3584,10240,0.023044443792766992
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,3584,7168,0.04059911105367873
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3584,6144,0.015754666593339708
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3584,6144,0.01869066721863217
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,3584,6144,0.03793599870469835
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3584,5120,0.017644445101420086
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3584,65536,0.0706204440858629
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3584,4096,0.01259022206068039
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,3584,5120,0.03623466690381368
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3584,4096,0.016819554898473952
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3584,3584,0.011979555918110741
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,3584,4096,0.033358222908443876
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3584,3584,0.016604445046848722
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3584,5120,0.014296000202496847
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3584,3072,0.01088711122671763
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,3584,3584,0.03303111261791653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3584,3072,0.015979554918077257
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3584,2560,0.008059555457697975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,3584,3072,0.03125155634350247
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3584,2560,0.01574222246805827
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3584,2048,0.0069919998447100324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,3584,2560,0.03129511078198751
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3584,2048,0.013704000247849358
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3584,1536,0.006287999865081575
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,3584,2048,0.02852355440457662
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3584,1536,0.012975999878512489
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,3584,1536,0.026206221845414903
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3584,1024,0.005161777966552311
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3584,1024,0.011994666523403592
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3584,768,0.004254222330119875
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,3584,1024,0.024710221423043147
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3584,768,0.012464000119103326
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,3584,768,0.02420888841152191
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3584,512,0.003549333247873518
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3584,512,0.0119991112086508
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3584,256,0.0031351111829280853
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,3584,65536,0.16530577341715494
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,3584,512,0.023593778411547344
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3584,256,0.01108977778090371
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3584,128,0.002867555452717675
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,3584,256,0.02165422174665663
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3584,128,0.009983110758993361
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3584,64,0.002667555585503578
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,3584,128,0.02032444377740224
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3584,64,0.009825777676370408
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3584,32,0.0027226667023367355
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3584,32,0.010234666367371878
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3584,65536,0.1192586686876085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3072,16384,0.027671999401516382
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,3072,16384,0.06093688805898031
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3072,16384,0.03178933262825012
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3072,12288,0.025230222278171118
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3072,10240,0.022129777404997084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3072,12288,0.023544889357354905
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,3072,12288,0.050316443045934044
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3072,8192,0.01979733341270023
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,3072,10240,0.04591200086805555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3072,10240,0.022008889251285132
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3072,7168,0.018519111805491976
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,3072,8192,0.03987377882003784
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,3072,7168,0.040391110711627536
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3072,6144,0.01816177699300978
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3072,8192,0.017979555659823947
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3072,7168,0.017210667332013447
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,3072,6144,0.037350220812691584
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3072,5120,0.013818666338920593
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3072,5120,0.017487999465730455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3072,6144,0.015410666664441427
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3072,4096,0.011936888926559024
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,3072,5120,0.03529422150717841
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3072,4096,0.016605334149466622
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3072,3584,0.011172444456153445
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,3072,4096,0.03382488754060533
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3072,65536,0.06545333067576091
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,3072,65536,0.16167110866970485
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3072,3072,0.010248888697889116
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3072,3584,0.01644799941115909
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3072,3072,0.015467555986510383
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,3072,3584,0.03283111254374186
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3072,2560,0.007768889268239339
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,3072,3072,0.031502223677105375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3072,2560,0.015044444137149386
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3072,2048,0.006817777951558431
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,3072,2560,0.030386666456858318
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3072,2048,0.012631999949614206
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3072,1536,0.006136000156402588
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,3072,2048,0.027927110592524212
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3072,1536,0.013497778111033969
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3072,1024,0.004638222356637319
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,3072,1536,0.025796444879637823
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3072,1024,0.011943110989199745
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3072,768,0.00388355553150177
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,3072,1024,0.02365866634580824
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3072,768,0.012510221865442065
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3072,512,0.0035146665241983202
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,3072,768,0.024826667375034753
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3072,512,0.011786667009194693
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3072,256,0.003141333245568805
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,3072,512,0.023804444405767653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3072,128,0.0028577778074476454
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3072,256,0.010496888723638324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,3072,256,0.02241155505180359
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3072,128,0.010248888697889116
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,3072,128,0.020326221982638042
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3072,64,0.002712888850106133
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3072,64,0.009927111367384592
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3072,32,0.0026977778308921387
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,3072,32,0.009655111365848118
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,3072,65536,0.10345866945054795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2560,16384,0.026735999517970618
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2560,12288,0.022664889693260193
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,2560,12288,0.05044088761011759
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2560,10240,0.02102577851878272
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,2560,16384,0.06047022342681885
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2560,12288,0.020769778225156996
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2560,10240,0.018360889620251127
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2560,8192,0.01901333365175459
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,2560,10240,0.04531022244029575
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2560,16384,0.026735110415352717
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,2560,8192,0.04134488768047757
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2560,7168,0.01889244384235806
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2560,8192,0.015623999966515435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,2560,7168,0.04070311122470432
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2560,6144,0.012860444684823355
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2560,6144,0.017824000782436795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2560,7168,0.014552889598740472
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,2560,6144,0.038183109627829656
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2560,5120,0.01181333346499337
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2560,5120,0.017210667332013447
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2560,4096,0.010460444622569613
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,2560,5120,0.035848001639048256
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2560,4096,0.01628977722591824
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2560,3584,0.009934222532643212
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,2560,4096,0.03396888905101352
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2560,3584,0.016025778320100572
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2560,3072,0.009192888935407003
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,2560,3584,0.032840887705485024
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2560,3072,0.015058666467666626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2560,2560,0.007311111523045435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2560,65536,0.06403377983305189
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,2560,3072,0.03086044390996297
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2560,2560,0.013914666242069669
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,2560,65536,0.1450551086001926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2560,2048,0.006659555352396435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,2560,2560,0.029127111037572224
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2560,2048,0.012799111505349478
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2560,1536,0.005662222289376789
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,2560,2048,0.026591110560629103
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2560,1536,0.012979555461141797
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2560,1024,0.004392888810899523
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,2560,1536,0.02552355494764116
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2560,1024,0.011692444483439127
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2560,768,0.0038968887594011095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,2560,1024,0.024114666713608637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2560,768,0.012372444073359171
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2560,512,0.0034471110751231513
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,2560,768,0.02395733363098568
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2560,512,0.01163999984661738
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2560,256,0.003109333415826162
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2560,128,0.0028044444819291434
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,2560,512,0.024217777782016333
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2560,256,0.010638222098350525
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,2560,256,0.021371554997232225
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2560,128,0.009778666827413771
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2560,64,0.0026640000028742682
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,2560,128,0.019967110620604623
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2560,64,0.010132444401582083
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2560,32,0.0026791110220882627
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2560,65536,0.09321422047085232
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2560,32,0.009895111123720804
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2048,16384,0.024826667375034753
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,2048,16384,0.06163644128375583
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2048,12288,0.021853332718213398
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2048,12288,0.017825777331988018
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2048,16384,0.02297244469324748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,2048,12288,0.047260443369547524
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2048,10240,0.02036355601416694
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2048,10240,0.016492444607946608
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2048,65536,0.05977866384718153
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,2048,10240,0.04711022310786777
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2048,8192,0.01886311173439026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2048,7168,0.013441777891582914
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,2048,8192,0.0410515566666921
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2048,7168,0.017992888887723286
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2048,8192,0.014226666755146451
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2048,6144,0.012092444631788464
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,2048,7168,0.039656000004874334
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,2048,65536,0.16436978181203207
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2048,6144,0.01752088963985443
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2048,5120,0.010741333166758219
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,2048,6144,0.03714044557677375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2048,5120,0.016899555921554565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2048,4096,0.009573332965373993
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,2048,5120,0.03598489032851325
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2048,4096,0.01521511044767168
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2048,3584,0.008996444443861643
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,2048,4096,0.03243644369973077
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2048,3584,0.01574133336544037
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2048,3072,0.008613333106040955
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,2048,3584,0.03202133377393087
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2048,3072,0.01463822192615933
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2048,2560,0.007135111424658034
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,2048,3072,0.02773333258099026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2048,2560,0.014262222581439547
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2048,2048,0.0063635553750726916
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2048,65536,0.07245333327187432
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,2048,2560,0.028103109863069322
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,2048,1536,0.025861332813898723
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2048,2048,0.013177778157922955
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2048,1536,0.004839111119508743
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,2048,2048,0.02667288813326094
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2048,1536,0.012873777912722694
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2048,1024,0.004032888760169347
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2048,1024,0.011702222128709158
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2048,768,0.003710222327046924
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,2048,1024,0.023927999867333308
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2048,768,0.01240888900227017
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,2048,512,0.024358222881952923
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2048,512,0.003387555480003357
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,2048,768,0.02420622275935279
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2048,512,0.011945777469211154
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2048,256,0.0030622221529483795
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2048,256,0.010661332971519895
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,2048,256,0.021588444709777832
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2048,128,0.002789333255754577
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2048,128,0.010253333383136326
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2048,64,0.0026382222357723448
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,2048,128,0.021213332811991375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,2048,32,0.0026702222724755607
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2048,64,0.009835555321640438
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,2048,32,0.009734222458468543
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1536,16384,0.018343999981880188
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1536,16384,0.023618666662110224
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1536,12288,0.020976000361972388
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,1536,16384,0.05606044663323296
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1536,10240,0.013764444324705334
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,1536,12288,0.04865511258443197
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1536,10240,0.019857777489556205
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1536,12288,0.015191111299726697
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1536,8192,0.011887110769748688
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,1536,10240,0.044455111026763916
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1536,8192,0.018398221996095445
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1536,7168,0.010764444039927589
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,1536,8192,0.04020977682537503
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1536,7168,0.01777777738041348
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1536,6144,0.009964444571071202
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,1536,7168,0.03842133283615112
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1536,6144,0.017172444197866652
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1536,5120,0.009298666483826106
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,1536,6144,0.035840001371171736
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1536,5120,0.01625777781009674
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1536,4096,0.008642666869693333
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,1536,5120,0.033943110042148165
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1536,4096,0.014465777410401238
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1536,3584,0.008456889126035903
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,1536,4096,0.031558222240871854
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1536,3584,0.014575999644067554
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1536,65536,0.055768887201944985
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1536,3072,0.008104888929261101
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,1536,3584,0.02997777859369914
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,1536,3072,0.027453333139419556
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1536,3072,0.01349688900841607
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1536,2560,0.009571555587980483
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1536,2560,0.01349955548842748
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1536,2048,0.008208889100286696
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,1536,2560,0.02815022071202596
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,1536,65536,0.14116533597310385
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1536,2048,0.012617777619096967
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1536,1536,0.006581333362393909
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,1536,2048,0.026153778036435444
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1536,1536,0.012715555727481842
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1536,1024,0.005201777650250329
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,1536,1536,0.026191110412279766
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,1536,1024,0.023990222149425085
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1536,1024,0.011763555308183035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1536,768,0.004481777962711123
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1536,65536,0.05357244279649523
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1536,768,0.012133333418104382
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1536,512,0.0039057777159743835
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1536,512,0.011902222202883827
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,1536,768,0.024499555428822834
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1536,256,0.003293333368168937
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,1536,512,0.023423999547958374
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1536,256,0.010739555789364709
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1536,128,0.003000000077817175
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,1536,256,0.021456888980335657
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1536,128,0.009889778163697984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1536,64,0.002834666727317704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,1536,128,0.021057777934604224
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1536,64,0.009786666267448002
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1536,32,0.00282222218811512
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1536,32,0.010077333284748925
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1024,16384,0.021746666895018682
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1024,12288,0.012325333224402534
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,1024,16384,0.04974222183227539
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1024,12288,0.01958488921324412
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,1024,12288,0.04283644424544441
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1024,10240,0.011110222174061669
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1024,16384,0.014142221874660917
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1024,10240,0.018411555223994784
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,1024,10240,0.04110488957828946
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1024,8192,0.010072888599501716
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1024,8192,0.017633777525689866
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1024,65536,0.04630044433805677
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1024,7168,0.00904977818330129
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,1024,8192,0.037257777320014104
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1024,7168,0.016117332710160148
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1024,6144,0.008371555142932469
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,1024,7168,0.03507644600338406
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1024,6144,0.015218666858143277
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1024,5120,0.00832088871134652
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,1024,6144,0.03472266594568888
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1024,5120,0.014656888114081489
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,1024,65536,0.11887910630967881
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1024,4096,0.007871111234029135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,1024,5120,0.030985779232449
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1024,4096,0.013960889644092984
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1024,3584,0.008049777812427944
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1024,65536,0.03918755716747708
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,1024,4096,0.029350221157073975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1024,3584,0.014273777604103088
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1024,3072,0.007015110717879401
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,1024,3584,0.028390222125583228
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1024,3072,0.013461333182122974
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1024,2560,0.014271999398867289
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,1024,3072,0.026517333255873785
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1024,2560,0.013542222479979197
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1024,2048,0.012836444709036084
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,1024,2560,0.027094221777386133
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1024,2048,0.012638222012254926
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1024,1536,0.01037333326207267
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,1024,2048,0.025455999705526564
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1024,1536,0.01276177747382058
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1024,1024,0.007821333077218797
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,1024,1536,0.024566221568319533
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1024,1024,0.011942221886581846
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1024,768,0.006513777706358168
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,1024,1024,0.02386755579047733
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1024,768,0.012248000337017907
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1024,512,0.0052488889131281115
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,1024,768,0.024365334047211543
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1024,512,0.01182577759027481
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1024,256,0.003932444585694207
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,1024,512,0.022993778189023335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1024,256,0.011030221978823343
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1024,128,0.0033484444850020935
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,1024,256,0.02225422196918064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1024,128,0.010115555591053432
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1024,64,0.003032888803217146
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,1024,128,0.020252444677882724
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1024,64,0.009850666754775578
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,1024,32,0.0030088888274298776
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,1024,32,0.00980888886584176
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,768,16384,0.011953777737087674
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,768,16384,0.02120977805720435
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,768,12288,0.009872888525327047
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,768,16384,0.04920977685186598
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,768,12288,0.01939911146958669
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,768,12288,0.04180711176660326
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,768,10240,0.009761778016885122
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,768,65536,0.10853866736094157
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,768,10240,0.01791555517249637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,768,8192,0.008767111433876885
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,768,10240,0.039064000050226845
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,768,65536,0.04461955693033007
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,768,8192,0.016519111063745286
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,768,7168,0.008288000192907121
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,768,8192,0.03374310996797349
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,768,7168,0.015467555986510383
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,768,6144,0.00796088907453749
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,768,6144,0.03032711148262024
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,768,7168,0.03275288807021247
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,768,6144,0.015081778168678284
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,768,5120,0.007819555699825287
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,768,5120,0.014490667316648694
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,768,4096,0.013991999957296582
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,768,4096,0.007248889240953658
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,768,3584,0.014157333307796054
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,768,65536,0.02956977817747328
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,768,5120,0.02896889050801595
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,768,3584,0.007156444092591603
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,768,4096,0.029021332661310833
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,768,3072,0.006788444601827198
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,768,3584,0.026983110441101923
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,768,3072,0.013361777696344586
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,768,2560,0.008534222013420528
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,768,3072,0.026531555586391028
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,768,2560,0.013538666897349887
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,768,2048,0.007351111206743453
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,768,2560,0.02572622232966953
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,768,2048,0.012677333421177335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,768,1536,0.0063688887490166565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,768,1536,0.012621333201726278
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,768,2048,0.024296000599861145
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,768,1024,0.004990222139490975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,768,1536,0.024932444095611572
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,768,1024,0.01178311142656538
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,768,768,0.004362666772471534
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,768,1024,0.023056000471115112
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,768,768,0.012359999948077731
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,768,512,0.003814222084151374
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,768,768,0.023787554767396715
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,768,512,0.011921777493423887
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,768,256,0.0032293332947625052
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,768,512,0.02377333409256405
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,768,256,0.010743111371994019
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,768,128,0.0029848888516426086
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,768,256,0.021216000119845074
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,768,128,0.009952888720565373
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,768,64,0.0027831111931138565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,768,128,0.02054577734735277
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,768,64,0.010124444133705562
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,768,32,0.0028008888992998335
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,768,32,0.009914666414260864
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,512,65536,0.04365066687266032
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,512,16384,0.009734222458468543
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,512,16384,0.020581333173645865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,512,12288,0.008664000365469191
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,512,16384,0.043567998541726
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,512,12288,0.018265777164035372
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,512,10240,0.008426666259765625
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,512,12288,0.03793155484729343
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,512,10240,0.017143110434214275
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,512,65536,0.09691022502051459
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,512,8192,0.007910221815109253
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,512,10240,0.03628977802064683
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,512,7168,0.015457777513398064
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,512,8192,0.015978667471143935
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,512,7168,0.0075253330998950535
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,512,8192,0.030505776405334473
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,512,65536,0.02324177821477254
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,512,7168,0.030185778935750324
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,512,6144,0.007205333146784041
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,512,6144,0.014924445086055331
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,512,5120,0.007156444092591603
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,512,6144,0.02886311213175456
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,512,5120,0.014467555615637036
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,512,4096,0.006435555716355641
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,512,5120,0.02697155541843838
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,512,4096,0.013967111706733704
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,512,3584,0.006872889068391588
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,512,4096,0.02663199934694502
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,512,2560,0.006157333238257303
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,512,3584,0.014152889450391134
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,512,3072,0.024911110599835713
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,512,3072,0.006592888798978593
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,512,3584,0.026661333110597398
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,512,3072,0.01352533366945055
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,512,2560,0.013467555244763693
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,512,2048,0.005504888792832692
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,512,2048,0.012673777838548025
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,512,2560,0.02513244416978624
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,512,1536,0.004868444469239977
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,512,2048,0.024102222588327196
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,512,1536,0.012758221891191272
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,512,1024,0.004135999828577042
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,512,1536,0.024977778395016987
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,512,1024,0.011844444606039258
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,512,768,0.0038195554580953387
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,512,1024,0.02292266653643714
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,512,768,0.012279999752839407
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,512,256,0.003168888803985384
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,512,512,0.02292266653643714
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,512,768,0.024293333292007446
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,512,512,0.003439999909864532
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,512,512,0.012064889073371887
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,512,256,0.0107342220014996
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,512,128,0.002862222285734283
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,512,256,0.021154666940371197
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,512,128,0.009829333258999718
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,512,64,0.0026622222115596137
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,512,128,0.020850666695170935
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,512,32,0.002720000015364753
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,512,64,0.010168888502650792
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,512,32,0.010001777774757808
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,256,65536,0.014371555712487964
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,256,65536,0.09053155448701646
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,256,16384,0.007887110941939885
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,256,65536,0.04274133178922865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,256,16384,0.019842666056421068
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,256,12288,0.007306666837798224
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,256,16384,0.03860977623197768
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,256,10240,0.006872889068391588
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,256,12288,0.017609778377744887
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,256,12288,0.034148444732030235
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,256,10240,0.016745777593718637
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,256,8192,0.006653333289755716
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,256,10240,0.03211377726660834
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,256,8192,0.015949333707491558
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,256,6144,0.006221333311663733
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,256,7168,0.006436444405052397
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,256,8192,0.030412445465723675
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,256,7168,0.015455111861228943
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,256,7168,0.028879112667507593
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,256,5120,0.006792888873153263
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,256,6144,0.014868444866604276
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,256,6144,0.027913779020309448
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,256,5120,0.014282666974597506
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,256,4096,0.0063564446237352155
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,256,5120,0.026899556318918865
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,256,4096,0.013975999421543546
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,256,3584,0.006783999916579988
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,256,4096,0.025920000341203477
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,256,3584,0.014189332723617554
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,256,3072,0.0063288890653186375
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,256,3584,0.02630222174856398
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,256,3072,0.01332711097266939
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,256,2560,0.006065777606434292
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,256,3072,0.02492977844344245
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,256,2560,0.013353777428468069
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,256,2048,0.005373333477311664
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,256,2560,0.025036444266637165
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,256,2048,0.012335999972290464
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,256,1536,0.004705777598751916
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,256,2048,0.023918223049905565
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,256,768,0.0036888888312710654
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,256,1536,0.012663111090660095
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,256,1024,0.003992889076471329
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,256,1536,0.02418400016095903
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,256,1024,0.01185155577129788
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,256,1024,0.022851554883850947
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,256,768,0.012372444073359171
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,256,512,0.0033271111961868075
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,256,768,0.023215111758973863
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,256,512,0.011770666473441653
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,256,256,0.002982222164670626
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,256,512,0.023690667417314317
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,256,256,0.010915555887752108
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,256,128,0.0027866665687825945
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,256,256,0.021261332763565913
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,256,128,0.010020444790522257
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,256,64,0.002607999990383784
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,256,128,0.019711110326978896
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,256,64,0.009664889011118147
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,256,32,0.0026195556339290407
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,256,32,0.010047111246320937
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,128,65536,0.01090666651725769
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,128,65536,0.04244533181190491
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,128,16384,0.006463999963468975
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,128,16384,0.019260444574885897
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,128,65536,0.08995288610458374
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,128,12288,0.006264888991912206
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,128,16384,0.03765866822666592
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,128,12288,0.017584888471497428
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,128,10240,0.006464889066086874
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,128,12288,0.034550223085615374
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,128,10240,0.016612443659040663
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,128,8192,0.0063697778516345555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,128,10240,0.03195377853181627
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,128,8192,0.015810666812790763
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,128,7168,0.006193777753247156
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,128,8192,0.0303422212600708
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,128,7168,0.01534222231970893
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,128,6144,0.006061333335108227
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,128,7168,0.028891556792789038
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,128,5120,0.014119999276267158
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,128,6144,0.014711111783981323
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,128,5120,0.006416888700591193
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,128,6144,0.02787999974356757
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,128,4096,0.006049777898523543
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,128,5120,0.027407111393080816
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,128,4096,0.013606222139464485
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,128,3584,0.0064479998416370815
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,128,4096,0.026449778013759192
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,128,3584,0.013642667068375481
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,128,3072,0.00619911112719112
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,128,3584,0.026294221480687458
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,128,3072,0.013080000049538083
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,128,2560,0.005688888745175467
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,128,3072,0.02488800055450863
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,128,2560,0.013218666944238873
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,128,2048,0.005048888838953442
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,128,2560,0.025072889195548162
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,128,2048,0.01238933371173011
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,128,1536,0.004453333301676644
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,128,2048,0.024650666448805068
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,128,1536,0.012583110895421771
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,128,1024,0.0037626665499475268
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,128,1536,0.0247706671555837
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,128,1024,0.011667555405033959
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,128,768,0.003469333259595765
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,128,1024,0.022820444570647344
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,128,768,0.012229333321253458
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,128,512,0.003302222324742211
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,128,768,0.02353777819209629
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,128,512,0.01200977795653873
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,128,256,0.0029475556479560006
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,128,512,0.022944889134830896
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,128,128,0.0027946666296985415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,128,256,0.010662222074137794
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,128,256,0.020584000481499564
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,128,128,0.01017955525053872
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,128,64,0.0026364444444576898
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8_block,1,128,128,0.020064888728989493
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,128,32,0.0026008888251251644
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,128,64,0.009653333160612319
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,128,32,0.010067555639478896
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,64,65536,0.008928889201747047
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,64,16384,0.006253333141406377
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,64,65536,0.041591111156675555
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,64,16384,0.019123555885420907
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,64,12288,0.006014222072230444
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,64,10240,0.00627377794848548
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,64,12288,0.017271111408869427
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,64,10240,0.01643200053109063
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,64,8192,0.006245333287451003
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,64,8192,0.015402666396564908
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,64,7168,0.006010666489601135
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,64,7168,0.015037332971890768
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,64,6144,0.005930666708283954
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,64,5120,0.006331555545330048
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,64,6144,0.01423466702302297
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,64,4096,0.005946666830115848
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,64,5120,0.013857777747843
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,64,3584,0.0063271112740039825
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,64,4096,0.013537777794731988
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,64,3072,0.006011555592219035
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,64,3584,0.013707555830478668
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,64,3072,0.012886222038004132
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,64,2560,0.005557333429654439
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,64,2560,0.012855110896958245
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,64,2048,0.005021333280536863
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,64,2048,0.012186666329701742
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,64,1536,0.004414222306675381
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,64,1024,0.003793777690993415
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,64,1536,0.012378666963842181
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,64,1024,0.011562666959232755
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,64,768,0.0034560000316964257
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,64,768,0.012163555456532372
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,64,512,0.0031902222997612427
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,64,256,0.0029048888633648553
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,64,512,0.011666666302416058
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,64,256,0.010840888652536603
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,64,128,0.002749333365095986
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,64,64,0.0025680000997251934
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,64,128,0.009827555881606208
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,64,32,0.0026133333643277488
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,64,64,0.009591999981138442
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,64,32,0.010126222338941362
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,32,65536,0.0075635554061995606
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,32,16384,0.006051555689838197
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,32,65536,0.04078933265474107
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,32,12288,0.006137777947717243
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,32,16384,0.01923022170861562
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,32,10240,0.006320888797442119
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,32,12288,0.01739555597305298
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,32,10240,0.016501333978441026
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,32,8192,0.006113777971929974
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,32,7168,0.006070222291681502
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,32,8192,0.015568888849682279
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,32,6144,0.005979555762476391
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,32,7168,0.015011555618709989
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,32,5120,0.0063786668082078295
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,32,6144,0.01463822192615933
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,32,5120,0.013966222604115805
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,32,4096,0.005906666732496685
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,32,3584,0.0063662222690052455
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,32,4096,0.01367644468943278
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,32,3072,0.00620266670982043
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,32,3584,0.01351199961370892
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,32,2560,0.005601777798599667
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,32,3072,0.01293777757220798
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,32,2560,0.012989333106411828
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,32,2048,0.005025777965784073
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,32,2048,0.012353777885437012
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,32,1536,0.004404444247484207
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,32,768,0.012110222544935016
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,32,1024,0.0037600000699361167
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,32,1536,0.012436444560686747
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,32,1024,0.011591999895042844
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,32,768,0.0034613334056403902
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,32,512,0.003140444349911478
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,32,256,0.002798222212327851
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,32,512,0.011955555942323474
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,32,128,0.0026657777941889236
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,32,256,0.01034400032626258
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,32,64,0.002530666689078013
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,32,128,0.010196444061067369
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,float16,1,32,32,0.002558222247494592
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,32,64,0.010015111002657149
VLLM,0.12.0,NVIDIA H200,gemm,vllm_default,fp8,1,32,32,0.010065777434243096
