framework,version,device,op_name,kernel_source,bmm_dtype,num_tokens,num_heads,latency
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,1,128,0.006876800209283829
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,1,128,0.010937599837779999
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,1,64,0.00857279971241951
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,1,32,0.004579199850559235
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,1,32,0.008771199733018875
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,1,16,0.004473600164055824
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,1,16,0.00856959968805313
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,1,8,0.0045056000351905824
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,1,8,0.008534400165081025
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,1,4,0.004569600149989128
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,1,4,0.008508799970149994
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,1,2,0.004524800181388855
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,1,2,0.008560000360012055
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,1,64,0.00639680027961731
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,1,1,0.004579199850559235
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,2,128,0.006735999882221222
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,2,64,0.00453759990632534
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,2,8,0.004464000090956688
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,2,16,0.004579199850559235
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,2,4,0.004464000090956688
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,2,128,0.012755200266838074
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,2,2,0.004636799916625023
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,2,2,0.008515200018882752
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,2,32,0.006403200328350067
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,2,1,0.004403200000524521
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,4,128,0.006707199662923813
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,4,128,0.012697599828243256
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,4,64,0.00456320010125637
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,4,64,0.008479999750852585
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,4,32,0.004623999819159508
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,4,32,0.008736000210046769
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,4,16,0.004527999833226204
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,4,16,0.00880960002541542
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,4,8,0.0043935999274253845
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,4,8,0.00862400010228157
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,4,4,0.004623999819159508
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,4,4,0.008566399663686752
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,4,2,0.004633599892258644
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,4,2,0.009055999666452407
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,4,1,0.004550400003790855
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,4,1,0.00857279971241951
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,8,128,0.006684800237417221
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,8,128,0.012905600666999816
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,8,64,0.0046431999653577805
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,8,64,0.008726400136947633
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,8,32,0.00451200008392334
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,8,32,0.008534400165081025
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,8,16,0.004371200129389763
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,8,16,0.008556800335645676
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,8,8,0.004518400132656098
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,8,8,0.00873280018568039
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,8,4,0.004543999955058098
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,8,4,0.008684799820184708
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,8,2,0.004495999962091446
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,8,2,0.008780799806118011
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,8,1,0.00451200008392334
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,8,1,0.007913599908351897
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,16,128,0.007302399724721909
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,2,64,0.008611200004816055
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,2,32,0.008668799698352814
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,1,1,0.007420799881219864
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,16,32,0.006172800064086914
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,16,128,0.01472959965467453
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,16,16,0.004611200094223023
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,16,64,0.008739200234413148
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,16,16,0.008537600189447403
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,16,8,0.008566399663686752
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,16,4,0.004278400167822838
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,16,32,0.009087999910116195
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,16,4,0.008739200234413148
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,16,2,0.006073600053787232
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,16,2,0.008921600133180618
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,16,1,0.00597120001912117
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,16,1,0.008819200098514557
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,32,128,0.008582399785518646
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,32,128,0.016748799383640288
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,32,64,0.006032000109553337
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,32,64,0.010655999928712846
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,32,32,0.006012799963355064
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,32,32,0.010592000186443329
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,2,16,0.008537600189447403
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,2,8,0.006742399930953979
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,32,16,0.006518399715423584
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,32,16,0.008963199704885483
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,16,64,0.004560000076889992
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,32,4,0.008591999858617782
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,32,8,0.00453759990632534
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,2,4,0.008591999858617782
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,32,8,0.008793599903583527
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,32,4,0.00488319993019104
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,32,2,0.004527999833226204
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,32,2,0.008505599945783615
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,32,1,0.004211200028657913
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,32,1,0.00867839977145195
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,48,128,0.008687999844551087
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,48,128,0.021052800118923187
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,48,64,0.006128000095486641
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,48,16,0.010620799660682679
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,16,8,0.004464000090956688
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,48,64,0.012617599964141846
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,48,4,0.004483199864625931
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,48,32,0.004419200122356415
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,48,32,0.010655999928712846
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,48,2,0.004467200115323066
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,48,16,0.004540799930691719
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,48,8,0.0044895999133586885
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,48,8,0.008710400015115739
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,48,4,0.008819200098514557
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,2,1,0.008627200126647949
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,48,2,0.008563199639320373
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,48,1,0.004479999840259552
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,64,32,0.004579199850559235
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,48,1,0.008908800035715102
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,64,128,0.006691200286149978
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,64,128,0.023004800081253052
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,64,8,0.004483199864625931
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,64,64,0.006457599997520447
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,64,64,0.013052800297737121
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,64,16,0.0046847999095916745
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,64,2,0.00451200008392334
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,64,32,0.011763200163841248
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,64,16,0.010515200346708298
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,64,4,0.004879999905824661
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,64,8,0.009548799693584442
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,64,4,0.008585599809885025
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,64,2,0.008739200234413148
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,64,1,0.004553600028157234
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,64,1,0.008595199882984161
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,80,128,0.008508799970149994
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,80,128,0.02507840096950531
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,80,64,0.0064191997051239015
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,80,64,0.016678400337696075
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,80,32,0.004662400111556053
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,80,32,0.012624000012874604
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,80,16,0.004479999840259552
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,80,16,0.01053439974784851
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,80,8,0.00459199994802475
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,80,8,0.008729600161314011
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,80,4,0.004508800059556961
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,80,4,0.008511999994516373
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,80,2,0.0044351998716592785
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,80,2,0.009071999788284301
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,80,1,0.00445760004222393
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,80,1,0.008588799834251403
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,96,128,0.008508799970149994
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,96,128,0.027113598585128785
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,96,64,0.006297600269317627
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,96,64,0.016758400201797485
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,96,32,0.006537599861621857
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,96,32,0.012636800110340119
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,96,16,0.004419200122356415
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,96,16,0.010815999656915664
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,96,1,0.004476799815893173
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,96,8,0.004639999940991402
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,96,8,0.008899199962615966
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,96,4,0.004569600149989128
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,96,4,0.009424000233411788
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,96,2,0.004598399996757508
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,96,2,0.008656000345945358
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,96,1,0.008764799684286118
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,128,16,0.004454400017857551
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,128,128,0.008489599823951722
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,128,128,0.033081600069999696
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,128,64,0.006703999638557434
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,128,64,0.018972800672054292
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,128,32,0.006351999938488007
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,128,32,0.014707200229167938
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,128,16,0.010992000252008438
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,128,1,0.00461760014295578
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,128,8,0.004515200108289719
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,128,8,0.00875839963555336
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,128,4,0.004467200115323066
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,128,4,0.008720000088214875
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,128,2,0.004499199986457825
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,128,2,0.008691199868917466
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,128,1,0.008803199976682663
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,160,128,0.010163199901580811
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,160,128,0.039241600036621097
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,160,8,0.004700800031423568
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,160,64,0.006566400080919266
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,160,4,0.0045311998575925825
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,160,64,0.022921599447727203
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,160,32,0.006137600168585777
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,160,32,0.015884800255298613
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,160,16,0.004639999940991402
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,160,16,0.010841599851846694
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,160,8,0.010524799674749374
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,160,4,0.01064639985561371
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,160,2,0.004566400125622749
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,160,2,0.008556800335645676
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,160,1,0.005699200183153152
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,160,1,0.008723200112581254
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,192,128,0.010543999820947647
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,192,128,0.041631999611854556
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,192,64,0.006761600077152252
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,192,64,0.024931199848651886
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,192,32,0.006438399851322174
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,192,32,0.016752000153064727
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,192,16,0.004912000149488449
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,192,16,0.01265919953584671
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,192,8,0.004515200108289719
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,192,8,0.010623999685049058
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,192,4,0.00445760004222393
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,192,4,0.008806400001049042
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,192,2,0.0045471999794244765
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,192,2,0.009587199985980987
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,192,1,0.0050432000309228895
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,192,1,0.008656000345945358
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,256,128,0.012246400117874146
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,256,128,0.053615999221801755
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,256,64,0.008528000116348267
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,256,64,0.031036800146102904
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,256,32,0.006700800359249115
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,256,32,0.021081599593162536
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,256,16,0.0045311998575925825
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,256,16,0.014803199470043183
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,256,8,0.004521600157022476
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,256,8,0.0107744000852108
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,256,1,0.008499199897050858
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,256,4,0.004502400010824204
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,256,4,0.008716800063848496
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,256,2,0.004873599857091904
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,256,2,0.008694399893283845
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,256,1,0.004582399874925614
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,320,128,0.01470080018043518
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,320,128,0.06498879790306092
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,320,64,0.008604799956083297
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,320,64,0.037411201000213626
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,320,32,0.006646399945020675
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,320,32,0.022841599583625794
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,320,16,0.0064800001680850984
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,320,16,0.014752000570297241
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,320,8,0.004630399867892265
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,320,8,0.012665599584579468
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,320,4,0.004396799951791763
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,320,4,0.010806400328874588
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,320,2,0.004630399867892265
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,320,2,0.01064319983124733
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,320,1,0.004447999969124794
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,384,64,0.04340479969978332
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,320,1,0.009536000341176987
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,384,16,0.006470400094985962
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,384,128,0.014886400103569031
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,384,128,0.07614719867706299
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,384,64,0.010751999914646149
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,384,32,0.008355200290679932
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,384,32,0.025033599138259886
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,384,8,0.006486400216817856
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,384,16,0.017187200486660004
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,384,8,0.012700800597667695
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,384,4,0.0046431999653577805
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,384,4,0.010470400005578995
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,384,2,0.005142400041222572
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,384,2,0.008723200112581254
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,384,1,0.004454400017857551
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,384,1,0.008761599659919739
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,512,128,0.01979839950799942
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,512,128,0.10031039714813232
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,512,64,0.01252799928188324
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,512,64,0.05377920269966126
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,512,32,0.008691199868917466
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,512,32,0.03105599880218506
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,512,16,0.006524799764156342
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,512,16,0.019619199633598327
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,512,1,0.004425600171089172
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,512,8,0.0064351998269557955
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,512,8,0.012831999361515046
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,512,4,0.004473600164055824
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,768,64,0.014735999703407287
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,512,4,0.010745599865913391
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,512,2,0.004879999905824661
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,512,2,0.010633599758148194
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,512,1,0.008620800077915191
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,768,128,0.02879360020160675
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,768,128,0.14571199417114258
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,768,64,0.07654719948768615
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,768,32,0.010787200182676315
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,768,32,0.043507200479507444
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,768,16,0.008739200234413148
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,768,16,0.026976001262664796
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,768,8,0.006460800021886826
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,768,1,0.010630399733781815
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,768,8,0.017286400496959686
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,768,4,0.006496000289916992
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,768,4,0.012787200510501862
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,768,2,0.004582399874925614
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,768,2,0.010758399963378906
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,768,1,0.005683200061321258
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,1024,128,0.035094401240348815
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,1024,128,0.19084800481796266
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,1024,64,0.01912959963083267
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,1024,64,0.09898239970207215
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,1024,32,0.012591999769210816
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,1024,32,0.05422080159187317
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,1024,16,0.008483199775218964
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,1024,16,0.033228799700737
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,1024,8,0.006518399715423584
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,1024,8,0.018848000466823576
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,1536,128,0.04785279929637909
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,1024,4,0.006492800265550614
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,1024,4,0.012761600315570831
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,1024,2,0.004656000062823296
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,1024,2,0.011372800171375274
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,1536,32,0.014688000082969666
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,1024,1,0.004447999969124794
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,1024,1,0.00870719999074936
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,1536,64,0.027296000719070436
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,1536,128,0.2806112051010132
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,1536,64,0.1450592041015625
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,1536,16,0.010598400235176086
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,1536,32,0.07611520290374756
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,1536,16,0.04353919923305512
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,1536,8,0.00868159979581833
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,1536,8,0.02507199943065643
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,1536,4,0.006547199934720993
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,1536,4,0.01671680063009262
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,1536,2,0.006486400216817856
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,1536,2,0.012678399682044983
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,1536,1,0.004735999926924706
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,1536,1,0.01067200005054474
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,2048,128,0.06210240125656128
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,2048,128,0.3726752042770386
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,2048,64,0.03504000008106232
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,2048,32,0.01875839978456497
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,2048,8,0.008739200234413148
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,2048,64,0.19079359769821166
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,2048,32,0.10008319616317748
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,2048,16,0.012691199779510498
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,2048,16,0.05413439869880676
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,2048,8,0.03332160115242004
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,2048,4,0.006505600363016129
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,2048,4,0.019446399807929993
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,2048,2,0.004745600000023842
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,2048,2,0.013609600067138673
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,2048,1,0.005628800019621849
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,2048,1,0.011884800344705581
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,3072,128,0.08799359798431397
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,3072,64,0.04801599979400635
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,3072,128,0.5523104190826416
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,3072,64,0.28027520179748533
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,3072,32,0.02712000012397766
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,3072,16,0.015024000406265258
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,3072,2,0.006447999924421311
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,3072,32,0.14572160243988036
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,3072,16,0.07618880271911621
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,3072,8,0.010793600231409073
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,3072,8,0.043356800079345705
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,3072,4,0.008339200168848038
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,3072,4,0.025158399343490602
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,3072,2,0.017552000284194947
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,3072,1,0.005868799984455109
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,3072,1,0.012547199428081513
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,4096,128,0.11343040466308593
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,4096,64,0.06150720119476318
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,4096,64,0.3702559947967529
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,4096,32,0.0344895988702774
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,4096,128,0.7344704151153565
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,4096,16,0.018774400651454925
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,4096,32,0.19062399864196777
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,4096,16,0.10051840543746948
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,4096,8,0.012918399274349212
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,4096,8,0.05382080078125
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,4096,4,0.008528000116348267
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,4096,4,0.033107200264930726
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,4096,2,0.006483200192451477
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,4096,2,0.01895039975643158
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,4096,1,0.005612799897789955
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,4096,1,0.014844800531864166
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,6144,128,0.1648416042327881
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,6144,64,0.08785600066184998
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,6144,32,0.047712001204490664
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,6144,64,0.5517375946044922
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,6144,8,0.07627519965171814
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,6144,128,1.106208038330078
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,6144,32,0.2803231954574585
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,6144,16,0.02744320034980774
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,6144,16,0.14569920301437378
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,6144,8,0.014892800152301789
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,6144,4,0.0106175996363163
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,6144,4,0.0435808002948761
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,6144,2,0.008479999750852585
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,6144,2,0.02616640031337738
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,6144,1,0.006470400094985962
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,6144,1,0.016838400065898894
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,8192,128,0.21754560470581055
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,8192,64,0.11349760293960572
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,8192,32,0.06190720200538635
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,8192,32,0.37114241123199465
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,8192,64,0.7314335823059082
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,8192,16,0.035011199116706845
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,8192,2,0.0084927998483181
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,8192,16,0.1900607943534851
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,8192,128,1.4915455818176269
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,8192,8,0.018931199610233308
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,8192,8,0.09904000163078308
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,8192,4,0.012479999661445617
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,8192,4,0.05384640097618103
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,8192,2,0.03316160142421722
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,float16,8192,1,0.006505600363016129
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_pre,default,fp8,8192,1,0.020838400721549986
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,1,128,0.004652800038456917
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,1,128,0.008729600161314011
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,1,64,0.004636799916625023
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,1,64,0.00891520008444786
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,1,32,0.006361600011587143
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,1,32,0.00878399983048439
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,1,16,0.006547199934720993
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,1,8,0.006224000081419945
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,1,8,0.008771199733018875
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,1,4,0.005100800096988678
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,1,4,0.008668799698352814
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,1,2,0.004662400111556053
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,1,2,0.006876800209283829
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,1,1,0.004556800052523613
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,1,1,0.00868159979581833
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,2,128,0.006159999966621399
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,2,128,0.008931200206279754
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,2,64,0.004595199972391129
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,2,64,0.008515200018882752
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,2,32,0.005427199974656105
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,2,32,0.008627200126647949
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,2,16,0.005203200131654739
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,2,16,0.0086496002972126
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,2,8,0.004623999819159508
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,1,16,0.008591999858617782
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,2,8,0.009059199690818786
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,2,4,0.008563199639320373
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,2,2,0.0047680001705884935
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,2,2,0.008595199882984161
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,2,1,0.004633599892258644
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,2,1,0.008640000224113464
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,4,128,0.007132799923419952
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,4,128,0.010623999685049058
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,4,64,0.00480320006608963
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,4,32,0.006191999837756157
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,4,32,0.012771199643611907
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,4,64,0.008710400015115739
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,4,16,0.00618240013718605
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,4,8,0.00623680017888546
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,2,4,0.004582399874925614
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,4,16,0.00870719999074936
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,4,8,0.00870719999074936
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,4,4,0.00453759990632534
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,4,4,0.008531200140714646
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,4,2,0.004543999955058098
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,4,1,0.008630400151014328
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,8,128,0.004742399975657463
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,8,64,0.004636799916625023
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,8,64,0.008691199868917466
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,8,16,0.0047775998711586
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,8,8,0.004742399975657463
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,8,4,0.004720000177621841
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,8,4,0.00862400010228157
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,8,2,0.0049472000449895855
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,8,2,0.008755200356245042
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,8,1,0.004428799822926521
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,8,1,0.008761599659919739
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,16,128,0.006473600119352341
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,4,1,0.004550400003790855
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,16,64,0.0046431999653577805
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,16,128,0.012870399653911591
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,16,32,0.0046271998435258865
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,16,32,0.009779199957847595
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,16,16,0.004927999898791313
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,16,16,0.009657599776983262
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,16,64,0.010579200088977813
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,16,8,0.004771199822425842
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,16,8,0.008873599767684936
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,16,4,0.0045471999794244765
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,16,4,0.008575999736785888
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,16,2,0.004675199836492538
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,8,32,0.00459199994802475
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,16,2,0.008617600053548813
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,16,1,0.0047391999512910845
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,32,128,0.006508799642324448
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,32,64,0.0044319998472929
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,32,128,0.014803199470043183
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,32,64,0.010844799876213073
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,32,32,0.004652800038456917
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,32,32,0.010527999699115753
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,32,16,0.004646399989724159
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,32,16,0.009945599734783173
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,32,8,0.0049472000449895855
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,32,8,0.010556799918413162
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,32,4,0.004454400017857551
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,32,4,0.008867199718952178
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,32,2,0.0045471999794244765
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,32,2,0.00851840004324913
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,32,1,0.004543999955058098
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,32,1,0.008579199761152267
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,48,128,0.006547199934720993
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,48,128,0.016832000017166136
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,48,64,0.004806400090456009
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,48,64,0.012675200402736665
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,48,32,0.005619199946522713
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,48,32,0.010995200276374817
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,48,16,0.004579199850559235
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,48,16,0.01055999994277954
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,48,8,0.004774399846792221
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,48,8,0.00867839977145195
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,48,4,0.004691199958324432
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,48,4,0.008803199976682663
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,48,2,0.004575999826192856
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,48,2,0.008630400151014328
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,48,1,0.004867200180888176
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,48,1,0.008703999966382981
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,64,128,0.006547199934720993
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,64,128,0.01887039989233017
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,4,2,0.00671359971165657
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,64,64,0.0047520000487566
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,64,64,0.014758400619029999
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,64,32,0.005382400006055832
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,64,32,0.010982400178909302
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,64,16,0.004687999933958053
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,64,8,0.004953600093722343
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,64,16,0.010576000064611435
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,64,8,0.009071999788284301
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,64,4,0.004540799930691719
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,64,4,0.008799999952316284
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,64,2,0.00448639988899231
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,64,2,0.008675199747085572
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,64,1,0.004419200122356415
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,64,1,0.008595199882984161
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,80,128,0.006620799750089645
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,80,128,0.02096319943666458
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,80,64,0.006249599903821945
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,80,64,0.014764800667762756
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,80,32,0.004639999940991402
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,80,32,0.012831999361515046
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,80,16,0.004668800160288811
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,80,16,0.01064319983124733
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,80,8,0.004534399881958961
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,8,8,0.008579199761152267
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,8,32,0.00859839990735054
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,8,16,0.008585599809885025
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,80,8,0.010761599987745285
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,80,4,0.004569600149989128
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,8,128,0.008876799792051315
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,80,4,0.008563199639320373
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,80,2,0.008582399785518646
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,80,1,0.005071999877691269
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,80,1,0.008524800091981888
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,96,128,0.024876800179481507
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,96,64,0.017132799327373504
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,96,32,0.004662400111556053
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,96,32,0.012854400277137756
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,96,16,0.01085439994931221
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,96,16,0.004534399881958961
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,96,8,0.004687999933958053
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,96,8,0.010527999699115753
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,96,4,0.004822399839758873
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,96,4,0.008684799820184708
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,96,2,0.00448639988899231
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,96,2,0.008921600133180618
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,96,1,0.004844800010323525
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,96,1,0.00856959968805313
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,128,128,0.006908799707889557
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,16,1,0.008563199639320373
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,128,128,0.029145601391792297
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,128,64,0.006400000303983688
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,128,64,0.019046400487422944
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,128,32,0.004851200059056282
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,80,2,0.00453759990632534
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,128,32,0.012988799810409546
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,128,16,0.010761599987745285
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,128,8,0.0045056000351905824
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,96,128,0.006598400324583054
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,128,8,0.01056319996714592
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,128,4,0.004403200000524521
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,96,64,0.006595200300216675
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,128,4,0.00947519987821579
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,128,2,0.004476799815893173
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,128,2,0.009161599725484849
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,128,1,0.004662400111556053
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,128,1,0.008511999994516373
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,160,128,0.00671359971165657
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,160,16,0.0046720001846551895
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,160,128,0.03528000116348266
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,160,64,0.006518399715423584
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,160,64,0.022486400604248048
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,160,32,0.006508799642324448
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,160,32,0.014912000298500061
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,160,8,0.004556800052523613
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,160,16,0.012700800597667695
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,160,8,0.010915199667215348
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,160,4,0.004604800045490265
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,160,4,0.010662399977445603
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,160,2,0.004396799951791763
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,160,2,0.008640000224113464
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,128,16,0.004534399881958961
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,160,1,0.004447999969124794
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,160,1,0.008867199718952178
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,192,128,0.008524800091981888
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,192,128,0.04127680063247681
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,192,64,0.006447999924421311
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,192,64,0.023081600666046143
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,192,32,0.00650240033864975
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,192,32,0.016659200191497803
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,192,16,0.004467200115323066
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,192,16,0.012716799974441528
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,192,8,0.004611200094223023
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,192,8,0.010608000308275222
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,192,4,0.0047136001288890835
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,192,4,0.008591999858617782
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,192,2,0.004732799902558327
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,256,128,0.050704002380371094
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,192,2,0.008687999844551087
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,192,1,0.004428799822926521
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,192,1,0.008723200112581254
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,256,128,0.008595199882984161
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,256,64,0.006473600119352341
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,256,64,0.029264000058174134
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,256,32,0.006294400244951248
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,256,32,0.018764799833297728
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,256,16,0.004614400118589402
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,256,16,0.014659200608730317
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,256,8,0.0045471999794244765
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,256,8,0.010604800283908844
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,256,4,0.004668800160288811
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,256,4,0.010566399991512298
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,256,2,0.004527999833226204
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,256,2,0.00851840004324913
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,256,1,0.004540799930691719
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,256,1,0.008579199761152267
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,320,128,0.009971199929714203
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,320,128,0.06239680051803589
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,320,64,0.006662400066852569
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,320,64,0.03516480028629303
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,320,32,0.006524799764156342
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,320,32,0.020934399962425233
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,320,16,0.004675199836492538
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,320,16,0.014934399724006652
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,320,8,0.004870399832725525
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,320,8,0.012681600451469422
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,320,4,0.004623999819159508
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,320,4,0.010604800283908844
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,320,2,0.0045855998992919925
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,320,2,0.009375999867916106
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,320,1,0.004550400003790855
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,320,1,0.008710400015115739
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,384,16,0.006521599739789963
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,384,128,0.010819199681282043
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,384,128,0.07555519938468933
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,384,64,0.006684800237417221
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,384,64,0.04137920141220093
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,384,32,0.00655680000782013
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,384,32,0.025113600492477416
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,384,16,0.017155200242996216
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,384,8,0.005795200169086456
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,384,8,0.013030399382114411
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,384,4,0.004806400090456009
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,384,4,0.01056319996714592
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,384,2,0.0046016000211238865
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,384,2,0.010623999685049058
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,384,1,0.00445760004222393
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,384,1,0.009308800101280212
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,512,128,0.017523199319839478
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,512,128,0.09822720289230347
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,512,64,0.00865280032157898
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,512,64,0.05090240240097046
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,512,32,0.006598400324583054
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,512,32,0.029100799560546876
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,512,2,0.004550400003790855
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,512,16,0.006524799764156342
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,512,16,0.018915200233459474
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,512,8,0.005209600180387497
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,512,8,0.012892800569534301
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,512,4,0.004729599878191948
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,512,4,0.010976000130176545
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,512,2,0.008684799820184708
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,768,64,0.07434239983558655
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,512,1,0.005107200145721436
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,512,1,0.008604799956083297
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,768,16,0.02496960014104843
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,768,128,0.029071998596191407
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,768,8,0.004665600135922432
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,768,128,0.14594240188598634
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,768,64,0.010623999685049058
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,768,32,0.008448000252246856
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,768,16,0.006431999802589417
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,768,32,0.03952319920063019
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,768,8,0.016812799870967864
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,768,4,0.005020799860358238
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,768,4,0.012627199292182922
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,768,2,0.004678399860858917
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,768,2,0.010806400328874588
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,768,1,0.004611200094223023
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,768,1,0.008710400015115739
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,1024,128,0.03715519905090332
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,1024,64,0.012723200023174286
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,1024,128,0.19081599712371827
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,1024,64,0.09864000082015992
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,1024,32,0.008579199761152267
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,1024,32,0.051523202657699586
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,1024,16,0.00647680014371872
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,1024,16,0.029190400242805482
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,1024,8,0.006604799628257751
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,1024,8,0.01905599981546402
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,1024,4,0.004620800167322159
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,1024,4,0.014665600657463074
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,1024,2,0.004569600149989128
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,1024,2,0.010566399991512298
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,1024,1,0.004508800059556961
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,1024,1,0.010716799646615982
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,1536,128,0.05283520221710205
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,1536,128,0.2809920072555542
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,1536,64,0.02698560059070587
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,1536,64,0.14534399509429932
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,1536,32,0.0106175996363163
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,1536,32,0.07428479790687562
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,1536,16,0.008566399663686752
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,1536,16,0.03958399891853333
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,1536,2,0.013414399325847625
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,1536,8,0.006457599997520447
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,1536,8,0.022924800217151643
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,1536,4,0.0046847999095916745
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,1536,4,0.016764800250530242
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,1536,2,0.006438399851322174
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,1536,1,0.00456320010125637
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,1536,1,0.01067200005054474
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,2048,128,0.06443520188331604
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,2048,64,0.03923200070858002
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,2048,128,0.3752608060836792
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,2048,64,0.1887488007545471
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,2048,32,0.01268479973077774
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,2048,32,0.09860159754753113
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,2048,16,0.008710400015115739
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,2048,16,0.05175679922103882
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,2048,1,0.004579199850559235
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,2048,8,0.00647680014371872
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,2048,8,0.029203200340270997
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,2048,4,0.006281600147485733
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,2048,4,0.018931199610233308
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,2048,2,0.006563200056552887
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,2048,2,0.012777599692344665
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,2048,1,0.010902400314807891
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,3072,16,0.010694400221109391
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,3072,128,0.09319040179252625
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,3072,8,0.008646400272846222
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,3072,64,0.0516319990158081
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,3072,4,0.006684800237417221
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,3072,128,0.5639455795288086
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,3072,4,0.02487040013074875
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,3072,64,0.2806560039520264
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,3072,32,0.026873600482940675
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,3072,1,0.012796799838542938
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,3072,32,0.14592000246047973
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,3072,16,0.07426880002021789
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,3072,8,0.041280001401901245
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,3072,2,0.006304000318050384
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,3072,2,0.01695999950170517
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,3072,1,0.004633599892258644
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,4096,128,0.11631360054016113
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,4096,64,0.06976959705352784
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,4096,8,0.0086496002972126
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,4096,32,0.03135679960250855
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,4096,128,0.7808095932006835
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,4096,64,0.3730783939361572
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,4096,32,0.19089280366897582
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,4096,16,0.012780800461769104
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,4096,16,0.09747200012207032
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,4096,8,0.051481598615646364
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,4096,4,0.006822399795055389
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,4096,4,0.03088639974594116
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,4096,2,0.006524799764156342
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,6144,64,0.08848320245742798
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,4096,2,0.018969599902629853
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,4096,1,0.00594559982419014
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,4096,1,0.014678399264812469
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,6144,128,0.1716096043586731
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,6144,16,0.1441759943962097
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,6144,32,0.0466623991727829
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,6144,128,1.2369983673095704
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,6144,32,0.28037118911743164
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,6144,64,0.5622911930084229
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,6144,16,0.027033600211143493
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,6144,8,0.010531199723482132
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,6144,1,0.016790400445461272
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,6144,4,0.00859839990735054
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,6144,8,0.07215359807014465
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,6144,4,0.04052160084247589
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,6144,2,0.006425599753856659
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,6144,2,0.024275200068950654
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,6144,1,0.004483199864625931
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,8192,128,0.2205984115600586
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,8192,64,0.11196160316467285
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,8192,32,0.0611840009689331
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,8192,64,0.7713247776031494
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,8192,32,0.37541120052337645
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,8192,16,0.03273600041866302
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,8192,2,0.006486400216817856
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,8192,128,1.6721887588500977
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,8192,16,0.18935359716415406
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,8192,8,0.01268479973077774
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,8192,8,0.09778879880905152
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,8192,4,0.008646400272846222
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,8192,4,0.05158079862594604
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,8192,2,0.030399999022483824
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,8192,1,0.006438399851322174
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,8192,1,0.01883520036935806
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,12288,128,0.32991681098937986
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,12288,64,0.16351360082626343
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,12288,32,0.0850655972957611
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,12288,16,0.04539200067520142
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,12288,8,0.02690880000591278
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,12288,64,1.206223964691162
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,12288,32,0.5597983837127686
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,12288,128,2.59366397857666
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,12288,16,0.2797919988632202
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,12288,8,0.138646399974823
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,12288,4,0.010585600137710571
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,12288,4,0.07425919771194459
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,12288,2,0.008595199882984161
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,12288,2,0.03952000141143799
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,12288,1,0.006387200206518173
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,12288,1,0.024726399779319765
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,16384,128,0.4362527847290039
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,16384,16,0.061680001020431516
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,16384,64,0.21368319988250734
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,16384,32,0.11196479797363282
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,16384,8,0.03115839958190918
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,16384,32,0.7813856124877929
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,16384,16,0.37218239307403567
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,16384,8,0.1908768057823181
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,16384,64,1.6635776519775392
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,16384,2,0.04967359900474548
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,16384,4,0.012681600451469422
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,16384,4,0.09862080216407776
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,16384,2,0.008556800335645676
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,16384,128,3.483235168457031
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,16384,1,0.006646399945020675
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,16384,1,0.02903040051460266
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,20480,32,0.13738880157470704
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,20480,128,0.5400415897369385
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,20480,64,0.2659807920455933
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,20480,16,0.07247040271759034
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,20480,16,0.4677248001098633
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,20480,8,0.04922240078449249
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,20480,32,1.000704002380371
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,20480,4,0.01892800033092499
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,20480,8,0.2342144012451172
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,20480,64,2.140118408203125
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,20480,4,0.12175359725952148
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,20480,2,0.008579199761152267
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,20480,2,0.0619488000869751
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,float16,20480,1,0.006707199662923813
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,20480,1,0.03540799915790558
SGLang,0.5.6.post2,NVIDIA B200,mla_gen_post,default,fp8,20480,128,4.358537673950195
