framework,version,device,op_name,kernel_source,gemm_dtype,m,n,k,latency
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,32,2560,0.006696888970004187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,7168,5120,0.03838933176464505
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,8192,2048,0.00850044439236323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,2560,512,0.005671111245950063
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,10240,3072,0.015270221564504834
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,768,10240,0.014527110589875115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,128,128,0.002124444478087955
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,7168,6144,0.01294222225745519
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1536,512,0.005943999936183293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,16384,10240,0.049379554059770375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2560,256,0.002854222224818336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,32,512,0.005607111172543631
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,65536,3072,0.245506657494439
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,5120,2048,0.006991110742092133
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,3072,768,0.0058124446206622655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,7168,3072,0.010856889188289642
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,5120,1024,0.004926222066084544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,256,3584,0.007735999921957652
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,7168,2048,0.04125866625044081
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,3072,1536,0.006353777729802662
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,7168,2560,0.01074488874938753
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1024,8192,0.008896888958083259
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,4096,2048,0.007288888924651676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,2048,10240,0.03238933285077413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,768,6144,0.022783999641736347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,512,3584,0.005551110953092575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3072,2560,0.007304000357786815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,12288,16384,0.05838311380810208
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,3072,5120,0.017608889275126986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,3072,256,0.005997333261701796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,256,1024,0.012411555482281579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,5120,65536,4.00768068101671
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3072,5120,0.011845333708657159
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,7168,16384,0.3633306556277805
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,128,7168,0.004884444591071871
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,4096,3584,0.010420444111029306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,12288,10240,0.043507556120554604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1024,768,0.005587555468082428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,12288,10240,0.0609440008799235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,2560,2560,0.014554666148291694
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,32,128,0.006311111152172089
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,3584,128,0.01440355512830946
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2048,16384,0.014877332581414117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,65536,8192,0.7068853378295898
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,16384,6144,0.07654311259587605
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,4096,2048,0.008226667013433244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1024,3072,0.007840889195601145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1024,64,0.008277333445019191
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,7168,1024,0.009914666414260864
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,512,1024,0.005904888941182031
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,2560,512,0.005567111074924469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,64,12288,0.013387555877367655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,6144,256,0.0052764444715446895
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,3072,10240,0.016866667403115165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,768,128,0.0021644443687465456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,8192,1024,0.04319733381271362
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,1536,1024,0.013788444300492605
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,10240,512,0.019113777412308585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,64,64,0.005555555638339784
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,6144,2048,0.006641777853171031
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2560,4096,0.008250666989220513
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,5120,2048,0.09392178058624268
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,7168,16384,0.02522666586769952
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,1536,7168,0.009713778065310584
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,5120,256,0.00563822231358952
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,10240,4096,0.018608889646000333
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1024,4096,0.00833155545923445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,768,768,0.003145777723855443
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,128,2048,0.009293333523803288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3584,10240,0.017204445269372728
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,2560,12288,0.019976889093716938
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,12288,2560,0.03226666649182638
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,16384,1536,0.030266665750079687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,4096,128,0.005939555664857228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,10240,64,0.007287110719415877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,8192,3072,0.009889778163697984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,6144,1536,0.006633777585294511
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,16384,3584,0.01587377819750044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,65536,1536,0.16944177945454916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,4096,7168,0.04212266537878248
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,4096,16384,0.024858666790856257
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,7168,2048,0.021552888883484736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,8192,128,0.005580444302823808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,3072,2560,0.019859555694792006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,4096,1536,0.016936888297398884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,5120,512,0.007989333735571967
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,8192,256,0.021772444248199463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,5120,16384,0.02379377848572201
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,4096,512,0.0059262220230367445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3072,256,0.0026133333643277488
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3072,64,0.008035555481910706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,10240,3072,0.012982221941153208
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2560,65536,0.05545599924193489
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2560,256,0.002599999929467837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,2560,1024,0.014895111322402954
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,10240,64,0.009098666409651438
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2560,8192,0.01420177850458357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,4096,65536,0.06831822130415174
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,4096,32,0.02481866710715824
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,32,4096,0.004587555511130227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,128,5120,0.016185777054892648
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,65536,2560,0.031480888525644936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,512,512,0.005582222094138463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,10240,512,0.003690666622585721
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,7168,2560,0.009650666680600908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,10240,3072,0.06614755259619819
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,65536,2560,0.08836533625920613
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,128,5120,0.021114667256673176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,32,1536,0.004497777670621872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,3584,12288,0.031572444571389094
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,128,256,0.0053137776752312975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,10240,128,0.00979466653532452
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,768,3584,0.018426666657129925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,6144,5120,0.05747822258207533
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,512,5120,0.052869333161248096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,16384,2048,0.012250666817029318
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,12288,64,0.0022657778528001574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2560,16384,0.09200800127453274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,5120,3584,0.00795377790927887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,256,5120,0.004910222358173794
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,2560,1024,0.014116444521480136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,1024,256,0.014896000425020853
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,7168,3072,0.026703110999531213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,4096,10240,0.017501332693629794
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,10240,5120,0.020647111866209242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,8192,512,0.008399110701349046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,128,1024,0.0029102222373088202
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,128,2560,0.016456888781653512
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,256,8192,0.07228089041180082
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,5120,1024,0.005698666804366642
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,256,16384,0.015311110350820752
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,12288,1536,0.027242667145199243
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,256,128,0.011418666276666852
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1024,1536,0.015340444114473132
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,768,8192,0.006974221931563483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1536,256,0.002528000002106031
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3072,5120,0.011071999867757162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,5120,3072,0.008377778033415476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,512,65536,0.09244266483518813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2560,65536,0.057592001226213246
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,8192,32,0.0029253332565228143
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,512,7168,0.011772444678677453
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1536,64,0.006606222026877933
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,768,1024,0.03257600135273404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,16384,10240,0.054309331708484225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,8192,128,0.015186667442321777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,6144,256,0.002502222235004107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3072,2048,0.007245333658324347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1536,12288,0.053370667828453906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,5120,3072,0.03258666727277968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,6144,512,0.0028159999185138275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,128,5120,0.015828443898095023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,6144,768,0.014509333504570855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,12288,1024,0.021789333886570398
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3584,1536,0.018787556224399142
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,64,16384,0.01588444411754608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,2048,6144,0.1541528834236993
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,256,2560,0.006959999601046245
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2048,768,0.006010666489601135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,5120,3072,0.056799113750457764
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,10240,8192,0.26317334175109863
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,512,5120,0.016920000314712524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,3072,7168,0.030227555168999568
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3072,32,0.002258666687541538
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,32,16384,0.015264888604482016
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,65536,4096,0.039592888620164655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1024,5120,0.009646221995353699
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,3072,3584,0.009196444518036312
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,2560,1536,0.01835644410716163
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,128,32,0.0022035555707083806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,65536,10240,0.08958577447467381
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,12288,64,0.007672888537247975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,5120,3584,0.08183466725879245
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3584,32,0.014012444350454541
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3072,6144,0.013245333400037555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2048,128,0.00600622221827507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,1536,3072,0.015578667322794596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,5120,2560,0.018627555833922494
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,7168,12288,0.023267555567953322
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,65536,1024,0.015885333220163982
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,7168,2560,0.026358221968015034
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,16384,10240,0.03322488731808133
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2560,8192,0.021896888812383015
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,10240,2048,0.012114666402339935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,12288,65536,0.1539671156141493
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,10240,32,0.006554666492674086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,10240,10240,0.033915556139416166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,1536,5120,0.041336887412601046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2560,16384,0.02492977844344245
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,3072,3584,0.023824888798925612
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,10240,7168,0.02442311081621382
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2048,128,0.005564444594913059
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,128,256,0.011353777514563667
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,16384,2048,0.015823110938072205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,768,2560,0.01516355574131012
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,128,768,0.005681777579916849
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,12288,2560,0.026305778159035578
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,128,128,0.0052897776994440295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,4096,2048,0.007424888511498769
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,256,4096,0.008026666939258575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,256,65536,0.07712444331910875
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,7168,3584,0.009407999614874521
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,16384,12288,0.0645244452688429
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,6144,5120,0.015304888288180033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,32,65536,0.06096000141567654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,12288,2560,0.012468444804350534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,1536,2048,0.024158222807778254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,4096,6144,0.05077155431111654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,6144,4096,0.01221777747074763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,7168,512,0.016917333006858826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,3584,1536,0.00627377794848548
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,256,12288,0.01346666696998808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,4096,8192,0.04391377833154467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1024,32,0.0021991110924217436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3072,512,0.0032293332947625052
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,65536,10240,0.10218133529027303
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2560,7168,0.009517333573765224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,16384,4096,0.019170666734377544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2560,7168,0.016798221402698092
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,5120,128,0.005912000106440649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,5120,512,0.005568000177542369
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3072,65536,0.06377866533067492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,16384,256,0.011847111086050669
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,12288,256,0.009027555584907532
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1024,65536,0.05798044469621447
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,3072,7168,0.009935999910036722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,256,16384,0.006666666517655055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,768,512,0.013208000196350945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,12288,64,0.006986666884687211
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2048,32,0.0021679999513758556
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,8192,1024,0.018242667118708294
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,3584,512,0.005565333283609814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,512,3072,0.014831999937693277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,8192,3072,0.019433778193261888
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,64,10240,0.012141333685980903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,6144,768,0.008078222473462423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,5120,3584,0.025641777449183997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,1024,10240,0.022064000368118286
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1536,7168,0.017627555463049147
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,5120,2048,0.006956444846259222
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,4096,7168,0.014742222097184924
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3584,65536,0.07297866874270968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,768,4096,0.008756444685988957
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,5120,256,0.005325333525737126
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2560,256,0.008573333422342936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,5120,10240,0.01773244473669264
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,512,2048,0.006704888823959563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,3584,768,0.01757600075668759
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,64,5120,0.015862221519152325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,65536,512,0.11742311053805882
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1024,1536,0.004458666675620609
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,512,10240,0.006638222270541721
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,10240,3584,0.016215999921162922
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,256,2048,0.006605333338181178
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,768,2048,0.006440889090299606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,4096,64,0.014779556128713818
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,6144,12288,0.03661866651640998
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,3584,512,0.014937778313954672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,3584,256,0.005974222388532426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,10240,16384,0.07955911424424914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2560,8192,0.013790222505728403
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1024,16384,0.08323999908235338
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,10240,256,0.03559199968973795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,6144,128,0.0025271111064487034
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3072,10240,0.016297777493794758
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,64,16384,0.04592533244027031
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,32,768,0.002838222309947014
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,8192,2048,0.023349333140585158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,8192,64,0.0031146665828095544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,512,768,0.006956444846259222
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,256,6144,0.017229333519935608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,12288,16384,0.07125244537989299
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1536,2560,0.009343999955389235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,128,2560,0.00488711107108328
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,12288,4096,0.01943911115328471
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,256,7168,0.00941955546538035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,6144,6144,0.014175999495718213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,4096,6144,0.026139555705918208
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,5120,2560,0.011183111204041375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,256,8192,0.020277332928445604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,5120,3072,0.021584888299306233
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,5120,2048,0.028274665276209515
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,128,4096,0.008039111064540016
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,4096,16384,0.023028444912698533
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,32,512,0.006682666639486949
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,7168,768,0.00851022203763326
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,512,4096,0.008049777812427944
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,8192,256,0.018616000811258953
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,10240,2048,0.010152888794740042
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,128,1024,0.0032702222880389956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,7168,1536,0.054591112666659884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,10240,10240,0.13934489091237387
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,2048,2560,0.014000889327791003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,7168,3072,0.01716444392999013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,2048,512,0.013184000220563678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3584,2048,0.006255110932721033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3584,2560,0.03989511066012912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,512,512,0.005647111270162795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,512,1024,0.013026666310098437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,6144,8192,0.15803289413452148
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,12288,65536,0.14616444375779894
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,16384,2560,0.013399110900031196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,6144,3584,0.012838222086429596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,768,5120,0.008567111359702216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,3584,2048,0.006656000183688269
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,3584,8192,0.012540444731712341
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,65536,512,0.35459287961324054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,1024,768,0.005547555370463266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,16384,3072,0.1149066686630249
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,65536,3072,0.09187466568417019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,8192,512,0.005703111075692707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,32,3072,0.004666666603750653
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,4096,512,0.018082666728231642
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,3584,6144,0.03725688987308078
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,512,32,0.0069617778062820435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2048,16384,0.01969688965214623
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,128,12288,0.027131555808915034
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3584,4096,0.011114666859308878
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,8192,65536,0.12816177474127874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,8192,512,0.003089777711364958
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,6144,12288,0.028140445550282795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,4096,2560,0.007335111498832703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,65536,6144,0.056714667214287647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,128,6144,0.005315555466545953
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,64,3072,0.004707555390066571
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,7168,16384,0.04858044452137417
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,12288,5120,0.19337777296702066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,10240,1024,0.00700711127784517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,7168,64,0.006623999940024481
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,3584,16384,0.021105777886178758
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,7168,16384,0.04380888740221659
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,7168,3584,0.013225778109497495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,2560,65536,0.12317511770460342
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,5120,1536,0.004589333302444882
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,65536,1024,0.01776444415251414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,4096,1536,0.017290666699409485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,65536,3072,0.08701155583063762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,12288,65536,0.14702399571736655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,1536,1536,0.0063031112982167145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,5120,12288,0.034992890225516424
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,10240,2560,0.018798222144444782
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,6144,1024,0.005970666805903117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,12288,3584,0.014491554763582019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,768,4096,0.007690666450394525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1536,8192,0.01146577795346578
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,128,65536,0.048672000567118325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,512,128,0.015544000599119397
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1024,12288,0.07695733176337348
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,32,32,0.005605333381228977
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,12288,768,0.00564444437623024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,6144,10240,0.016120000018013846
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,768,1536,0.015996444556448195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,7168,8192,0.044544888867272266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,2048,10240,0.12993066840701634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,4096,512,0.018358222312397428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,8192,512,0.017289777596791584
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,64,2048,0.007000000112586551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,512,8192,0.027628444963031348
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,8192,256,0.0070675553547011475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,7168,10240,0.02975733412636651
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,65536,1536,0.6080177625020345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,7168,3584,0.09879289070765178
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,7168,65536,0.7453529040018717
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,65536,2048,0.07585244708591037
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,3072,7168,0.011011555790901184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,7168,512,0.016557332542207506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,16384,512,0.029839999145931665
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,16384,65536,0.18071644836001924
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,6144,6144,0.02048000031047397
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,2048,1024,0.0151591118839052
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,1024,128,0.014500444134076437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1024,1024,0.0036675557494163513
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,12288,2560,0.06653066476186116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,512,2048,0.00628355559375551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,10240,5120,0.01591466698381636
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,64,32,0.005636444522274866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,3072,1536,0.006657777975002925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,128,7168,0.005658666706747479
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,65536,256,0.009557333257463243
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,128,12288,0.007640000018808577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,768,3072,0.007653333246707916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,1536,2560,0.006623999940024481
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,7168,65536,0.07735555701785617
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,65536,3072,0.037263999382654824
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,5120,12288,0.0647688905398051
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3072,3072,0.007654222349325816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,6144,64,0.005936000082227919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,1024,1024,0.013456888496875763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3584,512,0.0046640001237392426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,5120,3072,0.009504888620641496
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,768,3072,0.009379555781682333
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,128,65536,0.2741662131415473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,4096,16384,0.028756442997190688
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,128,10240,0.0369777778784434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,5120,2048,0.021949332621362474
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,128,8192,0.06593066453933716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2560,256,0.005260444349712796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,64,128,0.002159111201763153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,4096,6144,0.029134223858515423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3072,10240,0.014497778481907315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,5120,1024,0.017490666773584154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,6144,128,0.031996445523367986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,256,4096,0.006544888847404056
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,768,3584,0.007152000235186682
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,256,128,0.005564444594913059
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,10240,1024,0.020975111259354487
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,7168,1024,0.017504000001483493
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,10240,256,0.006585777633719974
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,3072,256,0.005616000129116907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,256,256,0.002144888871245914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,10240,10240,0.026143110460705225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,1024,10240,0.03714222378200955
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,10240,2048,0.023969777756267126
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,65536,512,0.06823199987411499
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,4096,3072,0.00740622232357661
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,4096,64,0.0018746666610240936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,10240,5120,0.019900444481107924
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,256,16384,0.017180444465743173
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,4096,8192,0.02587911155488756
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,4096,768,0.00590133335855272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,4096,64,0.0028826666788922418
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,10240,3072,0.012161778079138862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,16384,4096,0.015471110741297403
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3072,8192,0.18171111742655435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3584,5120,0.012456888953844706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,8192,768,0.018009778526094224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,128,10240,0.0125937776433097
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,256,512,0.0025537777692079544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2560,1536,0.02310577697224087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,2560,65536,0.06390488810009427
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,16384,8192,0.05996177593866984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3072,768,0.005967999911970562
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,10240,6144,0.11310222413804795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,2048,65536,0.11051999860339695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,3072,1024,0.005947555518812603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,512,1024,0.008001777860853408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,65536,1024,0.15734488434261745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,768,3072,0.016167999969588388
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,1024,16384,0.09031911028756036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,64,64,0.001863111017478837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,10240,5120,0.05052711235152351
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,5120,3584,0.012846222354306115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,128,10240,0.011751111182901593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,8192,1536,0.010616000327799054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,128,768,0.002809777855873108
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,512,16384,0.03155733479393853
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3072,8192,0.03289511137538486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,6144,2048,0.006631999793979857
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,4096,1024,0.016855110724767048
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,1536,10240,0.038811554511388145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,10240,5120,0.08110044399897258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,1024,65536,0.06357155243555705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,768,2560,0.015246222416559855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,16384,1536,0.00981244444847107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,3584,128,0.005336888962321811
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,4096,7168,0.013541333377361298
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,64,10240,0.012875555290116204
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,32,8192,0.010703999963071613
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,10240,3584,0.016946666770511203
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,5120,512,0.005572444448868434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,8192,512,0.017867555220921833
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,3584,65536,0.1575955549875895
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,4096,128,0.0032026666320032547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,32,16384,0.015812445018026564
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,64,2560,0.00452533322903845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1024,8192,0.008643555144468943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,4096,10240,0.01833866702185737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,1024,768,0.005636444522274866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,5120,256,0.00527022240890397
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,12288,6144,0.021361778179804485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,768,1024,0.005942222144868638
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,6144,5120,0.10816888676749335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,256,2560,0.00868266655339135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,768,65536,0.020587555236286588
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,16384,10240,0.034535109996795654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1024,4096,0.005972444597217772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1024,10240,0.010118222071064843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3072,256,0.00958933350112703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,6144,7168,0.039176887936062284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,1536,65536,0.06485511196984185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,10240,1024,0.006599999964237213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,6144,1536,0.007979555262459649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,6144,8192,0.03025777803526984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,5120,512,0.007982222570313348
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,7168,5120,0.01238222254647149
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,32,768,0.002874666617976295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,256,128,0.0022168887986077205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,8192,256,0.01685599982738495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,4096,256,0.00555288874440723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1536,5120,0.008655110994974772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,8192,12288,0.036751111348470054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3584,65536,0.07193866703245375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,128,1024,0.005614222337802251
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,1536,12288,0.03645777702331543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,512,3072,0.010728889041476779
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,65536,10240,0.1000231107076009
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3584,8192,0.014689778288205465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2048,256,0.005447111195988125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2560,2560,0.009676444033781687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,768,3072,0.010313777459992303
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,8192,64,0.006096000058783426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1536,65536,0.05907199780146281
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,512,65536,0.03151999910672506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,512,1024,0.013071999781661563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,256,4096,0.010355555348926121
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,128,12288,0.005199999858935674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,4096,256,0.004263999975389904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,5120,256,0.005588444570700328
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,16384,3072,0.0346684455871582
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,12288,2048,0.029990222718980577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,4096,3072,0.008349333372380998
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,1024,768,0.005576000031497743
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,3584,65536,0.09033777978685166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,8192,65536,0.1743884483973185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,256,768,0.0028719999310043124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,5120,1536,0.004679999831649992
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,1536,7168,0.033950222863091364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,2560,10240,0.045718222856521606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,4096,3584,0.027098667290475633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1024,10240,0.012443555725945367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,10240,512,0.006842666616042455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,16384,2560,0.034697777695126004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,256,6144,0.09879377815458511
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,6144,7168,0.03626311156484816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,5120,65536,0.07253599829143949
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,1024,2048,0.008035555481910706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,768,128,0.0022986665782001284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,6144,65536,3.7423973083496094
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1024,64,0.002895999906791581
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1024,8192,0.024311110377311707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,1536,5120,0.012271111210187277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,4096,32,0.0032115555885765287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,768,768,0.002895999906791581
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,8192,2048,0.024002666274706524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,128,3584,0.010775999890433418
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1024,65536,0.029511110650168523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,128,3584,0.019714666737450492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,2560,3584,0.02544622288809882
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,512,7168,0.005614222337802251
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1024,65536,0.10040266646279229
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,4096,128,0.006578666468461354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,32,4096,0.008416888614495596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,2048,2048,0.006978666616810693
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,5120,128,0.005682666682534748
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,16384,4096,0.04759555392795139
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,512,6144,0.017872888180944655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,6144,3072,0.009712888962692684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,256,2048,0.012757333616415659
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,5120,4096,0.028446220689349707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,32,32,0.001869333287080129
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2560,12288,0.026624000734753076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3584,32,0.005590222362014983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,65536,7168,0.2803022331661648
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,10240,3072,0.03187733226352268
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,768,1024,0.005954666684071223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,7168,512,0.012180444267061023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,8192,16384,0.02611644400490655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,12288,512,0.006598222172922558
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,10240,16384,0.05346222387419807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,10240,5120,0.023016888234350417
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,12288,3072,0.03601600064171685
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,3072,7168,0.029515554507573444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,3584,7168,0.03827644387880961
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,12288,2560,0.031202667289310034
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,128,12288,0.013081777426931592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,256,6144,0.016172443826993305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,12288,512,0.006592000110281839
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,5120,7168,0.34471999274359805
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,512,8192,0.04095822241571214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,3072,6144,0.012796444197495779
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,64,1024,0.002867555452717675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,128,7168,0.1131324503156874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1024,12288,0.011375111010339526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,512,12288,0.0647546648979187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,5120,5120,0.050126221444871694
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,16384,12288,0.08488888872994317
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,10240,12288,0.03976444403330485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,768,8192,0.006325333482689328
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,512,3584,0.007603555917739868
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,768,4096,0.008359111017651027
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,16384,5120,0.020879111356205408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,64,5120,0.006609777609507243
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,3584,7168,0.031153778235117596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,10240,3584,0.01311911145846049
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,512,768,0.0028808888875775864
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,4096,2560,0.029642668035295274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,16384,4096,0.02344177828894721
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2560,4096,0.009359999663299983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,65536,6144,0.06754222181108263
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,10240,8192,0.14130311542087132
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1024,3072,0.005653333332803514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,256,2560,0.017596445149845548
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,16384,2560,0.01682133310370975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,4096,256,0.0052791109515561005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,128,7168,0.017843556072976854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,2560,4096,0.008012444608741337
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,10240,65536,0.1432408889134725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,768,256,0.0052782222628593445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,512,6144,0.01796000036928389
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,16384,1024,0.03717422154214647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,768,2048,0.008166222108734978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,768,256,0.00573244442542394
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,256,10240,0.012108444339699216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,12288,65536,0.16265866491529676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,7168,64,0.006010666489601135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,3072,5120,0.02363288899262746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2560,65536,0.05512533254093594
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,65536,512,0.015260444747077094
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,7168,64,0.0032640000184377036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,768,65536,0.06250133117039998
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,256,1536,0.012885332935386233
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,12288,7168,0.030811554855770532
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,10240,2048,0.011180444724029966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,512,1024,0.012864000267452665
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,4096,65536,0.06738222307629056
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,16384,2048,0.07997333341174655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,12288,16384,1.1880728403727214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,16384,512,0.0069413334131240845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,65536,768,0.01607377827167511
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,256,128,0.010103999740547603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,512,4096,0.018615111708641052
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,16384,5120,0.028386665715111628
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2048,2560,0.006726222319735422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,3072,12288,0.029078221983379785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,32,512,0.00565422202150027
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2560,768,0.00590844452381134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,256,2560,0.004975999808973736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,16384,65536,0.21020800537533232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,65536,3072,0.05975466966629028
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3072,512,0.00360977773865064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,256,8192,0.012225777738624148
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,65536,7168,0.05969866779115465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,65536,64,0.02847733431392246
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,2048,512,0.014458666245142618
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3072,5120,0.012291555603345236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,1536,12288,0.026872000760502283
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,2560,3072,0.023741333021057978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,65536,10240,0.16152177916632757
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,768,128,0.002198222196764416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,3072,5120,0.008316444853941599
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,65536,5120,0.2366480032602946
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,65536,2048,0.08561155531141494
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,16384,256,0.006255110932721033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,2560,4096,0.00794400026400884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,128,5120,0.009334222310119206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,256,7168,0.005909333212508096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,128,6144,0.016900445024172466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1024,3584,0.008047111332416534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2048,4096,0.01311466677321328
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,8192,65536,0.09053599834442139
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,512,1536,0.007415999968846639
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,12288,10240,0.026888888743188646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1536,1024,0.004406222038798862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3584,7168,0.015256888336605497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,512,2560,0.014349333114094205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,3072,6144,0.01346577786737018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,65536,1536,0.02926666537920634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3072,768,0.0041706665522522396
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2560,64,0.0070017774899800616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,8192,16384,0.0470106667942471
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,4096,256,0.005941333456171884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,2560,16384,0.020831111404630873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,65536,7168,0.06338044669893053
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3072,8192,0.24037244584825304
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,10240,3584,0.014051554931534661
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,256,512,0.0052488889131281115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2560,7168,0.01977511081430647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,768,1024,0.003192000091075897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,65536,7168,0.06299377812279595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3072,5120,0.0240666667620341
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,768,1024,0.005896889087226655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1536,64,0.005888888819350137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2048,65536,0.6383928722805446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2560,768,0.027468444572554693
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,768,10240,0.011254222028785281
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,6144,3072,0.010310221877362993
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1536,3584,0.022091555926534865
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,10240,2560,0.012111999922328524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,1536,12288,0.016938666502634685
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,5120,768,0.016955556141005624
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,512,5120,0.009026666482289633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,7168,128,0.0025439999169773525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,128,32,0.0062871111763848206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1536,6144,0.009443555441167619
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,64,64,0.0019004444281260173
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,6144,1024,0.003544888976547453
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,12288,1024,0.0069466663731469055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,32,3072,0.03189599845144484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,8192,3072,0.05735911263359917
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,10240,10240,0.029841777351167467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,7168,1536,0.006288888967699475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,16384,4096,0.0188426673412323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,128,2048,0.01239733315176434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,5120,768,0.0058906666106647914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,1536,768,0.005552000055710475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,6144,6144,0.01239377756913503
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3584,12288,0.022413333257039387
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,4096,65536,0.06880711184607612
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,7168,65536,0.7386630905999078
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,768,1024,0.005559111220969095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2560,8192,0.014444443914625378
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,4096,7168,0.014352889524565803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,128,256,0.005643555687533484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,4096,3584,0.019567110472255282
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,2048,768,0.01648177703221639
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,10240,8192,0.025204444924990337
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,1536,1024,0.015538665983412
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,10240,2048,0.011069333387745751
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3584,64,0.002491555487116178
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1536,3072,0.005938666562239329
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,10240,8192,0.029319110843870375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3584,256,0.005568000177542369
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1024,8192,0.044703112708197705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,2048,3072,0.007287110719415877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,4096,6144,0.02625333269437154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,10240,3584,0.026363554928037856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,4096,6144,0.03204800022972955
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,128,4096,0.014442667365074158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,7168,512,0.007308444215191736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,1536,2560,0.021049777666727703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,7168,768,0.003524444583389494
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1024,128,0.005416888743638992
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,128,2048,0.011711111499203576
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,1024,3584,0.02094399929046631
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,128,3584,0.007677333222495184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,1536,6144,0.009020444419648912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,16384,256,0.009343110852771336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,7168,2048,0.010594666832023196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,6144,5120,0.012821333275900947
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1024,3072,0.008051555189821456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,768,128,0.006725333217117522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,6144,64,0.0022168887986077205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,2560,12288,0.027739556299315557
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,12288,3072,0.027519999278916255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,65536,8192,0.15309600035349527
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,65536,256,0.013594666288958656
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3584,6144,0.01314311143424776
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,2048,1024,0.01382844481203291
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,8192,2560,0.0103502223889033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,7168,768,0.028705779049131606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,3072,1024,0.014997333288192749
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,256,128,0.005922666854328579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,768,6144,0.0068488890926043195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,5120,1536,0.008598222500748104
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1024,512,0.003117333269781537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,32,256,0.005884444548024072
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,7168,1536,0.008683555656009251
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,12288,512,0.006973333656787872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,1024,2560,0.01752088963985443
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,1536,512,0.014197332991494073
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,768,512,0.006611555400821898
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,4096,12288,0.02053155501683553
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2048,4096,0.009022222624884712
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,256,10240,0.012091555529170565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,512,5120,0.006639110959238476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1536,256,0.0028373334142896864
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2048,16384,0.021339555581410725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,12288,1536,0.00871377769443724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3584,1024,0.0031840000301599503
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,12288,65536,7.0919138590494795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,2048,5120,0.04596799943182203
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,768,2560,0.00702044450574451
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,128,8192,0.00564533347884814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,16384,2048,0.06338400310940213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,65536,64,0.009728000395827824
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,8192,32,0.024482667446136475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,768,10240,0.007762666377756331
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,256,256,0.00527555536892679
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,1024,12288,0.12593244181738958
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,16384,3584,0.0159599996275372
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,10240,8192,0.12412177191840278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,2048,4096,0.022755554980701868
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,7168,7168,0.0325217776828342
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,65536,7168,0.07826489210128784
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,1536,8192,0.010639111200968424
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,16384,1536,0.010217777556843227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,3584,512,0.014154665999942355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1024,65536,0.024672889047198828
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3072,12288,0.017802667286660936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,128,512,0.011791110866599612
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,65536,1536,0.019001776973406475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1024,6144,0.009791110952695211
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,8192,32,0.004907555464241239
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,32,6144,0.011774222056070963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3072,512,0.0070471109615431885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,16384,768,0.011254222028785281
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,32,128,0.005531555662552516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2048,6144,0.008871999879678091
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3072,64,0.001822222231162919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,5120,1536,0.006477333191368315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,8192,4096,0.014739554789331226
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,6144,512,0.005897777775923411
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,2048,5120,0.00903733323017756
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,65536,2560,0.031726221243540444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,64,256,0.00795022232664956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,10240,128,0.006593777901596493
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,3584,12288,0.029139555162853662
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,7168,2048,0.006621333460013072
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1024,8192,0.008989333278603023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3584,2048,0.00544177782204416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,10240,768,0.0068897778789202375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,6144,64,0.005567111074924469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,3584,16384,0.02088622252146403
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,7168,7168,0.03156977891921997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3584,32,0.005562666803598404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2560,10240,0.016944888565275405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,2048,3584,0.01587199999226464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,768,32,0.005234666582610872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,5120,3584,0.013183111117945777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,4096,3072,0.11267199781205918
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2560,3072,0.007322666545708974
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,256,32,0.005212444398138259
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1024,7168,0.01071999967098236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,65536,5120,0.4549751281738281
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,4096,768,0.005615111026499007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,12288,128,0.018269333574506972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,1536,128,0.005445333404673471
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,5120,8192,0.30951378080579967
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,16384,1536,0.011056889262464313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,4096,1536,0.006631111105283101
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,768,2560,0.010656000011497073
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,512,5120,0.012773333324326409
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,10240,4096,0.01774577796459198
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,1024,7168,0.01923466722170512
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,16384,64,0.006236444330877728
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,6144,2560,0.01976355579164293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,5120,5120,0.015830222103330824
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,64,1536,0.006275555739800136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1536,128,0.0025066667132907445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,2048,4096,0.007617777420414819
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,256,2048,0.010439111126793755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,16384,768,0.006903111106819577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,128,256,0.011711111499203576
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1536,2560,0.008391111261314815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,256,65536,0.09800533453623454
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,2048,65536,0.06370577547285292
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,4096,3072,0.0074888889988263445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,768,512,0.008623110751310984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,128,5120,0.008621333373917473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1536,12288,0.02586666742960612
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,5120,2560,0.008083555433485243
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,12288,2560,0.05522666705979241
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,2560,2560,0.03257955445183648
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,2048,7168,0.010033778018421596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,5120,12288,0.028516444895002577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,65536,256,0.03376888897683885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,8192,8192,0.3161582152048747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,8192,7168,0.014149333039919535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,16384,16384,0.8898799684312609
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,4096,128,0.014878221684032016
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2560,1024,0.009677333136399588
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,3072,1024,0.006177777631415262
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,256,32,0.005631111148330901
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,768,65536,0.36765599250793457
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,32,1024,0.0032186667538351486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,7168,8192,0.015528000063366361
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,3584,5120,0.010710222025712332
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,128,3584,0.014469333820872836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,2048,2560,0.006999111009968652
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,16384,64,0.008344888687133789
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,10240,12288,0.041420443190468684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3584,2560,0.02885155545340644
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,256,128,0.002209777840309673
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3584,32,0.0025964443468385274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,1024,10240,0.022082666556040447
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1536,3072,0.006239999913507038
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2560,3072,0.0075777777367168
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,16384,8192,0.03432888785998026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,3584,3584,0.013986666997273764
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,2048,10240,0.013995556367768182
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,64,7168,0.010096000300513374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,7168,12288,0.019164444671736825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1024,65536,0.11671466297573513
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2048,5120,0.010823999842007955
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,32,1536,0.00328177772462368
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,2560,2560,0.01866755551762051
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,5120,3072,0.00850311087237464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,12288,10240,0.04480977853139242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,4096,768,0.012343111137549082
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,768,16384,0.022351110974947613
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,768,65536,0.07450133562088013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,12288,256,0.01684799955950843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,6144,7168,0.01719288859102461
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2048,7168,0.035353776481416486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,128,16384,0.015666666958067153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,768,1536,0.015127999915017022
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,65536,768,0.030493332280053034
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,16384,7168,0.07281333208084106
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,4096,768,0.0032133333798911837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,5120,12288,0.20161067114935982
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,8192,4096,0.025047111842367385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,12288,8192,0.03304622239536709
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,7168,3072,0.028322666883468628
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,7168,8192,0.029583109749688044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,10240,5120,0.020431111256281536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,6144,768,0.007291555404663086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,2048,512,0.014436443646748861
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1024,128,0.005954666684071223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3072,2048,0.00637155564294921
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,128,512,0.006663110935025745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,2560,65536,0.08669777711232503
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,3584,128,0.005576000031497743
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,4096,12288,0.017676444517241586
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,10240,16384,0.08853511015574138
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2048,10240,0.016173332929611206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,8192,768,0.01687111126052009
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1536,8192,0.12365510728624131
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,32,6144,0.005105777747101254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,6144,3072,0.02208799951606327
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,32,7168,0.005229333208666907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,65536,8192,10.655338711208767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,64,128,0.012051555845472546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,5120,3072,0.020875554945733812
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,1024,256,0.01479555500878228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1536,3584,0.006986666884687211
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,1024,512,0.01313777764638265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,64,768,0.002893333426780171
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,65536,1536,0.019359999232821994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,512,512,0.002546666603949335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,128,2560,0.004588444613748127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,8192,3584,0.011415111521879831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,12288,1536,0.027650665905740526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3072,3584,0.007640889121426477
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,16384,1024,0.05126133229997423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,1024,1024,0.014827556080288358
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,6144,5120,0.04072533382309808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,65536,256,0.013422222601042854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,2048,2048,0.006671111202902264
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3072,3584,0.009689778089523315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,256,10240,0.011703111231327057
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,10240,2560,0.011951111257076263
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,6144,768,0.007864889171388414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,65536,768,0.06899822420544095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,256,65536,0.06268178092108832
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,512,256,0.014500444134076437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,3072,16384,0.07545689079496595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,256,256,0.0021626665774318907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3072,16384,0.02104266650146908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,64,1536,0.003588444242874781
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,2560,16384,0.021555556191338435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,64,5120,0.004888888862397936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,3072,4096,0.007687110867765214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,768,16384,0.02678044471475813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,12288,10240,0.044066667556762695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,3072,128,0.015775110986497667
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,512,5120,0.00870666652917862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,4096,3584,0.018904889623324077
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,768,10240,0.012039110892348819
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,1536,1536,0.006260444306664997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,1024,512,0.005617777920431561
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2560,7168,0.029876444074842665
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,128,10240,0.011751111182901593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,16384,2048,0.011744000017642975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,6144,10240,0.016167110866970487
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,32,512,0.005584888988071018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,16384,7168,0.032758222685919866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,256,1536,0.006276444428496891
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,2560,768,0.005649777750174205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,16384,256,0.07139733102586535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,3072,65536,0.1346257792578803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,10240,768,0.0069448889957533936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,128,256,0.005227555417352253
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,4096,1536,0.004609777695602841
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,3072,128,0.015844444433848064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,6144,4096,0.012048888537618848
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,512,6144,0.009684444301658207
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,2560,16384,0.06980977455774943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,128,256,0.005602666487296422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,768,32,0.002096000024014049
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1024,512,0.002777777819169892
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,5120,256,0.015299555328157214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3584,1536,0.006295111030340195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,10240,16384,0.05132444368468391
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,64,1024,0.0032151111712058387
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3584,1024,0.006280888699822956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,128,3584,0.02093244426780277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,6144,65536,0.15499288505978054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,8192,2560,0.010055999788973067
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,6144,64,0.006666666517655055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2048,12288,0.011185777684052786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,1536,1024,0.018844443890783522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,65536,512,0.012755555411179861
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1536,3072,0.005479111025730769
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,2560,768,0.016952888833151925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,10240,2048,0.009740444521109263
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,64,2560,0.004959110998445087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,8192,128,0.008301333420806462
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,5120,7168,0.06640177965164185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,2048,768,0.01789688898457421
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3584,6144,0.01761244402991401
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,5120,12288,0.0620888868967692
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,65536,10240,0.09506311019261678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,8192,2048,0.009093333449628618
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,64,6144,0.004983110974232356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,128,256,0.005620444400442972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,256,2048,0.013537777794731988
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,8192,1536,0.009465778039561378
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1536,12288,0.09464444716771443
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,3072,6144,0.014616888430383472
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,1024,16384,0.16666577921973333
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,1024,7168,0.02586488922437032
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,32,5120,0.004956444518433677
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,1536,128,0.0053093334039052325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,65536,10240,0.20604888598124185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,1536,768,0.01737955543729994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2048,256,0.005593777944644292
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,256,5120,0.005110222018427319
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2560,3584,0.01926577753490872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,4096,256,0.01683022247420417
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,256,3584,0.006037333359320958
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2048,3584,0.00666133314371109
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,2048,768,0.0058559998869895935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,8192,1024,0.0063031112982167145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,256,2560,0.005294222384691238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,3072,6144,0.03425600131352743
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,3584,7168,0.027264888087908428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,3072,2048,0.006568888823191325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,512,128,0.003100444459252887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,2048,3072,0.007346666521496243
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,768,3584,0.007607999775144789
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,10240,64,0.0031679999083280563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2048,64,0.0021466666625605687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,8192,16384,0.22147822380065918
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,7168,768,0.005912000106440649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,8192,12288,0.06484711170196533
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,12288,7168,0.047727998760011464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,512,128,0.013412444127930535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,65536,3584,0.06573600239223905
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,5120,1024,0.02843911117977566
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,512,256,0.014446222119861178
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,32,32,0.0018560000591807896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,512,16384,0.0119991112086508
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,128,512,0.002469333302643564
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,5120,2048,0.012574222352769641
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,3584,5120,0.010461333725187512
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,512,16384,0.016148444679048326
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,3072,12288,0.017601778109868366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,7168,2560,0.008960888617568547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,2048,768,0.006587555425034628
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,512,3584,0.013089777694808112
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2048,768,0.018783110711309645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,256,2560,0.016656888855828177
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,2560,512,0.014479999740918478
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,768,16384,0.02093155516518487
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,8192,32,0.006311111152172089
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3072,3584,0.007642666498819987
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,4096,65536,0.24766755104064941
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,65536,10240,0.2157342169019911
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,4096,256,0.005567111074924469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3584,1024,0.0047457776963710785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,3584,1024,0.016527111331621807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,12288,1024,0.024892444411913555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,7168,2048,0.02204711072974735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,12288,256,0.006221333311663733
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2048,6144,0.009720889230569204
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,2560,65536,0.07202221949895223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,128,32,0.0018382221460342407
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,12288,16384,0.1470817724863688
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,16384,65536,0.6686408784654406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,16384,32,0.13184266620212132
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,128,16384,0.013091555900043912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1536,3072,0.012433778080675336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,512,1536,0.011723555624485016
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,512,768,0.013759111364682516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,7168,1024,0.0690408878856235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,8192,128,0.01475288967291514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,2560,12288,0.04601777924431694
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,1024,2560,0.017871111631393433
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,256,6144,0.022336888644430373
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,1536,128,0.014869333969222175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,1536,768,0.01646844380431705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1024,1024,0.0031884445084465873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,3584,5120,0.012713777522246042
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2560,4096,0.00794044468137953
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,256,1024,0.0029964444951878656
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,10240,6144,0.019876443677478366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,5120,256,0.0052382221652401825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,2560,65536,0.12632621659172907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3584,768,0.0038968887594011095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,768,6144,0.014499555031458536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,7168,1536,0.009958222508430481
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2048,2048,0.043475555049048535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1536,512,0.005621333503060871
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,8192,16384,0.05076355404324002
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1024,1024,0.0031928889867332247
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,3072,10240,0.040050665537516274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,4096,4096,0.007677333222495184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,3072,3072,0.017991110682487488
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,12288,64,0.006334222025341458
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,512,65536,0.07759822077221341
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,8192,32,0.0068488890926043195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,1536,1024,0.015839111473825242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,1024,2560,0.01789955629242791
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,5120,512,0.01570755574438307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1024,128,0.0022595555831988654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1024,7168,0.010040889183680216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,768,4096,0.008016889293988546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,10240,3072,0.012820444173283048
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,256,2048,0.012327999704413943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,10240,512,0.006648889018429651
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,16384,65536,0.1705626646677653
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,6144,4096,0.01091733326514562
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,12288,1536,0.007975111405054728
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,128,5120,0.008912888665994009
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2560,65536,0.13007022274865046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,6144,10240,0.017439110411538016
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1536,64,0.005555555638339784
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,128,5120,0.008449777960777283
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,2048,2560,0.00685777763525645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,6144,1024,0.01756533318095737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,768,3072,0.005571555760171678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,65536,5120,2.8584906260172525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,4096,12288,0.04227822356753879
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,6144,1024,0.01387466655837165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,64,4096,0.005954666684071223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,5120,3584,0.009169778062237633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,3072,512,0.014828445182906257
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,512,5120,0.006702222343948152
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,7168,65536,0.07456800010469225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,4096,7168,0.022896000080638464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3584,1024,0.004599110947714912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,7168,12288,0.07939910888671875
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,5120,32,0.0069004446268081665
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,10240,5120,0.08619288603464763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,4096,12288,0.02648799949222141
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,65536,16384,0.3467546568976508
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,128,1536,0.00868711123863856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,16384,3072,0.015379556351237826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,2048,5120,0.13141866525014242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,5120,4096,0.011651555697123209
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,8192,12288,0.049570666419135205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,7168,2048,0.012862222062216865
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2048,7168,0.00870311094654931
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,128,12288,0.007635555333561367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,12288,1024,0.007598222129874759
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,2048,16384,0.033470223347345986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,1024,12288,0.017688888642523024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,8192,6144,0.013246222502655454
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,32,2560,0.005158222383923001
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,3584,1024,0.006198222024573221
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1536,65536,0.058918224440680615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,128,64,0.005683555371231503
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,1024,6144,0.019271110494931538
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,5120,2560,0.050218668248918324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3584,2048,0.006584888945023219
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,256,512,0.008287111090289222
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,768,5120,0.017328000730938382
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,768,16384,0.0158924443854226
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,768,12288,0.02402755618095398
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,768,12288,0.013442666994200813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,32,128,0.0021048889805873237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3584,64,0.006144888699054718
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1024,4096,0.05021333363321093
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,3584,6144,0.036828445063696966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,4096,6144,0.013796444568369122
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1024,64,0.0021155555215146807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2048,8192,0.021218667427698772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,6144,1024,0.006575999988449945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,4096,3584,0.021034666233592566
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,4096,64,0.003588444242874781
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,1024,2048,0.006727111008432176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,6144,65536,0.1228053304884169
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,3072,128,0.005294222384691238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3072,8192,0.015626667274369132
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1536,768,0.007327111230956183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,4096,1024,0.00590133335855272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,128,2048,0.006641777853171031
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3072,7168,0.012824888858530255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,12288,1024,0.006308444258239534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,3584,768,0.005583111196756363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,64,256,0.008360000120268928
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2048,1536,0.03408711155255636
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,5120,4096,0.009567110902733272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,4096,6144,0.0096133334769143
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,10240,4096,0.014507555299335055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,65536,2048,0.051905777719285756
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,768,512,0.024311110377311707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,16384,3072,0.017431111799346078
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,3072,65536,0.06541777981652154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,2048,256,0.015172445111804538
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,512,3584,0.007637333538797166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,12288,2048,0.22899288601345488
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3584,7168,0.013872000078360239
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,8192,1024,0.006288888967699475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,65536,12288,0.3768044312795003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,512,16384,0.030151999659008447
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,5120,5120,0.032242665688196816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1536,12288,0.017840888765123155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,7168,2560,0.16792444388071695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,16384,16384,0.06680088573031955
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,8192,768,0.030333333545260962
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3584,8192,0.01571377780702379
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,10240,128,0.002191999927163124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,7168,65536,5.650946723090278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,6144,32,0.010686222049925061
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,16384,3584,0.061518225404951304
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,5120,10240,0.022063111265500385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,512,8192,0.02016533414522807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,16384,10240,0.07449155383639865
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,64,12288,0.004866666677925321
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,3072,3072,0.01616533266173469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,8192,32,0.002130666747689247
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1024,8192,0.012076444096035428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,256,4096,0.007859555383523306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3072,2560,0.0074577778577804565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,4096,8192,0.0420720014307234
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,7168,3584,0.011023110813564725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,6144,256,0.00282577777074443
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,10240,512,0.004126222183307012
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,5120,3584,0.010018666585286459
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1536,64,0.002179555594921112
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,3072,2560,0.02238222294383579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,64,7168,0.010024000373151567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,768,6144,0.00868088917599784
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,512,10240,0.011732444167137146
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2560,128,0.006270222365856171
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,768,2048,0.0069422221018208405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,128,16384,0.025383111503389146
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,16384,1536,0.011130666567219628
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1536,2048,0.006950221955776215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,32,128,0.005859555469618903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,65536,1536,0.021632888250880774
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,65536,5120,0.10102399852540757
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,65536,1024,0.2360773351457384
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,2048,128,0.006583111153708563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,7168,128,0.005873777800136142
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,7168,2048,0.03316355413860745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,32,12288,0.008771555291281806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,32,7168,0.015779554843902588
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,3072,768,0.018246221873495314
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,256,7168,0.01754844519827101
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,8192,256,0.016911110944218107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,16384,128,0.028628445333904688
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,16384,16384,0.04794666502210829
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,768,5120,0.00868711123863856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,10240,512,0.02442311081621382
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,12288,1536,0.02682311170630985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,16384,2048,0.03772799836264716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3072,32,0.0021679999513758556
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1536,7168,0.011332444846630096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,4096,10240,0.03817955652872721
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,128,3072,0.007244444555706448
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,12288,1536,0.009173333644866943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,65536,5120,0.12513689200083414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,8192,1536,0.029866665601730347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,256,512,0.006110222389300664
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,2560,2048,0.006477333191368315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,256,12288,0.013421333498424955
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,2560,5120,0.02533422244919671
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,5120,2048,0.04682666725582547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,7168,65536,0.09197066889868842
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,32,2048,0.004569777597983678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,128,2560,0.007234666910436418
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,12288,3584,0.014087110757827759
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,4096,128,0.005521777603361342
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,12288,7168,0.019938665959570143
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3072,4096,0.00795911086930169
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,10240,1536,0.024471999870406255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1024,128,0.005511111269394557
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,4096,1024,0.011824888487656912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,2560,3584,0.025913778278562758
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,10240,3072,0.01959288948112064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,16384,5120,0.1882906622356839
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,10240,16384,2.6531528896755643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,7168,3584,0.03331466515858968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,256,10240,0.01791111131509145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,768,7168,0.01593777702914344
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,32,2048,0.009053332938088311
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,3072,65536,0.06495377752516004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,256,256,0.008112000094519721
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,10240,65536,0.1356382237540351
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,768,32,0.005571555760171678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,10240,1536,0.008360000120268928
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1536,1024,0.004238222208287981
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,7168,1024,0.02218311031659444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,16384,1024,0.02684088879161411
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,2048,1536,0.006286222073766921
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,512,7168,0.06814755333794488
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,4096,65536,0.2958470980326335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,2048,5120,0.02949777907795376
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,3072,2560,0.006904888898134232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,3584,65536,0.12973777453104654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,8192,32,0.006248888870080312
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,512,6144,0.009290667043791877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,12288,6144,0.03809955716133118
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,2048,4096,0.027112000518374976
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,512,128,0.006629333313968446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,7168,7168,0.05345955491065979
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,2560,5120,0.0090524446633127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,5120,768,0.00318755561278926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,256,65536,0.17178755336337617
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,7168,4096,0.06884444422192044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1024,768,0.0033448889023727844
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,12288,1536,0.03906133439805772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,3072,256,0.005199111170238919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1536,32,0.005594666633341048
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,4096,128,0.002157333410448498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,4096,12288,0.017850667238235474
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3072,12288,0.01585422290696038
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,5120,6144,0.036350223753187395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,8192,6144,0.03906311260329352
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,10240,8192,0.027480888697836135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,2048,1024,0.014785778191354541
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,6144,16384,0.061399108833736844
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,128,6144,0.01715288890732659
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,512,4096,0.007381333245171442
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,12288,3072,0.02327377763059404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,8192,16384,0.03081866767671373
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,32,768,0.006993778049945831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,12288,1536,0.026586666703224182
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,768,768,0.005959110955397288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,5120,8192,0.017870222528775532
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3584,12288,0.019856888386938307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,16384,768,0.0069573331210348345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,10240,128,0.004218666503826777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,1024,8192,0.02028888960679372
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3072,2048,0.06368622514936659
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2048,8192,0.012837332983811697
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2048,65536,0.0428071121374766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,7168,7168,0.03155733479393853
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,1024,6144,0.01792622274822659
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,16384,2048,0.011436444189813403
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,6144,2048,0.018059555027219985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,4096,3584,0.013253333667914072
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,6144,1024,0.017602667212486267
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,12288,32,0.008748444418112436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2560,128,0.0022088889446523455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,12288,1024,0.0070471109615431885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,32,64,0.0018702221827374564
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,128,5120,0.008394666843944127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,2048,128,0.01556622154182858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,1024,256,0.014856000741322836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,16384,6144,0.06901422474119398
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,12288,65536,0.1461448934343126
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1024,64,0.0021964444054497615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,12288,2560,0.01847555571132236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,256,16384,0.01717066599263085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,512,2560,0.014487110906177096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,65536,768,0.019703111714786954
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3584,1536,0.004338666796684265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1536,65536,0.03402400016784668
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,8192,512,0.006256000035338932
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1536,8192,0.014256000518798828
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,65536,3584,0.0916657778951857
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,3072,7168,0.011644444531864591
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,8192,128,0.0021742222209771476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,512,8192,0.013069333301650153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,6144,65536,0.13256888919406468
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,128,256,0.0022746666024128595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,6144,6144,0.014479999740918478
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,256,7168,0.005607111172543631
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,16384,128,0.0022026666750510535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,128,64,0.006589333216349284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3584,3072,0.009361777868535783
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,6144,5120,0.012680889003806643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,32,768,0.0028400001012616684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1024,2560,0.00554666668176651
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1536,32,0.002022222305337588
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,2048,256,0.005575110928879843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,5120,10240,0.01739911072784
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,6144,6144,0.018255111243989732
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3584,10240,0.028991109795040552
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,512,32,0.0020435555941528743
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2560,5120,0.009004444711738164
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,10240,16384,0.054773333999845714
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2560,10240,0.015503999259736804
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,16384,256,0.018368888232443068
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,65536,12288,0.23113510343763563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,65536,3072,0.10163733032014634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,128,128,0.005214222189452913
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,8192,4096,0.013234666652149625
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,10240,128,0.008960000342792934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,768,2048,0.014861333701345654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,3584,128,0.014560000763999091
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,512,4096,0.00629333323902554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,65536,32,0.013967111706733704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3584,12288,0.2329191101921929
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,512,4096,0.00962222201956643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,128,128,0.006923555499977536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1536,512,0.004512000001139111
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,4096,128,0.0063164445261160536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3072,1024,0.035411556561787925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,12288,1536,0.007975111405054728
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,8192,4096,0.013144888811641268
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2560,10240,0.13028889232211643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,4096,256,0.006244444598754247
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,7168,8192,0.017625777257813346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,8192,512,0.0032248888164758682
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,128,3584,0.007608888877762689
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,5120,6144,0.026160000099076167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,16384,1024,0.008582221964995066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1536,7168,0.00927022182279163
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,4096,8192,0.12270400259229873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2048,12288,0.017271111408869427
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,3584,65536,0.08789333369996812
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,3072,1024,0.005969777703285217
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,256,2560,0.014130666851997375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,12288,2560,0.14860533343421087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,256,8192,0.010696888797812991
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,1024,12288,0.013447110851605734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,128,128,0.012741333080662621
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1536,65536,0.045647998650868736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,16384,4096,0.022686221533351477
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,256,4096,0.01748533381356133
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,768,3072,0.006119110931952794
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,10240,2560,0.026974222726292078
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,12288,7168,0.07297244336869982
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,3584,4096,0.007927999728255803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,16384,4096,0.04180977741877238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,4096,2048,0.006623111251327727
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,7168,65536,0.07522666454315186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,4096,7168,0.013170666992664337
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2048,512,0.005681777579916849
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,16384,16384,0.045287112394968666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,2560,65536,0.09720444679260254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,128,1536,0.010231110784742568
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,2048,3584,0.007532444265153673
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,7168,3072,0.007927999728255803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,6144,3072,0.01278133359220293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,3584,3072,0.007291555404663086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,2560,8192,0.08842577536900838
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,1024,1024,0.0059057776298787855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1536,16384,0.039177778694364764
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,10240,32,0.002534222271707323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,12288,10240,0.02645688917901781
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2048,2560,0.0057706667317284485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,1024,3584,0.007317333585686154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3072,4096,0.010447111394670276
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,1024,1536,0.006615110983451207
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,3584,8192,0.014519110321998596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,65536,12288,0.122761779361301
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,10240,3072,0.028076443407270644
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,128,2560,0.010382222632567087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,1536,512,0.0053173332578606075
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2560,6144,0.018616888258192275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,65536,1536,0.10896622472339207
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,5120,3584,0.030200001266267564
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,10240,128,0.017992888887723286
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,6144,10240,0.02347022294998169
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,1536,6144,0.030226667722066242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,6144,2048,0.007333333293596904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,1536,10240,0.2130675580766466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,256,128,0.007986666427718269
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1536,768,0.003863999827040566
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,32,64,0.0019511110666725372
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,10240,16384,0.03871555460823907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,4096,2048,0.01346666696998808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,768,8192,0.01998044384850396
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,7168,128,0.017495110630989075
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,10240,65536,0.14872177441914877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,3072,65536,0.06922844383451673
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,8192,4096,0.11047910981708103
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3072,65536,0.061921775341033936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,32,512,0.00589155571328269
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,128,65536,0.09779466523064508
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,10240,6144,0.02032177812523312
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,10240,65536,4.538431803385417
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,4096,65536,0.09482400284873115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,768,7168,0.005959110955397288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,7168,1536,0.006576889091067844
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3584,3584,0.012630222572220696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,12288,2048,0.08717599842283462
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,1024,512,0.005631111148330901
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,1536,16384,0.032999111546410456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,1536,65536,0.08858933051427205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,128,5120,0.008954666554927826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,10240,512,0.00865955568022198
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,4096,6144,0.01660711069901784
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,65536,16384,0.30506398942735463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,4096,12288,0.03198400139808655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2560,3584,0.007728888756699032
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,12288,6144,0.025578666064474318
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,6144,5120,0.010488889283604093
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,5120,1536,0.019567999574873183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,256,1024,0.005568888866239124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,16384,3072,0.014152889450391134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,16384,768,0.02161955502298143
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,256,65536,0.5209804640875922
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,768,768,0.016508445143699646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,10240,6144,0.2927466763390435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,10240,16384,0.039543999565972224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,128,16384,0.02578488820128971
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,7168,512,0.021736888421906367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2048,768,0.00463733325401942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,5120,12288,0.017975111802419026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,5120,128,0.03170933326085409
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,2560,8192,0.011101333631409539
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,4096,65536,1.2463484870062935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1536,32,0.018384888768196106
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,8192,4096,0.029225776592890423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,10240,12288,0.04411288764741686
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,256,12288,0.0051857779423395796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3584,3072,0.00833422193924586
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,1536,5120,0.027320888307359483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,768,8192,0.012256888879670037
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,6144,4096,0.012096000214417776
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,512,2560,0.006804444309737947
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,6144,7168,0.029380444023344252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,8192,2560,0.14825688468085393
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,768,4096,0.013144888811641268
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,3584,2048,0.015022221538755627
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,7168,7168,0.0431182218922509
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,65536,128,0.010024000373151567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,64,3072,0.007268444531493717
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,3584,2560,0.007310222420427535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,512,16384,0.017160000072585214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,768,256,0.0021608887861172357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,128,3072,0.015057777365048727
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,6144,10240,0.02650755478276147
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,16384,1536,0.017462222112549674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,256,65536,0.057632002565595836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,32,7168,0.005616000129116907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,128,512,0.002551111082235972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,2560,256,0.022170666191313002
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,16384,6144,0.02927288744184706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,7168,4096,0.032101333141326904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,128,64,0.0018702221827374564
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3072,3072,0.09244000249438816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,4096,128,0.0028168888141711554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,5120,128,0.002303999952144093
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,2560,12288,0.029389331738154095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,128,1536,0.012542222109105853
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,32,128,0.006630222416586346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,512,5120,0.009167999857001835
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,3072,10240,0.014049778381983439
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,256,3072,0.004675555560323927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,256,512,0.002526222210791376
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,10240,2560,0.16467289129892984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3584,16384,0.024988444315062627
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,8192,6144,0.015513777732849121
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,10240,2560,0.025416889124446448
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,2048,2048,0.007345777418878343
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,768,12288,0.024510223004553053
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,3584,1024,0.005664000080691443
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,1536,2560,0.017737777696715463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,2048,6144,0.009349333743254343
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1536,128,0.005471999860472149
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,1536,128,0.014542222023010254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,65536,3584,0.08601066801283096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1024,32,0.005611555443869696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,16384,8192,0.03995111253526475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,512,2048,0.0063457778758472866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,32,7168,0.0196106665664249
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,10240,1536,0.025407999753952026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,10240,1024,0.006649777707126405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,12288,5120,0.05727822250790066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,128,256,0.005591111050711737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1024,256,0.005399999933110342
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2048,7168,0.00870400004916721
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,12288,2048,0.011643555429246692
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,7168,2560,0.03947733177079095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,8192,3072,0.02444177700413598
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,5120,32,0.004956444518433677
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,7168,5120,0.02156977852185567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,2560,768,0.016857778032620747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2048,12288,0.01800622211562263
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,10240,8192,0.03096710973315769
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,768,7168,0.02254399988386366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2560,6144,0.01813155578242408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,4096,4096,0.023074666659037273
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,16384,4096,0.015519110692871941
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,512,1024,0.014667555689811707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,128,768,0.005912000106440649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,8192,7168,0.021728888154029846
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1024,16384,0.19976977507273355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,3584,1024,0.006650666809744305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2560,1536,0.03856177793608771
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,5120,12288,0.02793244520823161
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,12288,8192,0.08459199799431695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1536,256,0.002526222210791376
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,768,256,0.002556444456179937
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,6144,5120,0.03473155697186788
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,4096,8192,0.1185422208574083
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,768,4096,0.008013333711359236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2560,7168,0.018583999739752877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,768,512,0.002530666689078013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,12288,7168,0.023017777336968318
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,10240,2048,0.025475554996066626
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,12288,256,0.006636444479227066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,16384,2048,0.016226665841208566
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,4096,12288,0.024041778511471216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,16384,256,0.021016889148288306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3584,8192,0.015336889359686108
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,10240,16384,0.0654017792807685
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,10240,64,0.0022008888837363985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,256,256,0.0025742221623659134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,1024,256,0.014828445182906257
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,8192,12288,0.06604622470008002
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,2048,16384,0.030738668309317693
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,8192,7168,0.042768889003329806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3584,12288,0.03519644339879354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,65536,6144,0.11957421567704941
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,1536,12288,0.013454222016864352
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,512,3072,0.0052888890107472735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,12288,16384,0.08136000235875447
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3072,256,0.006620444357395172
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,64,3584,0.006030222194062338
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,16384,5120,0.24590489599439833
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,128,32,0.0019360000474585427
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,768,8192,0.010500444306267632
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,64,3584,0.007632888853549957
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1536,3072,0.012485333614879184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,768,256,0.002891555635465516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1536,64,0.0026177778426143858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,7168,256,0.005624889085690181
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,512,128,0.005312888986534542
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,6144,2560,0.022106667359670002
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,1024,128,0.014471999473041959
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,7168,4096,0.009362666971153682
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,12288,65536,0.1688471105363634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,256,2048,0.006952889263629913
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,768,256,0.006334222025341458
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,2560,3584,0.024774221910370722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,768,12288,0.07587733533647326
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,4096,4096,0.010006222460005019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,768,2560,0.006960888703664144
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,12288,4096,0.014183110660976835
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,2560,4096,0.01717066599263085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,12288,3584,0.2592382166120741
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,128,10240,0.012411555482281579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,2560,512,0.005551110953092575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,6144,256,0.005682666682534748
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,1536,2560,0.01793866687350803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,12288,7168,0.030601776308483545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,128,3584,0.00490488898422983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,16384,512,0.027804444233576458
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,512,3072,0.007280888656775157
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,64,65536,0.11776088343726264
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2560,16384,0.020957332518365648
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,768,4096,0.04827377862400479
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,128,512,0.013389333254761167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,4096,1536,0.007044444481531779
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,12288,2048,0.011370666325092316
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,256,64,0.0019191110299693213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,7168,3072,0.194724440574646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,10240,1536,0.008170666793982187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,7168,65536,0.1228133307562934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,12288,8192,0.07016177972157796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,128,65536,0.009901333186361525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,1024,768,0.011078221930397881
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,8192,65536,0.31040000915527344
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,5120,2560,0.007337777978844113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,3072,1024,0.010747555229398938
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,16384,8192,0.08217333422766791
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,16384,4096,0.037351999017927386
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,12288,32,0.038471112648646034
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3072,8192,0.01574844453069899
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1024,3072,0.005601777798599667
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3072,4096,0.0087333329849773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,10240,12288,0.043939554029040866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2560,7168,0.09254666831758286
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,5120,3072,0.10808444685406154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,5120,5120,0.026104888982243005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,2560,65536,0.14898044533199734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1024,7168,0.041011555327309504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,3584,10240,0.027452444036801655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,4096,7168,0.038535998927222356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,16384,128,0.018216000662909616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,128,8192,0.01035822182893753
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,256,3584,0.0203475554784139
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,2560,7168,0.023344889283180237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,16384,8192,0.04727022184265984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,3584,7168,0.09682666593127781
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3072,3584,0.048635555638207324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,2048,1536,0.010520888699425591
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,1536,3072,0.02270577847957611
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,512,1536,0.006583111153708563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2048,7168,0.0087013335691558
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,2560,5120,0.019262222780121695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,512,12288,0.013398222625255585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,16384,1536,0.017690666847758822
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,64,4096,0.006271111054552927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,10240,12288,0.04379111197259691
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3072,512,0.0031591111587153543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,8192,1024,0.008385777473449707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,10240,5120,0.034625778595606484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,12288,6144,0.06460799773534139
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,7168,7168,0.02906755606333415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,12288,65536,0.15838666756947836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,4096,16384,0.021568889419237774
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,10240,6144,0.02490577763981289
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,768,65536,0.06340799729029338
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,5120,256,0.0021964444054497615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1536,3584,0.010747555229398938
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,5120,4096,0.1404728889465332
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2560,512,0.0058808889653947614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,128,6144,0.009328000247478485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,65536,1536,0.030260443687438965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,32,16384,0.015184889237085978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,7168,7168,0.01533777846230401
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,1024,2560,0.014881778094503613
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,64,12288,0.006931555353932911
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,16384,8192,0.04289866818322075
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,4096,256,0.016183111402723525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,768,65536,0.18075199921925864
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,65536,2560,0.056941330432891846
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,3584,256,0.005988444305128521
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,16384,7168,0.09658933348125881
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,4096,4096,0.06754399670494927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3072,5120,0.015308444698651632
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,5120,5120,0.00941777808798684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,2560,2560,0.00813244448767768
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,1024,128,0.015501333607567681
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,65536,512,0.021341333786646526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,2048,3584,0.010091555615266165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2560,1024,0.00629688882165485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,768,512,0.005316444569163852
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,12288,16384,0.04757510953479343
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,4096,3072,0.015461333923869662
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,12288,6144,0.044241779380374484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,4096,256,0.00462311092350218
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,16384,2048,0.028935998678207397
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,64,12288,0.013148444394270578
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,10240,8192,0.034880889786614314
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,2560,4096,0.008043555749787224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,512,3072,0.02717688845263587
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,4096,2048,0.008024888734022776
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2560,5120,0.012785777449607849
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,128,32,0.005212444398138259
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,256,1024,0.006083555519580841
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,6144,12288,0.06329955657323201
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,16384,6144,0.15913421577877468
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,64,512,0.002489777695801523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,2048,3072,0.021351999706692163
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,64,3072,0.0112631113992797
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,256,2560,0.008007999923494127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,8192,16384,0.41858310169643825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,1024,256,0.014841778410805596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2048,512,0.0024951110697454875
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,12288,4096,0.01701155636045668
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1536,7168,0.01000711073478063
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,4096,2048,0.006258666515350342
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,5120,8192,0.023696889479955036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,16384,12288,0.11669600009918213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,65536,1536,0.02184533410602146
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,6144,3584,0.010453333457310995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,1024,7168,0.009721777505344814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,8192,5120,0.012614222036467658
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,8192,65536,0.2484968768225776
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,65536,12288,0.26029245058695477
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,16384,5120,0.2506133450402154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,5120,3072,0.007701333198282454
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1024,256,0.005598222215970357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,512,1024,0.013112000293201871
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2048,1536,0.00832711077398724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,1024,6144,0.009639110830095079
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,7168,1536,0.0064222220745351576
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,256,1024,0.005622222191757626
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,16384,2048,0.014863111906581454
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2560,3072,0.007256000406212277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3072,3072,0.006712889091836081
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,4096,256,0.005672889037264719
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2048,256,0.0025191110455327565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,32,32,0.0018995555324686898
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,6144,256,0.0035386664999855887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,3584,6144,0.02660177813635932
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,1024,2560,0.006711999989218182
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,7168,7168,0.013129777378506131
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,16384,512,0.006937777830494775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,10240,12288,0.0348639984925588
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1024,6144,0.010245333115259806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,1536,5120,0.01941866676012675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,1024,256,0.005251555393139522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2560,65536,0.06198133362664116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,16384,7168,0.045253333118226796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,10240,3584,0.026192888617515564
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,12288,3072,0.015975111060672335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,2560,1536,0.006663110935025745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,2560,3584,0.009536888864305284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,12288,5120,0.0967644453048706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2560,10240,0.015527110960748462
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,10240,3072,0.014889778362380134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,768,7168,0.010957332948843638
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,5120,6144,0.0592213339275784
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,128,12288,0.02472266720400916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,6144,16384,0.03429422113630507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,256,65536,0.04919644527965122
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,6144,16384,0.02569777766863505
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,2048,3072,0.034675555096732244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,1536,3584,0.007324444750944774
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,65536,3584,4.198696136474609
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,10240,4096,0.032279110617107816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,512,8192,0.020668443706300523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,2048,10240,0.034620443979899086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,6144,2048,0.006986666884687211
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,1536,16384,0.06732444630728827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,7168,1024,0.0723173353407118
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,12288,10240,0.03123555580774943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,12288,256,0.0210533340771993
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,65536,256,0.012888888518015543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,2048,512,0.005620444400442972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,6144,65536,0.1706391175587972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,768,256,0.007720000214046902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,64,1536,0.006481777876615524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,1536,2048,0.008200888832410177
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,2048,1536,0.01867555578549703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,3584,16384,0.07866488562689887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,12288,3072,0.03206400076548258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,256,7168,0.01072799993885888
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,2560,2560,0.018603555030292936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,10240,64,0.006263111200597551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,8192,16384,0.08151911364661323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1536,65536,0.12532267305586073
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,3584,1024,0.010830221904648675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3072,512,0.003915555361244413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,2048,12288,0.031007111072540283
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,128,768,0.002847111059559716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,65536,1536,0.030481778913074072
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,128,8192,0.01051999959680769
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,10240,3072,0.014179555906189812
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3584,64,0.005607111172543631
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,128,5120,0.015795555379655626
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,2048,65536,0.08703377511766221
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,3072,2048,0.03304444419013129
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,128,12288,0.022679110368092854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,3584,4096,0.0461368891927931
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,64,6144,0.01854933301607768
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,64,32,0.006632888896597757
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,256,5120,0.007299555672539606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3584,3584,0.012883555557992725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,4096,512,0.008630221916569604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1024,512,0.0069777775141927935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,512,10240,0.02130844526820713
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,12288,2560,0.019606222709019978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,256,8192,0.01127911110719045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1536,128,0.005191111316283544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,512,1024,0.006240889016124938
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,2560,8192,0.012903999951150684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,64,768,0.007373332977294922
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2560,3584,0.013400000002649097
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,10240,3584,0.01517777807182736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,12288,32,0.006296000132958095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,6144,64,0.005668444352017508
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,12288,3584,0.027994665834638808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1024,768,0.011476444701353708
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,7168,768,0.012748444245921241
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,64,3584,0.011420444481902652
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,4096,1024,0.007628444168302748
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,512,7168,0.03986400034692552
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,64,8192,0.011376889215575324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,10240,256,0.02513155506716834
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,768,768,0.015499555402331881
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1536,1536,0.006591111007663939
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,3072,2560,0.007315555380450354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,10240,8192,0.022666666242811415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,3072,2048,0.01554844445652432
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,128,3072,0.004588444613748127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,6144,128,0.0052435555391841466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3584,256,0.005569777968857024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,7168,3072,0.009332444104883406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,7168,4096,0.11310222413804795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,512,16384,0.015476443701320224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,256,6144,0.014085332552591959
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,768,8192,0.02057422200838725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,8192,256,0.002492444382773505
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,1536,16384,0.047019554509056934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,512,5120,0.009008000294367472
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,65536,8192,0.14437244998084173
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,65536,256,0.02481066683928172
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,2560,7168,0.022580444812774658
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,128,5120,0.004955555415815777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,65536,256,0.045732445187038846
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,65536,512,0.18461155891418457
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,6144,7168,0.04055022199948629
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1536,65536,0.03433155682351854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,5120,64,0.005241777747869492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,2048,2048,0.0069431112044387394
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,16384,768,0.022299556268586054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,2048,512,0.014488889111412896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,768,3584,0.00998577806684706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,3072,12288,0.05348089006212023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,256,256,0.013746666411558786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3072,768,0.006242666807439592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1536,256,0.005229333208666907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,5120,32,0.006693333387374878
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,3584,65536,0.06773600313398573
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,8192,16384,0.05854666895336575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,4096,1024,0.009688888986905416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,128,5120,0.015224888920783997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,128,16384,0.030928889910380047
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,10240,4096,0.017861333158281114
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,65536,8192,4.978118048773871
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,512,12288,0.02609955602222019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1536,10240,0.031015998787350122
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,7168,512,0.01072355525361167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,4096,16384,0.02771200074089898
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,10240,8192,0.03815911213556925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1536,1024,0.007065777977307637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,10240,1536,0.017716444200939603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,4096,2048,0.011828444070286222
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,8192,768,0.034309334225124784
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,65536,1024,0.03869688842031691
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,3584,3072,0.016925333274735343
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,8192,128,0.005937777873542573
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,1024,1024,0.006612444503439798
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,4096,2560,0.007394666473070781
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,128,1024,0.005948444621430502
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,8192,1536,0.020995555652512446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,5120,4096,0.055732442273033984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,256,6144,0.006214222146405115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,1024,10240,0.029498666524887085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,256,7168,0.01035377797153261
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,8192,768,0.009886222581068674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3072,3072,0.00793422261873881
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,3072,2048,0.006605333338181178
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,512,16384,0.017134222719404433
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,6144,16384,0.12669955359564886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,4096,256,0.025410667061805725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,768,1024,0.005959110955397288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,16384,256,0.0063591111037466265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,1024,2048,0.007010666860474481
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,512,16384,0.14256355497572157
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,12288,3072,0.012452444268597497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,7168,256,0.002830222249031067
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,8192,1024,0.014676445060306124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1536,2560,0.007642666498819987
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,512,12288,0.013382222089502545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,10240,768,0.018645332919226754
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,1536,10240,0.018209778600268893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3584,10240,0.25220266977945965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2048,2560,0.0052782222628593445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,768,1024,0.022125333547592163
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,6144,2560,0.020601777566803824
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,256,128,0.0021484444538752237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,2048,16384,0.023904000719388325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,5120,1536,0.018202667435010273
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,10240,512,0.019310222731696237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,128,12288,0.013417777915795645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,10240,6144,0.019948444432682462
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,5120,6144,0.23738845189412436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1536,7168,0.011408000356621213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,2560,16384,0.060264892048305936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1536,1024,0.0032213332338465583
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,7168,3072,0.02314844396379259
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1536,1024,0.003510222252872255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,256,512,0.0028311111446883944
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,256,256,0.012807110945383707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,64,6144,0.004597333156400257
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,10240,1536,0.03362044360902574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,16384,1024,0.031376888354619346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1536,16384,0.01311377767059538
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,8192,4096,0.04155200057559543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,16384,768,0.019225777851210702
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,512,1536,0.003567111161020067
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,10240,3584,0.013450666434235044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,768,3584,0.007624000310897827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,6144,3072,0.009001777403884465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,65536,2560,0.10388888915379842
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3584,2048,0.0052435555391841466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,512,16384,0.08468800120883518
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,4096,1024,0.006285333385070165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,2048,768,0.027663111686706543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,768,64,0.0021466666625605687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,7168,2048,0.008295111358165741
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,12288,7168,0.03496977686882019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,1536,10240,0.012855999999576144
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,256,8192,0.010718222293588849
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1536,10240,0.01109777804878023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,8192,512,0.003023111157947116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,1536,128,0.005245333330498801
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,768,7168,0.07236088646782769
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,3584,1536,0.040207111173205905
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2560,768,0.0029048888633648553
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2560,256,0.010876444478829702
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,128,2560,0.00738666703303655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,12288,1024,0.031799111101362444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,128,2560,0.007332444190979004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,1024,128,0.005960000058015187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,5120,8192,0.017504889104101393
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,16384,2560,0.014833778142929077
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,10240,16384,0.04184622234768338
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,6144,7168,0.013660444153679742
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,4096,10240,0.15639200475480822
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,64,1024,0.006299555715587404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,4096,1536,0.00960888879166709
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,8192,65536,0.1547244389851888
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,16384,512,0.00701333334048589
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,64,3072,0.009722666607962714
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,128,4096,0.008329778081840938
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,512,5120,0.005597333527273602
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,32,12288,0.013481777575280933
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,7168,256,0.0052764444715446895
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,10240,4096,0.019646222392718
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,3072,768,0.005590222362014983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,10240,256,0.003923555629120933
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,3584,12288,0.04198666744761997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1536,32,0.0069520001610120135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,3072,2560,0.00721777809990777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,10240,12288,0.03375466664632162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3072,2560,0.00630044440428416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,5120,1024,0.0046977777447965415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,512,32,0.002128888956374592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,4096,512,0.003530666646030214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,6144,1536,0.0053084443012873335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,4096,5120,0.010121777653694153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,128,7168,0.010058666268984476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,128,12288,0.10168177551693386
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2560,32,0.0052577778697013855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3584,4096,0.13530843787723118
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3584,32,0.0023093333260880574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,128,1536,0.0037271111375755737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,512,1536,0.015563555889659457
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,3584,12288,0.15222044785817465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,512,512,0.00258933338854048
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,2048,2560,0.016954667038387723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1536,1024,0.0032035555276605818
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,256,10240,0.011819555527634092
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,512,1536,0.0042168887125121225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,2560,1024,0.005959110955397288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,4096,3584,0.010395555860466426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2048,1536,0.004118222329351637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3584,128,0.002502222235004107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1024,5120,0.009015111459626092
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,1536,12288,0.01309866706530253
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1024,4096,0.0069475554757648045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2048,3072,0.01942311061753167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,32,2048,0.004558222161398994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,2560,10240,0.034887999296188354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,7168,12288,0.10049244430330063
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,768,1536,0.0038568890757030914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,16384,5120,0.05371822251213921
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,1024,3584,0.007627555893527136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,8192,3584,0.1335386700100369
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,32,3584,0.007625777688291337
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,32,2560,0.007259555160999298
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,1536,2560,0.018920888503392536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1536,2048,0.0045422220395671
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,8192,2560,0.010045333041085137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,128,128,0.0020871110674407748
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,7168,32,0.021789333886570398
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,12288,2560,0.011728000309732227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,5120,512,0.005585777676767773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,10240,3584,0.1776906649271647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,10240,5120,0.053341332409116954
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,512,3584,0.007314666277832455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,128,5120,0.00833422193924586
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,256,4096,0.012755555411179861
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,256,1024,0.0081404447555542
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,12288,8192,0.02307644486427307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,6144,2048,0.007280000381999546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,3584,65536,0.09411733018027411
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,16384,7168,0.14739110734727648
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3072,256,0.003527111063400904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,1536,3584,0.007967111137178209
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,128,1024,0.002875555513633622
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,3584,2560,0.016504888733228047
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,10240,768,0.005187555733654234
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2560,8192,0.053652445475260414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,5120,2560,0.007293333609898885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,8192,1536,0.02296000056796604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2560,1536,0.00812977800766627
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,1536,65536,0.10851644145117866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,128,2048,0.004639111045334074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,4096,768,0.0036151111125946045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,12288,128,0.020603555772039626
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,4096,5120,0.01239466667175293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,8192,16384,0.029447111818525527
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,2048,3584,0.0240328891409768
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,12288,768,0.011388444238238864
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,1536,1024,0.006593777901596493
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,4096,768,0.005909333212508096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,4096,10240,0.03747289048300849
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,10240,8192,0.023260444402694702
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,12288,5120,0.0192266669538286
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,2560,5120,0.01348444405529234
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,768,10240,0.011776888536082374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,3584,2048,0.046889778640535146
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,65536,4096,0.1198542250527276
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,768,512,0.011767110890812345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2048,2560,0.007246221933099959
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,3072,16384,0.048060443666246205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,16384,16384,0.2002168893814087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,128,512,0.005592000153329637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3072,768,0.027116444375779893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,512,3072,0.005616888817813661
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3072,32,0.005564444594913059
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,256,5120,0.012387555506494312
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,256,8192,0.013758222262064615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,4096,3584,0.028773334291246202
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,256,8192,0.008070222205585903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,3584,768,0.018575110369258456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,10240,2048,0.010410666465759277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,16384,256,0.008364444805516137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,65536,128,0.0053048887186580235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,2560,1024,0.005976000179847081
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,512,12288,0.027266666293144226
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,4096,1024,0.02388622197839949
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,7168,4096,0.01570311188697815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,64,2048,0.014476444986131458
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2560,32,0.0021911110315057966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,8192,128,0.00666933341158761
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,6144,128,0.005942222144868638
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,128,32,0.0019120000716712738
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2560,8192,0.22001066472795275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,10240,6144,0.024692444337738886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,12288,1024,0.007903111477692923
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,6144,2048,0.008916444248623317
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3072,16384,0.05974044402440389
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,5120,256,0.0028844444702068963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,65536,1024,0.029256890217463177
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,4096,10240,0.047629333204693265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,1536,512,0.01456622282663981
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,6144,65536,0.08161511023839314
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,4096,1536,0.012130666938092975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,2560,7168,0.027764444549878437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3072,7168,0.012747555143303342
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,768,8192,0.010693333215183683
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,65536,1024,0.01662577854262458
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,7168,32,0.0021342221233579847
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,65536,12288,0.2940097649892171
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2048,768,0.007271111011505127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,5120,8192,0.014816888504558139
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,256,3072,0.0075786668393347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3072,512,0.0027875554644399216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,8192,768,0.006263999889294307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,7168,8192,0.03187111020088196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,6144,1536,0.019667555888493855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,7168,768,0.0075911109646161394
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,65536,768,0.021830222672886316
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,7168,3584,0.008924444516499838
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,8192,16384,0.0640488862991333
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,6144,2560,0.008311111066076491
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,7168,12288,0.05839288896984524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,5120,3072,0.021818665994538203
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,256,65536,0.05311377843221029
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,5120,2048,0.007768889268239339
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,12288,6144,0.045336888896094434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,128,7168,0.005199111170238919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,32,1536,0.0059057776298787855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,256,512,0.002536000063021978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,6144,12288,0.8582035700480143
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3072,2048,0.009229333036475712
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,64,8192,0.05598400036493937
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,10240,2048,0.016879111528396606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,768,2048,0.04811288913091024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,4096,65536,0.5290444691975912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3584,32,0.005565333283609814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,128,5120,0.01538044379817115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,65536,3584,0.05590044127570259
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,4096,6144,0.038880000511805214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,8192,4096,0.01111288865407308
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,5120,7168,0.035755554835001625
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,1024,65536,0.06384088595708211
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,6144,12288,0.05483822358979119
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1536,2560,0.004955555415815777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,32,10240,0.004934222333961063
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,2048,768,0.016533333394262526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,7168,5120,0.03578222129080031
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,32,3072,0.007610666255156199
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,6144,65536,0.142067551612854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,8192,128,0.01483022173245748
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,12288,65536,0.2664462195502387
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3072,256,0.0028577778074476454
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3072,12288,0.018954666124449838
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,8192,3072,0.022184888521830242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3072,512,0.003481777591837777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,7168,2048,0.13490845097435847
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,64,7168,0.005193777796294955
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,4096,4096,0.009242666264375051
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,16384,6144,0.03158044483926561
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2048,8192,0.1443893379635281
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,16384,65536,1.0146844651963975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,128,512,0.006364444477690591
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,4096,512,0.005653333332803514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2560,5120,0.010074666804737514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,8192,10240,0.03385688861211141
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,512,6144,0.009125333693292405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,768,768,0.005919999960396025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,6144,512,0.005622222191757626
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,256,1536,0.0035422220826148987
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,4096,128,0.005294222384691238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,768,128,0.0021662221600612006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,128,12288,0.015308444698651632
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,65536,6144,0.38346134291754824
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,8192,2560,0.010079111489984725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,256,6144,0.00889511075284746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,32,65536,0.17461866802639434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,12288,3072,0.013288000391589271
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3072,12288,0.016557332542207506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1536,64,0.0028613333900769553
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,16384,2048,0.03481066558096144
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2048,7168,0.020617778102556866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,3072,2560,0.006993778049945831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,16384,2560,0.01538044379817115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,5120,65536,0.15717066658867732
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,2048,10240,0.02533955540921953
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,6144,1024,0.018820444742838543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,10240,16384,0.1581395599577162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3584,2048,0.008298666940795051
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,2048,256,0.015121777852376303
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,2560,16384,0.057555556297302246
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,768,768,0.005592000153329637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,4096,5120,0.011408888631396823
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,4096,10240,0.01642400026321411
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3072,8192,0.06551644537183973
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,12288,128,0.020796444680955675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,65536,5120,0.26020532184176975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,256,65536,0.06140977806515164
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1536,128,0.002230222233467632
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1536,4096,0.00853066643079122
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,5120,2560,0.013624000052611033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,5120,5120,0.014440000057220459
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,65536,16384,0.17030133141411674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,128,512,0.0058853332367208265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,6144,10240,0.01889600025282966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,8192,12288,0.02611200014750163
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,768,2048,0.00444444434510337
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,4096,512,0.025783111651738484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,2048,16384,0.0636355545785692
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,3584,2048,0.015846222639083862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,2560,5120,0.008357333640257517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,16384,3072,0.037744889656702675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,16384,8192,0.03144888745413886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2048,256,0.00554933316177792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,768,512,0.005558222118351195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3584,512,0.0069679998689227635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,128,16384,0.03603555427657233
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,128,65536,0.011552000211344825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,12288,1536,0.03618399964438544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,32,512,0.0024862223201327855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1024,4096,0.007647111184067196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1536,768,0.0031840000301599503
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2560,16384,0.031001776456832886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,2560,5120,0.008658666577604082
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,768,768,0.00350222239891688
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,7168,1024,0.019715555840068393
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,6144,512,0.005640000104904175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,6144,3072,0.01665866706106398
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,16384,65536,0.6263795428805882
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,2048,12288,0.29120977719624835
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,2048,4096,0.027580445011456806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,5120,3072,0.008663111262851292
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,10240,4096,0.025303999582926433
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,64,65536,0.0069448889957533936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,128,65536,0.009676444033781687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,128,2560,0.01309511148267322
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,4096,256,0.00590133335855272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1536,128,0.005571555760171678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,3584,256,0.015208888385030957
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,2560,512,0.014175110393100314
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,65536,1536,0.3308186795976427
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,1536,7168,0.014695111248228284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,1024,16384,0.016143111719025504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3072,128,0.0052693333062860705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,7168,8192,0.017816000514560275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,65536,3584,0.07915822002622816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,65536,6144,0.18636266390482584
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,6144,65536,0.2012417713801066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,4096,512,0.01514400045077006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,12288,768,0.008709333009190029
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,768,3584,0.006232000059551663
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,3072,4096,0.007669332954618666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,5120,1024,0.005737777799367905
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,16384,32,0.0021679999513758556
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,2560,3584,0.01824977828396691
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,12288,6144,0.03036266565322876
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,10240,256,0.003633777714437909
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,3584,256,0.0053093334039052325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,7168,8192,0.01867111192809211
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,64,16384,0.015627556376987033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,32,6144,0.009345778160625035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2048,2560,0.04920622375276354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,64,32,0.001834666563404931
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,7168,1536,0.006599111275540457
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,7168,3584,0.1608062187830607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2048,6144,0.0094035557574696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,6144,5120,0.10230577654308742
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,10240,12288,0.029612445169025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3584,8192,0.02008800043000115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,1024,2560,0.007267555428875818
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2560,12288,0.017633777525689866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,32,256,0.005657777604129579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2560,512,0.003253333270549774
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,16384,3072,0.01807733376820882
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,1536,768,0.016531555189026725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,512,12288,0.016140444411171805
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,768,1536,0.006254222244024277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,6144,8192,0.01945955554644267
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,16384,6144,0.048432889911863536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,32,5120,0.012901333471139273
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,8192,65536,0.14153244760301378
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,4096,256,0.016512889001104567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,12288,12288,0.0488755570517646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,6144,10240,0.041179554329978094
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3584,32,0.011815999945004782
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,16384,7168,0.06619022289911906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,32,3072,0.007909333540333642
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,5120,6144,0.1542960007985433
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,512,5120,0.019311999281247456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,10240,64,0.006600889066855113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1024,768,0.0032666667054096856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,3072,1536,0.006267555471923616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,16384,5120,0.02814133299721612
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,7168,512,0.0029884444342719186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,5120,3072,0.013083555632167392
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,6144,16384,0.022617777188618977
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,512,65536,0.6125857565138075
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2048,3584,0.008880889250172509
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,1024,1536,0.006263999889294307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3072,2048,0.005222222043408289
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,7168,128,0.01478844384352366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,768,256,0.02202133337656657
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,7168,3584,0.023360000716315374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,3584,6144,0.03604622350798713
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,128,7168,0.00999111102686988
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,4096,3584,0.009304000271691216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,65536,12288,0.41625155342949766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,2048,3584,0.03531288769510057
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,10240,512,0.012739555703269111
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,8192,32,0.011042666931947073
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,1024,4096,0.08560622400707668
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,16384,7168,0.02161600026819441
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,65536,7168,0.12678578164842394
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,6144,1024,0.040849778387281634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,8192,16384,0.03941866755485535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,16384,4096,0.03600355651643541
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,4096,7168,0.017675555414623685
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,6144,3072,0.00850399997499254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,32,4096,0.011429333024554782
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,8192,4096,0.04433066646258036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,1536,1024,0.006923555499977536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,256,16384,0.2447484334309896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,3584,6144,0.10672533512115479
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,256,6144,0.0052355556852287715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,1536,2048,0.01514933341079288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,2560,2560,0.018194667167133756
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,8192,3584,0.024438222249348957
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,6144,3584,0.05367288986841837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,256,8192,0.010813333094120026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,1536,16384,0.06773244672351413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,32,16384,0.015822221835454304
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,10240,256,0.005947555518812603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,7168,8192,0.09556000100241767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,10240,1024,0.007558222446176741
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,768,3072,0.010017777482668558
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,16384,12288,0.052222222089767456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3584,64,0.0063324446479479475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,1536,256,0.018958222534921434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,32,128,0.005670222143332164
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,2048,2048,0.02849422229660882
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,65536,3584,0.04228355487187704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,4096,2560,0.02236088944805993
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1024,512,0.0028204443968004654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,512,256,0.014838222000333997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,3584,512,0.007330666813585494
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,4096,512,0.005590222362014983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,12288,512,0.0062871111763848206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,256,1536,0.003921777837806278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,1024,7168,0.011393778026103973
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,8192,256,0.017685333887736004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,32,6144,0.009690666364298927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,1536,1536,0.006350222147173352
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,16384,16384,0.07074666685528226
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,8192,3072,0.017122666041056316
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,1536,512,0.005595555735958948
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,5120,512,0.015170666906568738
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,256,768,0.007977777885066139
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,65536,2048,0.0774293343226115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,64,512,0.005604444278611078
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,10240,4096,0.13704800605773926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1024,1024,0.0032622222271230486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2048,3584,0.011404444773991903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,7168,768,0.018727999594476488
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,1024,8192,0.02296000056796604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,12288,1024,0.007012444237867991
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,4096,128,0.02564977771706051
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,8192,2560,0.008011555506123437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,16384,2560,0.04177422324816386
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,128,4096,0.007840889195601145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,64,2048,0.006608888920810487
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3584,128,0.005280888742870755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3584,128,0.005963555640644497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,7168,2560,0.009272889130645329
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,12288,5120,0.35631823539733887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,768,512,0.002974222310715251
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,16384,1024,0.008418666819731394
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,768,7168,0.010079111489984725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,512,10240,0.022325333621766832
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,64,768,0.0027173333283927706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,512,12288,0.02476088868247138
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,10240,256,0.01789955629242791
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,64,3584,0.007664889097213745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,6144,12288,0.190628449122111
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,6144,768,0.018571555614471436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,12288,64,0.005921777751710679
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3072,1024,0.004920000003443824
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,7168,1024,0.0062977779242727495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,5120,256,0.017613333132531907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,8192,4096,0.010085333552625444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2048,8192,0.012460444536474017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,8192,256,0.01699733402993944
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,12288,3584,0.018045332696702745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,3584,2048,0.022685334086418152
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,256,5120,0.018654222289721172
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,8192,2048,0.009076444639099969
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,6144,10240,0.017323555217848886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3072,1536,0.004078222231732475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,512,8192,0.005620444400442972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1024,768,0.0031964443624019623
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,6144,12288,0.026947556270493403
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,6144,12288,0.01868355605337355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,32,10240,0.011767999993430244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1536,128,0.009372444616423713
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,32,7168,0.019374221563339233
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,1024,512,0.006937777830494775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1536,12288,0.18431733714209664
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,6144,8192,0.03049955434269375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1536,512,0.006170666466156642
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,7168,3584,0.014320888453059726
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,5120,16384,0.13579289118448892
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,6144,16384,0.05486489004558987
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,6144,4096,0.008692444198661381
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,65536,8192,0.07642310857772827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3072,65536,0.06256622076034546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,5120,4096,0.09103288915422227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,768,7168,0.00620888877246115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,6144,1024,0.030194666650560167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,32,1024,0.0029173334025674393
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,12288,4096,0.020208888583713107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2560,128,0.002879111096262932
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,2048,256,0.015118221441904703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,3584,6144,0.028924445311228435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,10240,768,0.004907555464241239
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,8192,256,0.002537777854336633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,10240,12288,0.05316977699597677
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,256,2560,0.009869332942697737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1536,32,0.005906666732496685
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,4096,7168,0.013074666261672974
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,4096,256,0.01735555628935496
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,6144,12288,0.019159111711714003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,6144,6144,0.012224888636006249
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,5120,3072,0.013559111290507846
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,768,7168,0.011359111302428775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,6144,128,0.0028355556229750314
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,10240,7168,0.03455111053254869
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,256,12288,0.04272800021701389
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,5120,128,0.005590222362014983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,1536,1536,0.022338666849666174
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,2560,6144,0.010106667048401302
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,512,7168,0.02199999988079071
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,256,1536,0.006600889066855113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,7168,32,0.011748444702890186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,6144,10240,0.03122044603029887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2048,6144,0.009480000370078618
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,5120,10240,0.01826844447188907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,65536,8192,0.1776346630520291
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2560,2048,0.015761777758598328
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,256,16384,0.021105777886178758
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,65536,4096,2.1450523800320096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,64,1536,0.011141333315107556
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,3072,6144,0.038141333394580416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,5120,12288,0.0458568897512224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,10240,16384,0.039170665873421565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3584,12288,0.018073777357737224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,64,32,0.0018177777528762817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,512,1536,0.003903999924659729
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1536,128,0.0022044444663657085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,6144,3584,0.022823111878501043
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,128,6144,0.022312889496485393
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,16384,1024,0.01871555546919505
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,512,16384,0.029537777105967205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,8192,128,0.0021866667601797315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3584,4096,0.009415999882751042
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,768,5120,0.00902844468752543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,65536,64,0.02685333291689555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,64,5120,0.004937777916590373
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,32,32,0.0019288888821999233
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,2048,128,0.014813333749771118
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,512,65536,0.10530489020877415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3584,6144,0.011420444481902652
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,7168,256,0.003313777761326896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,10240,32,0.005950222412745158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,512,4096,0.008005333443482717
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,2048,4096,0.016511110795868766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,10240,3584,0.05096266667048136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,8192,64,0.007648000286685095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,65536,12288,0.10984355873531765
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,768,256,0.0052675555149714155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,5120,64,0.003170666595300039
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,2560,4096,0.007601777712504069
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1024,7168,0.010336000058386061
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,32,1536,0.00794044468137953
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,256,32,0.002152888932161861
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,16384,256,0.006251555350091722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,768,8192,0.024712888730896845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,256,3072,0.005850666513045629
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,10240,5120,0.05008444521162245
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,10240,5120,0.042825778325398765
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1024,5120,0.008935111264387766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,6144,128,0.003165333428316646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,12288,256,0.017511111166742113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,16384,3072,0.03792266713248359
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,65536,16384,1.941979514227973
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,3072,7168,0.009988444546858469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,7168,256,0.0070159998204973005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,32,256,0.005647999958859549
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,16384,10240,0.029498666524887085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,5120,16384,0.022655111220147874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,32,16384,0.01590488851070404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,3584,2048,0.006618666566080517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,6144,64,0.018589332699775696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,128,128,0.0022337778160969415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,5120,1536,0.02552000019285414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2048,6144,0.01665599975321028
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1024,10240,0.015519110692871941
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2560,768,0.003216888962520493
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,7168,1024,0.04441155658827888
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,6144,6144,0.12048178248935276
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,512,2048,0.004571555389298333
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,65536,1024,0.02237244447072347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,128,768,0.002867555452717675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,10240,64,0.002236444503068924
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,1024,128,0.020971554848882887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,12288,768,0.006180444525347815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,3072,10240,0.0473911099963718
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,256,256,0.013099555340078143
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3584,16384,0.0823155575328403
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,32,65536,0.008357333640257517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,4096,768,0.005607111172543631
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,768,2048,0.020233778489960563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,12288,1536,0.009667555491129557
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,3584,768,0.007343110938866933
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,768,12288,0.02973244587580363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,7168,2048,0.007905777957704332
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,512,8192,0.01314933349688848
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,12288,65536,0.15169155597686768
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1536,3072,0.005618666609128316
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,768,1024,0.013842666314707862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,3584,10240,0.014708444476127625
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,65536,4096,0.6735155317518445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,4096,8192,0.03253422180811564
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,256,16384,0.015279110934999255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3072,7168,0.01761155492729611
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,3584,4096,0.02184088859293196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,10240,2048,0.028479999966091577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,4096,12288,0.019305777218606737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,5120,12288,0.0823777781592475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,512,128,0.013974222871992322
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,8192,8192,0.05936621957355075
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,5120,5120,0.13421689139472112
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,8192,6144,0.019401777121755812
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,10240,512,0.006974221931563483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,5120,2560,0.008258666429254744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,4096,3584,0.026424888107511733
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,64,12288,0.005431111074156231
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,512,2560,0.014504889647165934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,16384,128,0.0028488888508743713
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,256,6144,0.017287110288937885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,65536,1024,0.02749155627356635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,2560,3584,0.03887822230656942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,2048,12288,0.05272444420390659
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,512,2048,0.030376000536812678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,4096,65536,0.06765866941875882
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,10240,16384,0.06890133354398939
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,5120,3072,0.011743110915025076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,16384,65536,0.19474489159054229
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,32,256,0.0021786666992637846
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1024,768,0.0028640000770489373
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3072,64,0.005572444448868434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,64,7168,0.013162666724787818
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,256,12288,0.014469333820872836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,256,16384,0.01463733282354143
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2048,32,0.005585777676767773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2560,64,0.002114666625857353
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,5120,3584,0.010078222387366826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,7168,256,0.01068800025516086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,6144,12288,0.027330666780471802
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,10240,8192,0.02566755480236477
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1536,6144,0.009335111412737105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,1024,3072,0.016255999604860943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,10240,512,0.011326221956147088
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,1536,3072,0.007263110743628607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,65536,5120,0.11137955718570286
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,768,512,0.0052693333062860705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,8192,3584,0.009786666267448002
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,4096,1024,0.02036622166633606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,512,256,0.016843555702103507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,3072,2560,0.023382221659024555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,4096,2560,0.0069617778062820435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,768,3072,0.007975999679830339
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1536,8192,0.01093155559566286
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,16384,4096,0.04205422268973457
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,7168,7168,0.03048622277047899
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,65536,2560,0.06634577777650622
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,3584,3584,0.06877244181103177
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,256,32,0.0018684443914228016
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,10240,65536,0.15272888872358534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,768,3584,0.00811288919713762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,6144,10240,0.01851111153761546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,768,1024,0.005906666732496685
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,768,32,0.002151111140847206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1536,2048,0.0070017774899800616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,3584,2048,0.0155395550860299
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,512,512,0.005412444472312927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2560,32,0.0052435555391841466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,3072,1024,0.015604444675975375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,5120,8192,0.015083556373914083
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,128,1024,0.003593777616818746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,2048,1024,0.006089777582221561
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,64,1536,0.0035777779089079965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,7168,6144,0.04849955439567566
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,6144,2560,0.009939555492666032
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1536,2048,0.005943999936183293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,3072,6144,0.036673777633243136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,12288,768,0.00904799997806549
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,512,7168,0.018274666534529794
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,32,4096,0.012991111311647626
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,5120,6144,0.01148000028398302
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,3584,7168,0.023714666565259297
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,5120,1024,0.01756177842617035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2560,4096,0.008448888858159384
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,7168,4096,0.013175110850069257
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3072,512,0.007589333587222629
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,4096,3072,0.012081777883900536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,12288,7168,0.1303244431813558
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,512,6144,0.005606222069925732
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,768,16384,0.016187555260128446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,2560,5120,0.028568890359666612
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,65536,1024,0.06346044275495741
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,64,8192,0.011122666299343109
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,8192,3072,0.017646221650971305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,8192,128,0.00903911143541336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,128,16384,0.052609778112835355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,6144,3584,0.010570666856235929
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,12288,8192,0.051670223474502563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,65536,256,0.034986668162875705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,128,2048,0.008343111309740279
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,7168,128,0.005331555588377847
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,32,64,0.0019173332386546666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,4096,32,0.00564533347884814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,256,3072,0.00739644467830658
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,2048,6144,0.01389600005414751
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,1536,256,0.005274666680230035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,2048,2560,0.015917332635985482
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,6144,12288,0.09639200237062241
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3584,16384,0.023336889015303716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,128,2560,0.007326222128338284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3584,7168,0.011824888487656912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,6144,256,0.007284444239404466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,8192,10240,0.01685244507259793
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,256,8192,0.00941422250535753
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,12288,1024,0.006592888798978593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,12288,7168,0.046210666497548424
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,7168,7168,0.040009776751200356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,128,16384,0.016157334049542744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,256,64,0.0019120000716712738
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,1536,256,0.0052533331844541765
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2560,65536,0.10414755344390869
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,65536,2560,0.2501804563734266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,65536,8192,5.456117418077256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,768,768,0.028139554791980322
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,128,16384,0.01626044511795044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,32,10240,0.020977778567208186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,16384,8192,0.02702222267786662
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,6144,64,0.008685333033402761
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,5120,10240,0.0209840006298489
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,2560,3584,0.02345422241422865
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,5120,128,0.00797155582242542
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1024,10240,0.009745777481132084
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,10240,65536,0.14188178380330405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,10240,768,0.0069653333889113525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,10240,2560,0.011426666544543372
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,5120,5120,0.010924444430404238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,1024,12288,0.025120000044504803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,3584,2560,0.008048888709810045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,6144,4096,0.00906844437122345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,8192,6144,0.017651556266678702
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2560,16384,0.020801777640978496
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,12288,7168,0.05330933464898003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,4096,3072,0.009750222166379293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,6144,3584,0.019298666053348117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,5120,32,0.0021377777059872947
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,1536,768,0.005630222045713001
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2560,5120,0.13562133577134874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2560,1536,0.006622222148709827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,256,4096,0.0063528890411059065
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,8192,64,0.0022542222092549005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,256,8192,0.12840978304545084
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,64,64,0.0019422223170598347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,2560,8192,0.017288888494173687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,8192,3584,0.011776888536082374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,16384,12288,0.03312355611059401
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,1024,4096,0.007997333175606197
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,8192,3072,0.010845333337783813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,6144,4096,0.010421333213647207
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,256,2560,0.011745777395036487
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1024,32,0.005959110955397288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1536,768,0.002895111011134254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2048,6144,0.01701244380739
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,3584,1024,0.005943999936183293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,8192,64,0.006970666348934174
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,512,3584,0.02424533334043291
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,128,3584,0.004641777939266629
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3072,3584,0.011667555405033959
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,65536,1536,0.03149688906139798
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,8192,768,0.008718222379684448
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,5120,32,0.005283555636803309
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,6144,16384,0.09503555297851562
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,12288,16384,0.04388533367051018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,12288,1024,0.006203555398517185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,10240,3584,0.015307555596033732
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,10240,2560,0.011824000212881299
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,768,7168,0.13145777914259169
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,512,2048,0.006963555183675554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,5120,12288,0.41031466590033633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,3072,3584,0.007911110917727152
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,4096,65536,0.18552445040808785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,65536,256,0.012465777496496836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,8192,65536,0.09152088562647502
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,128,2048,0.00646577775478363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,768,65536,0.05968088573879666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,512,128,0.006647999915811751
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,7168,12288,0.07744622230529785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,5120,2048,0.018252443936136033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,3072,4096,0.008055110772450766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,65536,7168,0.06502844227684869
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,12288,65536,0.1499155494901869
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,7168,3072,0.007994666695594788
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2560,7168,0.010398222340477837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,3072,16384,0.020463999774720933
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,1024,768,0.006646222124497096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,2048,256,0.008791111409664154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2560,12288,0.027172444595230952
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,512,3072,0.005792888916201062
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,8192,7168,0.05828977955712212
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,5120,4096,0.042378667328092784
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,16384,8192,0.06605333089828491
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1536,7168,0.009377777576446533
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,5120,6144,0.03910133242607117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,12288,32,0.003906666818592283
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,8192,7168,0.09458221991856892
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2560,1024,0.004295999805132548
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,4096,3072,0.11631555027431911
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,2048,512,0.0053048887186580235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3072,1024,0.011249778171380361
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,16384,4096,0.0221413340833452
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,128,3584,0.0047600000268883174
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,256,4096,0.004622222234805425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,256,6144,0.00942666663063897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,128,8192,0.02751911183198293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,65536,7168,4.238811492919922
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,128,64,0.010019555687904358
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,5120,10240,0.24043112330966523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,12288,10240,0.03645955522855123
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,12288,16384,0.10014133320914374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2048,5120,0.009015111459626092
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,64,768,0.005960888746711943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,5120,16384,0.03342755635579427
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,1536,128,0.015633778439627755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,65536,10240,0.1643128924899631
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,4096,512,0.005731555736727185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,3584,2560,0.007000000112586551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,2560,1536,0.01792266633775499
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,4096,6144,0.012551999754375882
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,6144,8192,0.03221777743763394
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,64,3072,0.004948444250557157
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,32,3072,0.005592888842026393
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,5120,4096,0.16448177231682673
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,5120,65536,0.2028435601128472
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,65536,8192,0.6695431073506674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,3072,2048,0.006608888920810487
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,65536,8192,0.07224977678722806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,128,12288,0.010403555300500657
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,256,1536,0.006256888724035687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2048,256,0.006583999842405319
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2048,32,0.0021893332401911416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,768,10240,0.023022222850057814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,12288,3072,0.049376888407601245
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,8192,64,0.015994666351212394
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,16384,65536,0.22242045402526855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,5120,65536,0.18884889284769693
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,512,1536,0.006334222025341458
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,65536,2560,0.0946577787399292
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,2560,10240,0.2692506578233507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,16384,5120,0.017464000317785475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,12288,128,0.006304889089531369
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,65536,6144,0.05317777726385328
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,5120,3584,0.007628444168302748
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3584,128,0.0024862223201327855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,64,32,0.005578666511509154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,12288,2560,0.28324445088704425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3072,1024,0.006264888991912206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,12288,65536,0.1704568862915039
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,32,768,0.0028239999794297745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,512,4096,0.007649777664078607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1024,7168,0.007992888490358988
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,256,10240,0.03340977761480544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,4096,10240,0.018240888913472492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,128,3072,0.007649777664078607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,6144,3072,0.008858666651778752
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,10240,6144,0.031054221921496924
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1536,32,0.0035004446076022256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,12288,2560,0.012715555727481842
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,4096,32,0.0021608887861172357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,768,128,0.013829333086808523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,5120,512,0.02976622184117635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,3584,16384,0.03699200020896064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,768,7168,0.030633777379989624
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,4096,16384,0.023766222927305434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,10240,12288,0.044659554958343506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,1536,10240,0.02685333291689555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2560,1024,0.006367110957702001
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,7168,2560,0.009272000028027428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,8192,768,0.007651555869314406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,128,512,0.0069102222720781965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,512,4096,0.017479999197853934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,65536,8192,0.0937617752287123
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,768,128,0.014155555102560254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,5120,64,0.0063768890168931745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2560,3584,0.00832088871134652
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,7168,2048,0.0070257774657673305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,5120,4096,0.010476444330480365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,10240,1024,0.020953777763578627
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3584,5120,0.08556889163123237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,10240,128,0.015458666616015963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,10240,2048,0.10092800193362766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,10240,65536,0.14225510756174722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2048,4096,0.007947555846638149
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,512,2560,0.0069546666410234236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,32,2048,0.008665777742862701
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,256,512,0.011726222104496427
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,768,16384,0.02862666712866889
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,2048,8192,0.025395555628670588
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,12288,2048,0.009342222577995723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2048,8192,0.021207110749350652
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,10240,8192,0.04010755485958523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,5120,16384,0.023423110445340473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,768,2560,0.006908444480763541
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,7168,512,0.008691555923885768
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,32,5120,0.005964444329341252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,1536,2048,0.006308444258239534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,768,768,0.015498666299713982
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,6144,5120,0.2243786652882894
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,65536,1536,0.0721564425362481
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,3584,12288,0.04847466614511278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,256,3584,0.01891644464598762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,7168,4096,0.009367999931176504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2048,3072,0.00564533347884814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,7168,1024,0.005931555396980709
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,3072,1024,0.006576889091067844
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,7168,4096,0.01240533341964086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,12288,3584,0.014663111832406787
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,7168,4096,0.026759111218982275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,6144,2048,0.006917333437336816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,12288,2560,0.10097066561381023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,768,5120,0.007621333003044128
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,256,768,0.003175111073586676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,512,2560,0.006976889239417181
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,512,1536,0.006166222194830577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,4096,16384,0.035614222288131714
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3584,1536,0.006281777802440856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,4096,3072,0.00797155582242542
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3072,128,0.002442666639884313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,12288,16384,0.06322844160927667
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,12288,12288,0.12938488854302302
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,1536,2048,0.006627555522653792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,65536,7168,0.16543289025624594
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,32,16384,0.015480000111791821
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,10240,7168,0.02642222245534261
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,2048,2560,0.04450044367048475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,2048,7168,0.010096000300513374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,5120,5120,0.012422222230169507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2048,3072,0.007601777712504069
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,7168,8192,0.03615644574165344
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,10240,3584,0.013987556099891663
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,2560,16384,0.02053777707947625
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,256,1536,0.012814222110642327
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1536,256,0.005960888746711943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,16384,128,0.006007110906971826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,7168,1536,0.01331555595000585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,128,5120,0.02015555567211575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,3584,3584,0.02132533325089349
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,16384,10240,0.04309066798951891
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,512,65536,0.08131733205583361
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,16384,2048,0.11152888668908013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,7168,6144,0.012358222570684222
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,768,2560,0.01594044433699714
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,12288,768,0.036271111832724676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3584,2048,0.00665155549844106
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,65536,2560,1.2536453670925565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,65536,4096,0.054471111959881253
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,256,12288,0.022060443957646687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,12288,2560,0.013411555853154926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,6144,768,0.05490755703714159
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,16384,65536,0.2113164398405287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,8192,768,0.01896711190541585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1536,32,0.0022177776942650476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,4096,3072,0.007693332930405934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,7168,1536,0.00924177798959944
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,7168,4096,0.012544000314341651
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,128,4096,0.01943822205066681
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,256,2048,0.013412444127930535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,512,7168,0.005613333235184352
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,8192,3584,0.0099982221921285
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2560,10240,0.025625778569115534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,128,768,0.002842666581273079
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,768,3072,0.005543111099137201
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,10240,256,0.007401777638329401
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1536,16384,0.03696799940533108
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,3584,2048,0.020661332541041903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,6144,256,0.005889777921968036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,6144,512,0.0063564446237352155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,10240,128,0.010465777582592435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,256,4096,0.007999999655617608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,6144,64,0.005554666535721884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,6144,5120,0.026305778159035578
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1024,768,0.0037466668420367767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,4096,64,0.006179555422729916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,512,64,0.0041946665280395085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,65536,1024,0.015925332903862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,7168,65536,0.08461955520841812
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,2560,1536,0.01553600033124288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,8192,512,0.019498666127522785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1536,2560,0.007301333049933116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2560,2048,0.006595555692911148
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,3072,8192,0.01219377749496036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,5120,5120,0.02387733260790507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,16384,768,0.006777777853939269
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,65536,4096,0.05065600077311198
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,2048,256,0.015107555521859063
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,7168,1536,0.006610666712125142
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,128,5120,0.006605333338181178
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,7168,768,0.020940444535679285
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,5120,8192,0.04611022273699442
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,512,2048,0.004583111239804162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,1536,12288,0.04860533277193705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1024,5120,0.008767111433876885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,2560,256,0.016151999433835346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,10240,8192,0.21235199769337973
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,768,1536,0.007274666594134436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,7168,128,0.0042088888585567474
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,65536,12288,0.17653066582149932
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,1024,1024,0.013456888496875763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,3584,3584,0.007308444215191736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,6144,8192,0.03088444471359253
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,2048,12288,0.015846222639083862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,768,8192,0.03613422314325968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,128,7168,0.010103999740547603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,6144,5120,0.022858666049109563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,12288,128,0.00628977765639623
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,1024,65536,0.3995262251959906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,5120,4096,0.008013333711359236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,768,2560,0.0053182223604785065
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,32,4096,0.004425777743260066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,256,6144,0.005250666704442766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,512,2048,0.007066667079925537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,16384,2048,0.032650666104422674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2048,64,0.0022728888110982045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,5120,64,0.007657777931955125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,10240,5120,0.01890222231547038
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,2560,6144,0.03236888845761617
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,4096,768,0.005603555589914322
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,12288,65536,0.23373066054450142
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,65536,5120,0.12975200017293295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,16384,64,0.0069724445541699724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,512,4096,0.007617777420414819
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3072,6144,0.012716444830099741
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2048,64,0.002157333410448498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,128,2560,0.016534222496880423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,512,2048,0.0069751110341813825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,128,512,0.005952888892756567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,7168,4096,0.012728888955381183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,6144,1024,0.010009778042634329
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,16384,512,0.03407911128467984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,1024,7168,0.02331377731429206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,12288,10240,0.036280890305837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,1024,4096,0.016193777322769165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,7168,128,0.005576000031497743
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,512,10240,0.021681777305073206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,1024,4096,0.01684177749686771
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,65536,768,0.0314302212662167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,7168,7168,0.01591999994383918
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,2560,256,0.015371556083361307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,128,7168,0.009376889301670922
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,5120,10240,0.022090666823916968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,10240,7168,0.022351110974947613
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,5120,1024,0.006099555641412735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,2048,512,0.013762666947311826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,768,1024,0.015815110670195687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,16384,3072,0.28935111893547905
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2048,768,0.011517333487669626
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,2048,3584,0.009356444080670675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,16384,32,0.0032017777363459268
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,12288,3584,0.03427822060055203
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,256,5120,0.008707555631796518
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,4096,10240,0.029097778929604426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,256,16384,0.023727999793158636
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,5120,4096,0.027817779117160376
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2560,12288,0.2362142139010959
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,7168,12288,0.02032711108525594
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,64,1024,0.006314666734801398
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,768,256,0.0021608887861172357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,1024,65536,0.08066222402784559
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2048,768,0.00591822216908137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,128,1536,0.006263999889294307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,128,6144,0.009657777845859528
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,10240,7168,0.026318222284317017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,2560,3072,0.007269333634111617
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,1024,5120,0.008655110994974772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,1536,12288,0.013442666994200813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,2048,512,0.005700444595681296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,12288,16384,0.061623109711541064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,5120,3584,0.05057777629958259
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,10240,768,0.01871288816134135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,4096,4096,0.021009777983029682
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,16384,10240,0.055586669180128306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,10240,10240,0.04031199879116482
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2560,12288,0.02325244413481818
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,32,4096,0.004545777622196409
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,8192,256,0.01073155552148819
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,768,64,0.005619555711746216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1536,6144,0.014688889185587565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2048,64,0.0021048889805873237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,8192,2560,0.022896000080638464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,512,65536,0.06237155861324734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,2560,16384,0.0355022218492296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2560,4096,0.008354666332403818
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,4096,4096,0.007988444632954067
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,12288,6144,0.4421493212381999
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,7168,256,0.005294222384691238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2560,12288,0.29263554679022896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1024,10240,0.010438222024175854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1024,10240,0.016902221573723685
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,32,32,0.0052639999323421055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,8192,128,0.021556443638271753
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,7168,12288,0.09180711375342475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,64,768,0.00562755556570159
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,65536,2560,0.2626844512091743
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,1536,256,0.005607111172543631
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,128,1024,0.005993777679072485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,2048,1536,0.014488000008794995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,2048,6144,0.008995555341243744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,32,512,0.0063786668082078295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,8192,32,0.00253866674999396
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,6144,8192,0.03553244471549988
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,65536,6144,0.2837288909488254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,4096,256,0.015982222225930955
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,16384,2048,0.03253422180811564
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,5120,768,0.006002666635645761
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,768,12288,0.03277511066860623
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,65536,2048,0.05301422211858961
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1024,3584,0.0075253330998950535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,128,6144,0.023160000642140705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,1024,2560,0.006631111105283101
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,512,3584,0.007600000335110559
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,8192,2560,0.025455110602908667
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2560,768,0.005623111294375525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,128,1024,0.013065777719020844
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,8192,128,0.025560888979170058
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,512,12288,0.007618666523032718
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,2048,768,0.036672890186309814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,4096,7168,0.011737777955002256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,1536,1024,0.005948444621430502
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,768,4096,0.007680888805124495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,64,4096,0.008578666382365758
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,6144,1536,0.006962666908899943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,4096,256,0.005572444448868434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,65536,128,0.05172355638609993
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,2048,4096,0.020055999358495075
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,4096,768,0.0028942223224374983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,10240,1024,0.024433778391944036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,65536,1536,0.02383733292420705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,1536,2048,0.01464355488618215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,8192,12288,0.30020178688897026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,10240,10240,0.03500266538725959
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3584,1024,0.006260444306664997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,1024,256,0.023375999596383836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1024,512,0.00665244460105896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,8192,6144,0.017755554782019723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,6144,16384,0.037872887319988675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3584,7168,0.020642666353119742
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1024,256,0.005738666488064661
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,512,512,0.012220444778601328
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,5120,2048,0.020607999629444547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,6144,7168,0.017829333742459614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,12288,12288,0.051983998881445996
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3584,65536,0.0732097758187188
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,10240,2048,0.027698665857315063
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,768,768,0.005935110979610019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,10240,6144,0.03537777728504605
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,768,256,0.005602666487296422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,2048,768,0.005708444449636672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,65536,3584,0.09133866760465835
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,512,3584,0.008034666379292807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,8192,10240,0.04299022091759575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,6144,5120,0.016958221793174744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,4096,7168,0.03075466553370158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,65536,32,0.17193867100609672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,32,256,0.002181333386235767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,4096,10240,0.28725067774454754
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,3072,128,0.00527022240890397
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,16384,3584,0.014671110444598727
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2048,512,0.005604444278611078
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,512,4096,0.016959110895792644
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,768,256,0.003589333345492681
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,7168,128,0.015463110473420886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,768,5120,0.005593777944644292
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3072,3584,0.056836446126302086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,6144,7168,0.01759822169939677
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,1024,1536,0.006234666539563074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,4096,3584,0.022319111559126113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,12288,3072,0.013617777989970313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,2048,5120,0.051797333690855235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,32,64,0.0022400000856982335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2560,512,0.006000888844331105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1024,12288,0.012669333153300814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,65536,2560,0.38192444377475315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,256,128,0.010002666877375709
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,512,768,0.018221333622932434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,4096,5120,0.04319377740224203
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1024,10240,0.10874666770299275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,16384,7168,0.056440002388424344
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,1024,12288,0.03665688965055678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,8192,32,0.024373332659403484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1536,128,0.002165333264403873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,65536,3072,0.06770488950941297
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,32,768,0.002883555574549569
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,16384,7168,0.03793066740036011
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,2560,768,0.005620444400442972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,512,3584,0.015505777464972602
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,768,3584,0.007692444655630324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,256,1536,0.0036684444381131064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,64,3072,0.004856889032655292
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1536,8192,0.011120888921949597
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3584,10240,0.015485333071814643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,768,3072,0.01538044379817115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,12288,12288,0.03602666656176249
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,256,5120,0.008373333348168267
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3072,12288,0.01828799976242913
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,12288,6144,0.027985778119828966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,65536,10240,0.19690577189127603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,6144,2560,0.0069422221018208405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,10240,256,0.005629333357016246
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,768,1024,0.00563822231358952
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,12288,1536,0.009509333305888707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,2560,512,0.005574222240183089
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1024,1536,0.006355555521117316
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,2048,3584,0.025063999825053748
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,256,3072,0.007687999970383114
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,8192,768,0.03616888986693488
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,65536,1024,0.054251554939481944
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,4096,10240,0.0414373344845242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,12288,16384,0.06186578008863661
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1024,64,0.005618666609128316
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,1536,65536,0.34970932536655003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1536,1024,0.006314666734801398
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,12288,5120,0.03919911053445604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,64,512,0.0028640000770489373
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,64,12288,0.05977955791685316
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,64,4096,0.004594666676388847
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,128,16384,0.016201777590645682
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,5120,3072,0.02108000053299798
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,10240,16384,0.04383377896414863
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,1536,1536,0.01735111077626546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,768,4096,0.08080266581641303
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,12288,65536,1.718099594116211
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,7168,6144,0.02039377722475264
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,1024,7168,0.00943288869327969
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,4096,1024,0.006315555423498154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,64,256,0.002242666565709644
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,6144,512,0.03638222151332431
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,12288,64,0.006695999867386288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1536,12288,0.030181331766976252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,1536,2560,0.019702222612169053
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,768,2560,0.004920888692140579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,5120,12288,0.0609440008799235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,512,12288,0.013187555803192986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2560,6144,0.01518133282661438
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,512,2560,0.004915555732117759
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,128,3584,0.007662222617202335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,12288,32,0.00700711127784517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1536,512,0.006027555714050929
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,16384,1536,0.036885334385765925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,768,5120,0.013811555173661975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,5120,5120,0.016304888659053378
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,768,3584,0.01795644395881229
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,64,1024,0.005951111101441913
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,2048,3072,0.021398221453030903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,1024,3072,0.016873778568373788
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,16384,512,0.0192586663696501
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,12288,2048,0.011843555503421359
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,4096,32,0.0026337777574857077
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,4096,64,0.002234666711754269
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,512,64,0.002228444442152977
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,6144,1536,0.018944889307022095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,16384,1024,0.02637422250376807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,64,128,0.0022213332768943575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,1536,4096,0.02581688927279578
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,6144,7168,0.01532444523440467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,768,1024,0.017854221993022494
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,8192,128,0.005628444254398346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,7168,3072,0.010786666638321347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,768,16384,0.056051552295684814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,512,10240,0.011768889096048145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1024,128,0.005959110955397288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,6144,5120,0.036291556225882635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,128,4096,0.017691555950376723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,256,6144,0.017668444249365065
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,512,3584,0.00612088872326745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,10240,7168,0.03883644607332017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,4096,12288,0.39335023032294375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,5120,2560,0.007431999676757389
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,8192,65536,0.16986221737331816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,8192,512,0.005939555664857228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,768,6144,0.021324444148275588
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,3072,6144,0.022710222336981032
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,10240,1024,0.007621333003044128
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,12288,128,0.015779554843902588
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,128,65536,0.06944888830184937
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,256,4096,0.01960711181163788
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,65536,10240,1.0883252885606554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,4096,1024,0.030928889910380047
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,1024,6144,0.0284826656182607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,6144,1024,0.00583022211988767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,12288,512,0.019555555449591745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,7168,3072,0.012862222062216865
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,3072,768,0.04107111030154758
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2560,5120,0.012562666502263812
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,7168,3584,0.029832889636357624
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,128,7168,0.005541333307822545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,12288,1536,0.008351999852392409
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,1024,10240,0.06835377878612943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,65536,8192,0.09400800201627944
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,5120,8192,0.029995554023318823
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,128,4096,0.007905777957704332
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,12288,1536,0.008533333738644918
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,4096,6144,0.010144000252087912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,5120,1024,0.057984888553619385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,128,65536,0.0640044477250841
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,8192,12288,0.09873866372638279
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,8192,1024,0.00794577764140235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,7168,768,0.007728888756699032
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,256,6144,0.027273777458402846
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,64,4096,0.01715377800994449
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,768,2048,0.0148062225845125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,2048,3072,0.010775999890433418
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,1536,10240,0.021034666233592566
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,6144,16384,0.6649049123128256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,768,4096,0.01797155539194743
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,8192,6144,0.029750221305423315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,1536,128,0.01482311056719886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,8192,128,0.002432888787653711
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3072,256,0.01364177796575758
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,512,3072,0.007262222468852997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,4096,65536,0.11153066158294678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,6144,4096,0.01108088923825158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,4096,8192,0.026675555441114638
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,256,6144,0.00629688882165485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,10240,5120,0.018179555733998615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,3584,3584,0.017278222574128043
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,12288,256,0.018205333087179396
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1536,8192,0.010015111002657149
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,64,2048,0.008420444197124904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2048,12288,0.01863911085658603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,12288,256,0.0035626664757728577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,10240,8192,0.027590221828884546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1536,4096,0.012255111502276527
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,4096,2560,0.0069351109365622205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,12288,2560,0.013278221918476952
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,5120,8192,0.014460444450378418
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,768,12288,0.023800889650980633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,8192,512,0.010028444230556488
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,5120,768,0.005919999960396025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,64,6144,0.00941066692272822
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,128,7168,0.01718755563100179
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,6144,128,0.0031351111829280853
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,768,256,0.015808888607554965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,768,768,0.004882666799757215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,7168,2560,0.03415200114250183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3072,2560,0.0221413340833452
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3584,768,0.004238222208287981
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,6144,16384,0.03395466672049628
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,4096,128,0.01478844384352366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,12288,65536,0.5804070896572536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,5120,256,0.006600889066855113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,4096,2048,0.019204444355434842
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,12288,32,0.002151111140847206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2560,2048,0.006706666615274217
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,16384,3584,0.04224444429079691
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,3072,256,0.015166223049163818
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2048,7168,0.12577066156599256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2048,2560,0.005882666756709416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,16384,8192,0.030525333351559106
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,65536,5120,0.058427552382151283
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,64,64,0.0018595556418100994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,12288,65536,0.13684622446695963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1536,512,0.007289778027269576
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,8192,12288,0.03565066721704271
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,12288,512,0.02593333356910282
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,128,16384,0.241338676876492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3072,3584,0.0074337778819931885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,512,16384,0.018214222457673814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,65536,10240,0.14554222424825033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1536,256,0.005616000129116907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,768,768,0.01552977826860216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,16384,65536,0.25612711906433105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1536,1536,0.008367111285527548
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3072,512,0.00535111129283905
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,16384,8192,0.9195671081542969
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,5120,768,0.003195555466744635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,8192,12288,0.43863733609517414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,10240,8192,0.031735112269719444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,512,65536,0.09188177850511338
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,64,12288,0.0196435550848643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,64,3072,0.017855111095640395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3584,4096,0.010967999696731567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,768,16384,0.016768889294730294
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,2560,10240,0.01294133315483729
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,4096,3584,0.008062221937709385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1536,16384,0.06806577576531304
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,128,4096,0.01757866640885671
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,1536,2560,0.021615111165576514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,512,4096,0.007592889169851939
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,32,7168,0.005239999956554837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,5120,64,0.005959110955397288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,64,1024,0.0028853333658642242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,4096,6144,0.01277599980433782
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2560,1536,0.006907555378145642
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,16384,16384,0.04880977670351664
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1536,7168,0.011110222174061669
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1024,1024,0.014206222361988492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,12288,6144,0.03485688898298476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,12288,1536,0.04998044504059685
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,6144,64,0.002206222257680363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,128,65536,0.07014755407969157
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3072,16384,0.12418222427368164
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,5120,65536,0.42669423421223956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,6144,512,0.005912000106440649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,65536,256,0.027366222606764898
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,256,7168,0.009864889085292816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3584,7168,0.013824000126785703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,7168,1536,0.006710222197903528
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,768,256,0.0063457778758472866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,768,768,0.007292444507280986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,6144,1024,0.01771377854877048
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,32,2560,0.005444444302055571
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,10240,1536,0.009881777895821465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,16384,4096,0.037793778710895114
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,65536,1024,0.06556000312169392
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,1536,128,0.015862221519152325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,65536,12288,0.5553270975748698
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,8192,3072,0.0090524446633127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,1024,8192,0.014217777384652032
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1024,2560,0.007321777443091075
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,5120,3072,0.00924355536699295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,32,4096,0.004610666798220741
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,2048,6144,0.020871111088328894
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,256,128,0.006569777925809224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2048,512,0.0028897778441508612
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,4096,7168,0.012727110750145383
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1536,16384,0.0147706667582194
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,7168,128,0.0052737775776121355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,4096,1536,0.01613333324591319
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,128,5120,0.004254222330119875
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,256,8192,0.01147466649611791
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,16384,2560,0.012345777617560493
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,7168,12288,0.04592355423503452
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,10240,8192,0.49398846096462673
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,10240,512,0.003936000168323517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,8192,5120,0.017498667041460674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,7168,1024,0.005933333188295364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,6144,2560,0.007675555845101674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,10240,10240,0.059593776861826576
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2560,1536,0.00490488898422983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,8192,256,0.005560889012283749
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,7168,65536,0.13647288746303982
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,16384,64,0.002242666565709644
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,64,4096,0.004607111215591431
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2048,4096,0.007983999947706858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1536,6144,0.01146044416560067
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,16384,128,0.006984000404675801
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,256,512,0.006931555353932911
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,4096,4096,0.013967111706733704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,5120,7168,0.03945066531499227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,12288,256,0.007698666718271043
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,2048,1536,0.006622222148709827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,16384,5120,0.04751111070315043
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,1024,1024,0.006955555743641323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3584,8192,0.020943111843532983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,512,4096,0.0177520000272327
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,5120,256,0.005616888817813661
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,256,65536,0.07286666499243842
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,1024,4096,0.00794311116139094
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,10240,32,0.006294222341643439
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,256,5120,0.004970666848950916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2048,1536,0.00628977765639623
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,5120,65536,0.06924266947640313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,768,2560,0.007277333074145847
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2560,1024,0.004981333182917701
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,5120,256,0.004945777770545748
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3072,2560,0.013209777573744455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,7168,6144,0.029843555556403265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3584,65536,0.07124444511201647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,256,12288,0.028579556279712256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,65536,3072,0.06008444229761759
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,2560,65536,0.08454310894012451
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,256,512,0.007041778001520369
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,65536,4096,0.10410844617419773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3584,6144,0.08633955319722493
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3072,768,0.0029324444217814338
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,1536,1536,0.016560888952679105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,1024,7168,0.009727111293209923
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,6144,2048,0.007672000262472365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,64,512,0.00259911103381051
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,64,16384,0.01574488812022739
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3072,256,0.0022417778770128884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,768,128,0.018816888332366943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,16384,2048,0.02033244404527876
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2560,12288,0.01958933307064904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,128,128,0.006246222390068903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,3072,6144,0.031141334109836157
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,5120,3584,0.008607111043400234
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,128,256,0.005623111294375525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,5120,12288,0.04481777879926893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,256,2560,0.03241599930657281
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2048,256,0.00564533347884814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,5120,7168,0.015390222271283468
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,5120,16384,0.04677155613899231
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,128,4096,0.00831288927131229
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,768,12288,0.013762666947311826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,10240,3584,0.014762666490342883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,10240,65536,5.949184841579861
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,65536,768,0.02565244502491421
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,6144,768,0.03322844372855292
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,128,128,0.0052577778697013855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,2560,2048,0.008378666308191087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3072,10240,0.01899377836121453
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2560,512,0.011012444065676795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,65536,6144,0.14918133947584364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,128,512,0.005481777919663324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,6144,10240,0.028149333265092637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,5120,10240,0.01591999994383918
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1024,12288,0.0372924440436893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,16384,128,0.006150222072998683
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,6144,10240,0.02607288956642151
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,3072,512,0.006620444357395172
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,6144,1024,0.006239999913507038
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,512,256,0.006024888820118374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3584,65536,0.2122622198528714
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,2560,10240,0.01763911048571269
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,16384,3072,0.018435556027624343
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,8192,3584,0.012191111014948951
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,256,1536,0.006679111056857639
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,7168,12288,0.04507199923197428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,128,64,0.0022675556441148124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,7168,5120,0.018096889058748882
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,1024,768,0.0058959999846087555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3072,16384,0.019367110398080613
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,4096,6144,0.012487110992272695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,10240,1024,0.007229333122571309
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3072,128,0.0026417778184016547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,12288,512,0.02639466689692603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,2560,6144,0.012120000190205045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,768,10240,0.00814133303032981
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1536,512,0.0052560000783867305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,8192,12288,0.042415112257003784
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,64,65536,0.0069395556218094295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3072,768,0.006313777632183499
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,6144,12288,0.0222062220176061
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,256,5120,0.024033778243594702
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,10240,128,0.006638222270541721
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2560,3584,0.007592889169851939
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3584,3584,0.011042666931947073
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,16384,4096,0.019187556372748483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2048,3584,0.0076142218377855085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1024,32,0.005567111074924469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,16384,16384,0.045019557078679405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1024,256,0.002352888799375958
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,12288,768,0.007648888561460707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,8192,64,0.0020968889196713767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,4096,16384,0.05704977777269152
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3584,768,0.006808888994985157
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1024,128,0.0021724444296624926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2048,256,0.002489777695801523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,2048,6144,0.009691555466916826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,3072,128,0.006503110958470239
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,2048,7168,0.037709332174725003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,4096,4096,0.020213334096802604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,10240,65536,0.19001332918802896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,768,1536,0.00665866666369968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,1536,16384,0.023698666029506262
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,10240,12288,0.03212177753448486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,256,2560,0.006981333096822103
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,65536,3584,0.03511200017399258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,6144,4096,0.02427822185887231
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,512,3072,0.05998488929536608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,10240,1536,0.152712000740899
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,128,12288,0.013400889105266996
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,3584,10240,0.016189333465364244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,6144,12288,0.5069404708014594
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,7168,4096,0.0120666664507654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,5120,1024,0.009872000250551436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,64,64,0.005571555760171678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,7168,2048,0.023818666736284893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,4096,5120,0.011834666960769229
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,65536,10240,0.1625937753253513
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,5120,16384,0.023804444405767653
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,6144,6144,0.0530426667796241
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,8192,7168,0.04404177930619982
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,6144,6144,0.012736000120639801
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,1024,12288,0.04251022140185038
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,768,8192,0.02100444502300686
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,256,768,0.005938666562239329
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,3072,6144,0.021677333447668288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,16384,1024,0.03374222252104018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,7168,12288,0.029370665550231934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,768,768,0.01548266741964552
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,1536,3584,0.02476177778508928
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,3584,7168,0.01036177741156684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,12288,2560,0.10918844408459133
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,128,64,0.005697777701748743
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,8192,65536,0.08260266648398505
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,1536,10240,0.027160889572567407
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,1024,7168,0.009773333039548662
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1536,64,0.0035217776894569397
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,6144,1024,0.017874666386180453
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,65536,1536,0.32597067621019155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,6144,8192,0.04450577828619215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,512,1024,0.003181333343187968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3584,5120,0.011095110740926532
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,10240,5120,0.017961778574519686
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,65536,768,0.016504888733228047
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,128,7168,0.015818667080667283
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,7168,2048,0.013835555149449242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,512,256,0.0052355556852287715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,128,12288,0.013302221894264221
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,65536,512,0.012516444755925072
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2048,10240,0.015636444091796875
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,128,512,0.005927111125654644
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,12288,1024,0.007618666523032718
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,2048,4096,0.01813066667980618
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3072,1024,0.0031431110368834603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,6144,12288,0.05144266618622673
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,16384,5120,0.07198933098051283
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3584,6144,0.012683555483818054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,512,12288,0.013321778012646569
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,4096,256,0.026167111264334783
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,1536,768,0.02568622264597151
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,5120,2560,0.016414221790101793
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,128,5120,0.013070222404268054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3584,1536,0.006575999988449945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,768,3584,0.03240622083346049
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,256,7168,0.010024000373151567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,12288,16384,0.10082488589816624
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,128,8192,0.011335111326641507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,4096,7168,0.013191111385822296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,128,3584,0.014155555102560254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2048,8192,0.01255022237698237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,65536,2560,0.8821075757344564
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,65536,3072,0.042709334029091724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,8192,4096,0.019440000255902607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,128,64,0.00999111102686988
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,6144,5120,0.016959110895792644
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,2048,4096,0.025787555509143408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,512,64,0.00665244460105896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,8192,768,0.005942222144868638
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,2048,1536,0.006609777609507243
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,256,1536,0.005640888793600931
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,10240,12288,0.11941511101192898
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,3584,8192,0.029358221424950495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,2560,6144,0.03267466690805223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,128,3584,0.005919999960396025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,128,7168,0.009697777529557547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,4096,6144,0.023918223049905565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,7168,3584,0.010784000158309937
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,512,4096,0.00592533333433999
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,3584,4096,0.007984889050324758
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,8192,2560,0.008948444492287105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,3584,1024,0.01719111038578881
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,7168,256,0.02141688929663764
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,256,2048,0.013567999833159976
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,512,3072,0.007266666326257918
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,3584,1024,0.016238222519556682
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,512,3072,0.011427555647161273
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,7168,6144,0.03180888957447476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,2560,128,0.01585244470172458
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,10240,2048,0.025456888808144465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,1024,5120,0.0381022228135003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,16384,5120,0.01906044450071123
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,8192,2048,0.010439111126793755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,3072,16384,0.04152977797720168
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,256,64,0.001814222170246972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,10240,7168,0.210225780804952
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3072,65536,0.07387110922071669
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,7168,16384,0.07216533025105794
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,512,7168,0.010741333166758219
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2560,2560,0.005615111026499007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2560,1024,0.004552000098758274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,2048,65536,0.1011893351872762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,768,12288,0.017913776967260573
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,768,256,0.0023911111056804657
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,128,4096,0.007650666766696506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,16384,4096,0.020656888683636982
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3584,12288,0.029531555043326482
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3584,4096,0.01591022147072686
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,5120,512,0.0035093331502543557
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2560,65536,1.0388657251993816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2048,3584,0.014557333456145393
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,8192,3072,0.015231110983424716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,5120,4096,0.06618399752510919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,12288,3584,0.014053333136770459
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,16384,3584,0.037590222226248846
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,6144,64,0.017848889032999676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,8192,4096,0.015619556109110514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,256,7168,0.00868888861603207
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,7168,64,0.002340444467133946
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,256,2560,0.0099973330895106
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,128,5120,0.01692800058258904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,2560,16384,0.03260088960329691
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,768,7168,0.009877333210574256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,7168,3584,0.010661332971519895
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,10240,1536,0.008936000367005667
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,128,3072,0.008184000021881526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,16384,16384,0.045716444651285805
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,3584,12288,0.36275821261935765
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,12288,512,0.01994222237004174
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,8192,10240,0.07414666811625163
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,128,1024,0.003175111073586676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,2048,12288,0.08985422054926555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2560,10240,0.01667644414636824
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,512,7168,0.007607111500369177
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,12288,3584,0.018946665856573317
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,8192,64,0.006248000181383557
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,8192,512,0.005611555443869696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,16384,768,0.030117332935333252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,3584,4096,0.028196444114049275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,3584,16384,0.03791022300720215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,512,3584,0.015196444259749519
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,768,128,0.002532444480392668
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,256,512,0.012175111307038201
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,4096,2048,0.007038222418891058
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,3584,768,0.017880888448821176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1024,4096,0.010649777948856354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,256,3072,0.007896888587209914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,2048,128,0.005937777873542573
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,256,5120,0.016033778587977093
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,8192,5120,0.012792000340090858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,5120,5120,0.010838222172525195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,12288,65536,13.882146199544271
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3584,16384,0.03372088736957974
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,10240,65536,0.31445956230163574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,16384,3072,0.01866044435236189
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,128,65536,0.02760800056987339
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,7168,2048,0.008640888664457533
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,5120,2560,0.12200266785091823
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,5120,1536,0.011013333168294696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,12288,1536,0.009153777526484596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,8192,64,0.009347555538018545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,8192,256,0.005560889012283749
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,512,768,0.0028408887899584244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,1024,16384,0.029073778125974867
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3072,2048,0.007104000283612146
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1536,768,0.0038559999730851916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,1024,512,0.00592533333433999
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3072,768,0.005906666732496685
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,4096,7168,0.04419022136264377
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3072,768,0.006259555617968242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,4096,2560,0.013344888885815939
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,8192,3072,0.030010667112138536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,7168,2560,0.02751200066672431
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,65536,3584,0.3355404535929362
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,5120,8192,0.014254222313563028
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,12288,512,0.006688000013430913
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,5120,2048,0.013703111145231458
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,4096,4096,0.01036000003417333
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,512,5120,0.009600888523790572
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,16384,7168,0.03417777683999803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,10240,10240,0.03915199968549941
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2560,16384,0.0214044451713562
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,256,3584,0.00868800034125646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,64,16384,0.017059556312031217
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,65536,64,0.004271111140648524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,12288,12288,0.12065154976314968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,8192,768,0.005970666805903117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,128,16384,0.042931556701660156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,768,2560,0.0069884442620807225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,5120,3584,0.03503466645876566
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,1536,256,0.014775110615624322
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1024,64,0.0052879999081293745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,2560,2560,0.016678222351604037
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,768,32,0.0025617778301239014
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3072,2048,0.005003555367390315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,1536,5120,0.01935822268327077
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,8192,128,0.0022604444788561925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,6144,8192,0.022135999467637803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,4096,6144,0.020623111062579684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3584,2048,0.0068986668354935115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,12288,768,0.017495110630989075
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,4096,5120,0.008632889224423302
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,3072,2048,0.006623111251327727
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,10240,3584,0.016541333662139047
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,65536,12288,0.2467573218875461
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,65536,3072,0.0682435565524631
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2560,65536,0.05770311090681288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,7168,4096,0.011875555747085147
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,8192,10240,0.028227554427252874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,65536,5120,0.10745955838097467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,512,10240,0.006928000185224745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,16384,65536,0.22074133820003936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,256,7168,0.009698666632175446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,4096,8192,0.02349244389269087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,768,128,0.005419555637571547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1536,16384,0.016737777325842116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,8192,3072,0.030025776889589097
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,32,256,0.005392888767851724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,16384,32,0.0028737777223189673
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,3072,128,0.014820444915029736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,10240,12288,0.0713919997215271
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,128,64,0.0018311111877361934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,3072,6144,0.025568889247046575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,65536,8192,0.19508888986375597
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1024,7168,0.017308443784713745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,16384,3584,0.01479111115137736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,16384,2048,0.01311911145846049
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,3072,2560,0.023391111029518977
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,65536,128,0.16998755931854248
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,10240,128,0.006353777729802662
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,12288,12288,0.03251111176278856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,16384,10240,0.0303111109468672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,5120,6144,0.01369511087735494
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,4096,256,0.006561777657932705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,7168,32,0.007639110916190677
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,16384,1536,0.009670221971140968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,4096,5120,0.02142133315404256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,1536,128,0.012076444096035428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,2560,3584,0.10428088903427124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,64,256,0.005224888937340842
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,12288,256,0.01310755560795466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1024,16384,0.02086755633354187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,256,512,0.00590133335855272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,64,4096,0.0052560000783867305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,3584,3584,0.030151999659008447
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,32,7168,0.01035288886891471
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,6144,1536,0.00629333323902554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2560,5120,0.015936889582210116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3584,10240,0.01754577789041731
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,1536,256,0.005255110975768831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,256,1024,0.005894222193294101
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,3584,10240,0.12979911433325872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,10240,256,0.0053013331360287136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,128,2048,0.012747555143303342
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,2048,1536,0.01751377847459581
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1024,3072,0.007599111232492659
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,3584,10240,0.017315554949972365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,256,2048,0.013429332938459186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,32,12288,0.03106044398413764
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,10240,768,0.006243555496136348
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,4096,2048,0.02092711130777995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3072,768,0.003519111209445529
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,3584,65536,0.13789421982235378
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,7168,2560,0.006978666616810693
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1024,4096,0.0443173348903656
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,256,2048,0.006909333169460297
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,16384,8192,0.041033777925703264
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,768,3072,0.016703110602166917
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,3584,512,0.00552977787123786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3072,65536,3.4609519110785594
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,512,768,0.010038221875826517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,2560,16384,0.4195973343319363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,16384,3072,0.04411822226312426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,768,3072,0.015121777852376303
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,512,6144,0.008998221821255153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2560,65536,0.05617600017123752
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,5120,128,0.01478844384352366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2560,6144,0.011398221883508893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,2560,65536,0.06670844554901123
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2048,1024,0.0031902222997612427
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,16384,2560,0.03395911057790121
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,7168,32,0.005938666562239329
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,7168,5120,0.013793778088357715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,2048,8192,0.034972445832358465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2560,512,0.005611555443869696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,3072,2560,0.016170667277442086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,4096,2560,0.027338667048348322
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2560,3072,0.0075031113293435835
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,2560,1536,0.01649599936273363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,2560,65536,0.09103288915422227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,16384,65536,0.41565778520372176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,512,8192,0.011559999651379056
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,10240,4096,0.1427626609802246
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,256,7168,0.01185155577129788
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,16384,3072,0.03609600000911289
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,65536,512,0.21266667048136392
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1024,64,0.0029128889242808023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,768,3584,0.007354666789372762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,32,12288,0.0046160001721647055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,3072,10240,0.02752888864941067
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,10240,3072,0.02581866747803158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1536,16384,0.011886222494973077
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3072,3584,0.013722666435771518
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,128,1024,0.00600355573826366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,3584,512,0.005597333527273602
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3584,6144,0.01499555508295695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,7168,2560,0.009894222021102905
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,8192,1536,0.02184177769554986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,10240,8192,0.05253333184454176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,8192,256,0.005582222094138463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,1536,16384,0.017793777916166518
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2560,32,0.005983110931184556
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1536,10240,0.09116355578104655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,2560,2048,0.015951999359660678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3584,3584,0.1167644394768609
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,65536,2048,0.026735110415352717
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1024,5120,0.00556355549229516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3584,1024,0.019712888532214694
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,128,3584,0.008662222160233391
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,3584,65536,0.4913164244757758
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,6144,3584,0.07710666788948907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,10240,256,0.0058906666106647914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,1024,1536,0.014780445231331719
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1024,2560,0.007148444652557373
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,5120,3072,0.007864889171388414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,3072,16384,0.02703999976317088
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,16384,5120,0.07009333372116089
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,512,10240,0.012038222617573209
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,5120,2560,0.054973334074020386
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,6144,128,0.0069120000633928515
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,2048,7168,0.015089778436554803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,6144,3584,0.007852444218264686
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2048,3584,0.007956444389290279
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,256,2560,0.009328000247478485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,32,6144,0.004929777648713854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,64,512,0.0058204444746176405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,128,12288,0.005260444349712796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,32,10240,0.02499288817246755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1536,65536,0.0335804455810123
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1536,5120,0.01585155559910668
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,65536,10240,0.0815911094347636
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,32,4096,0.006075555665625467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,64,7168,0.01593866613176134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3072,6144,0.01906844476858775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,128,2048,0.006683555742104848
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,1536,128,0.01553066737122006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,2048,2048,0.012190221912331052
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,128,12288,0.013797333670987023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,7168,6144,0.03855289022127787
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,8192,65536,0.9551351335313586
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,3072,512,0.02439733346303304
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,16384,5120,0.02630311085118188
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,8192,5120,0.0421377784676022
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2560,65536,0.2000782224867079
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,5120,1536,0.041264888313081526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,6144,65536,0.3895920117696126
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,6144,6144,0.1849671072430081
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,5120,256,0.0055440002017551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,1536,5120,0.018719111879666645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,1024,8192,0.02569155560599433
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,128,2560,0.02158755560715993
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2560,1024,0.007887110941939885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1024,7168,0.019913777709007263
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3072,3072,0.007274666594134436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,5120,3072,0.007782222496138678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,768,10240,0.02882844540807936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,16384,5120,0.04800711075464884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,6144,2560,0.021925333473417494
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,16384,2048,0.12101244926452637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,12288,2560,0.01387111097574234
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,65536,5120,0.08195110824373034
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,12288,6144,0.02369600037733714
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,16384,4096,0.02208799951606327
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,512,256,0.005567111074924469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,2560,1536,0.016890666551060148
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,16384,1536,0.17401599884033203
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,512,12288,0.011503111157152386
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,32,12288,0.013423110875818463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,3072,16384,0.061234666241539844
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,512,6144,0.006308444258239534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2048,128,0.0041893331540955436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1536,3072,0.02812355425622728
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1536,512,0.005229333208666907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,8192,128,0.0058906666106647914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,12288,10240,0.0444435543484158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,65536,3072,0.08577511045667861
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2560,4096,0.012749333348539142
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,5120,512,0.005547555370463266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,12288,8192,0.02276533345381419
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1024,128,0.0021662221600612006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,128,7168,0.017535999417304993
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,12288,3584,0.013131555583741931
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,64,12288,0.009345778160625035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,1536,5120,0.06568088796403673
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,64,64,0.0021751111166344746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,7168,6144,0.3880559868282742
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,4096,10240,0.09672177500194973
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,128,2048,0.016520000166363187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,65536,10240,0.19798578156365287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,10240,8192,0.036687999963760376
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,512,3584,0.012679110798570844
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,1024,65536,0.08370311392678155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,8192,1536,0.006648889018429651
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,2048,3584,0.016160888804329764
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,12288,5120,0.02453511125511593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,12288,7168,0.04050222370359633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,5120,7168,0.018624888526068795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,1024,8192,0.010729778144094678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,2560,3072,0.015857777661747403
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,16384,2048,0.017624000708262127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,5120,8192,0.0941351122326321
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,64,8192,0.005663110978073544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,64,6144,0.01406311160988278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,2048,256,0.01586844523747762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3072,3072,0.011408888631396823
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,768,768,0.00665866666369968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1536,10240,0.01240711079703437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,512,12288,0.02737866673204634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,16384,6144,0.08103822337256537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,7168,2048,0.007607111500369177
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,4096,3072,0.007251555720965068
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,16384,4096,0.04366755485534668
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,32,64,0.005595555735958948
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,7168,768,0.055895997418297656
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,4096,5120,0.021881777379247878
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,64,10240,0.00489777781897121
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,5120,32,0.0031582222630580268
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,1024,512,0.005409777578380372
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,5120,768,0.007242666350470648
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3072,256,0.005647999958859549
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,512,3584,0.00794044468137953
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,8192,1536,0.03358488943841722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,5120,8192,0.017901332841979135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,8192,256,0.005567111074924469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,10240,4096,0.01628977722591824
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,768,128,0.005567111074924469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,5120,12288,0.040185779333114624
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,16384,768,0.01997155613369412
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,1024,65536,0.6220435566372341
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,3072,10240,0.01440177857875824
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1024,7168,0.0857786668671502
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,5120,2560,0.10994933048884074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,32,7168,0.006136000156402588
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,2048,4096,0.017729777428838942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,12288,8192,0.03505955470932855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,65536,16384,0.23929866154988608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,64,65536,0.010095111197895473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,16384,2048,0.0710373322168986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,5120,3072,0.007745777567227681
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,64,32,0.0018826667219400406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2560,768,0.005940444353553984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,2560,12288,0.06899555524190266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,2560,3072,0.022089777721299067
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1536,3584,0.007698666718271043
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,512,10240,0.022357333037588332
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,4096,1536,0.016911110944218107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,12288,3072,0.12094310919443767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3072,2560,0.007886221839321984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,1536,2048,0.007335111498832703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2048,128,0.005954666684071223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,128,2048,0.012794666820102267
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,256,2560,0.010409778190983666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,12288,7168,0.02639377779430813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1536,10240,0.014831999937693277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1536,5120,0.008671999805503422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,512,12288,0.014180445008807711
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,768,10240,0.012128888732857175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,3072,8192,0.033144887950685285
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,65536,1024,0.015110222829712762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,32,64,0.0018364443547195857
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1536,3584,0.008284444610277811
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,128,4096,0.014578666951921252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,12288,4096,0.023570666710535686
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,64,64,0.0052720002002186244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3072,10240,0.016940444707870483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,768,3584,0.018231110440360177
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,32,1024,0.0031235555393828284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,7168,12288,0.018343110879262287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,128,256,0.002151111140847206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,8192,6144,0.02642755541536543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,512,16384,0.017323555217848886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,5120,3584,0.010001777774757808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3584,6144,0.027121777335802715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1024,4096,0.00794311116139094
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,128,5120,0.008001777860853408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,3584,10240,0.015287111202875773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,7168,6144,0.09863822327719794
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3072,3584,0.018696889281272888
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,1024,16384,0.05622844563590156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,32,64,0.001957333336273829
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1024,8192,0.007720888488822513
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,5120,3584,0.09400622049967448
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,64,2048,0.004882666799757215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,256,4096,0.006583999842405319
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,256,128,0.0052373334765434265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,65536,32,0.00998755544424057
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,6144,6144,0.025416000021828547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,256,32,0.0018666666001081467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,32,65536,0.009739555418491364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,256,12288,0.013068444199032254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,8192,32,0.01034044474363327
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,5120,32,0.002500444443689452
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,6144,5120,0.010287111004193624
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1536,768,0.005567111074924469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,3584,1024,0.017608000172509086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,7168,65536,0.0862648884455363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1024,128,0.008336889247099558
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,16384,1024,0.009272889130645329
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,6144,1024,0.022683555881182354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,16384,6144,0.02012977831893497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,65536,5120,2.700920952690972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3072,6144,0.020940444535679285
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,768,3584,0.022156443860795762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,256,12288,0.05196711089875963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,128,4096,0.005615111026499007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,512,7168,0.009706666900051964
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,768,1536,0.006731555693679386
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2560,32,0.017916444275114272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,768,1536,0.008601778083377415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3584,512,0.005635555419656966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,1024,12288,0.020068445139461093
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,64,256,0.007040888898902469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,6144,5120,0.012840888566441007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,4096,12288,0.01757600075668759
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,512,2560,0.026220444175932143
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,64,4096,0.00810933361450831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,3072,12288,0.0177520000272327
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,256,2048,0.019690665933820937
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,2048,3584,0.007600000335110559
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,512,16384,0.016151999433835346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,6144,16384,0.034360888931486346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,16384,16384,0.10877333084742229
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,768,32,0.0021555556191338436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,65536,3584,0.041103111373053655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,12288,256,0.007273777491516537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,10240,768,0.010399111443095738
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,768,8192,0.013010666602187686
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,5120,3072,0.007287999822033777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,8192,768,0.020997333857748244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,6144,12288,0.2626720004611545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,64,8192,0.005586666779385672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2048,128,0.0021786666992637846
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3584,12288,0.3025040096706814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3584,512,0.011729777687125735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,12288,256,0.023433778021070693
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,512,3072,0.014854222536087036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,768,7168,0.010431110858917236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,768,16384,0.01495822270711263
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,6144,256,0.016937777400016785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,768,768,0.008980444735950893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,1536,5120,0.026029333472251892
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,512,768,0.0038373333712418876
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,3072,65536,0.06557599703470866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,128,128,0.01105955574247572
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,65536,768,0.015071999695565967
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3584,2048,0.034303999609417386
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,128,7168,0.010829333629873065
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1536,4096,0.012097777591811286
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,3072,2560,0.018382221460342407
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,2048,12288,0.03244977858331468
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,64,10240,0.050072001086341016
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,10240,10240,0.7969688839382596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,10240,6144,0.0343466666009691
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,768,3072,0.008912888665994009
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,128,12288,0.00832444429397583
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,64,64,0.008613333106040955
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,10240,512,0.004787555585304896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,512,512,0.006544000158707301
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,16384,8192,0.05784000290764702
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,7168,64,0.005231110999981563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,16384,128,0.0147724449634552
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,768,65536,0.023415999280081853
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,8192,1024,0.006258666515350342
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,1536,10240,0.011707555916574268
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,8192,1536,0.022286223040686712
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,10240,5120,0.047157333956824414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,256,8192,0.005904888941182031
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,6144,768,0.01907911068863339
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,12288,5120,0.017156443662113614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,1024,3584,0.02238133384121789
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,6144,1024,0.006657777975002925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,8192,1536,0.006851555572615729
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,512,3584,0.015142222245534262
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,7168,8192,0.06687821944554646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,10240,5120,0.2428142229715983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,768,7168,0.010002666877375709
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,4096,7168,0.04084622197681003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,768,16384,0.0496711101796892
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,256,6144,0.0076044441925154785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2048,4096,0.008410666551854875
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,256,2048,0.0042257776690853965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,5120,7168,0.015982222225930955
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2560,8192,0.010985777609878115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,10240,3584,0.015081778168678284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,5120,16384,0.040842665566338435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,8192,1536,0.006888888776302338
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,1536,128,0.014478221535682678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,128,2560,0.013112000293201871
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3584,5120,0.011691555380821228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,5120,128,0.014762666490342883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,7168,3072,0.028322666883468628
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,5120,256,0.007632888853549957
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,256,8192,0.018608889646000333
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,12288,3584,0.03403288788265652
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,256,16384,0.008808000220192803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,7168,16384,0.0571955574883355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,65536,768,0.04352533486154345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,3072,2560,0.0069448889957533936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,2560,8192,0.047826667626698814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,256,12288,0.019533332851197984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,8192,32,0.0025840000145965153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,16384,2560,0.013243555194801755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,4096,7168,0.07112977902094524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,256,8192,0.010641777680979835
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,64,6144,0.0063475556671619415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,3584,768,0.018664000762833487
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2560,512,0.005601777798599667
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,768,3584,0.015808888607554965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,8192,10240,0.21316799852583143
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,10240,65536,0.20100977685716417
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,32,2560,0.004932444542646408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,6144,6144,0.040720889965693154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,32,10240,0.005030222237110138
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,65536,3072,0.07805066638522677
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,2560,8192,0.06805688805050321
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3072,1536,0.011559999651379056
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,2048,16384,0.04656444324387444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,65536,7168,0.09142044517729019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,2048,2560,0.022660444180170696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,16384,1536,0.028663999504513208
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,64,1536,0.006613333192136552
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,128,2048,0.01276177747382058
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,4096,256,0.005565333283609814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,16384,1536,0.10012088881598578
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,4096,4096,0.02936089038848877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,768,65536,0.10044266780217488
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,512,16384,0.008629333641793992
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,7168,3584,0.04017422265476651
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,16384,1536,0.08932711018456353
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,16384,256,0.017186666528383892
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,64,64,0.007992000215583378
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,6144,2048,0.007762666377756331
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,2048,4096,0.009344889058007134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,256,3072,0.004930666751331753
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1536,1024,0.006205333189831839
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,6144,12288,0.02623822291692098
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,5120,5120,0.032680888970692955
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3584,512,0.005607111172543631
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,6144,65536,0.07555022504594591
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,16384,256,0.0031911111954185697
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,3072,8192,0.024056888288921777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1536,4096,0.007346666521496243
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,256,5120,0.008372444245550368
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,6144,5120,0.016996444927321542
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,65536,6144,0.06750310791863336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,4096,10240,0.05602044529385037
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,3072,1024,0.014615999327765571
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,4096,3072,0.012481778032249875
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,8192,128,0.0021582223061058256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,16384,10240,1.0237973531087239
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2560,6144,0.24966933992173937
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,128,2560,0.026183999247021143
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,4096,1024,0.024130667249361675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,65536,8192,0.17970666620466444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3072,32,0.005610666755172942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,256,32,0.002042666698495547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,128,2048,0.004581333448489507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,256,8192,0.010437332921557956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,12288,256,0.017252445220947266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3584,6144,0.012684444586435953
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3584,65536,0.06794933478037517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,512,3072,0.007335999773608313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,7168,3072,0.015255999233987598
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,256,7168,0.008353778057628209
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3072,1536,0.007540444533030192
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,65536,12288,0.13082488377888998
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,512,2048,0.004554666578769684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1536,10240,0.026055999928050574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,32,512,0.002516444358560774
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,768,1536,0.014848889576064216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,512,65536,0.2904328770107693
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,3584,4096,0.019631110959582858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,128,3072,0.005006222261322869
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,128,3584,0.008231110870838165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3072,2048,0.0069848886794514125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,512,64,0.005958222266700532
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,65536,256,0.022286223040686712
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,2560,256,0.015831111205948722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,256,2560,0.007286222444640265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2560,3584,0.00832177781396442
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1024,16384,0.029828445778952703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1536,768,0.0035217776894569397
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,512,768,0.0068417779273457
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3072,12288,0.19030222627851698
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,256,16384,0.011512888802422417
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,3584,768,0.018226666582955253
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1024,2560,0.006940444310506185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3584,32,0.0031617778456873367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,2048,1536,0.017575111654069688
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,10240,8192,0.02236533330546485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3584,256,0.005647111270162795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,6144,12288,0.019476445184813607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,7168,10240,0.020770667327774894
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,128,16384,0.026068444053332012
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,256,1024,0.0063271112740039825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,128,16384,0.015473778049151102
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,12288,3072,0.02736888825893402
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2048,10240,0.010474666953086853
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,4096,128,0.00564444437623024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,5120,10240,0.053929779264662
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,5120,2048,0.006700444552633498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,3072,1024,0.015850666496488783
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,8192,10240,0.021075555019908484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,128,32,0.005453333258628845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,6144,2560,0.0075937774446275495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3072,2048,0.006625777731339137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,128,16384,0.07612711191177368
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,16384,512,0.015861334072219003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1536,65536,0.13211111227671304
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,768,16384,0.015483554866578845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,16384,7168,0.347126219007704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,128,2048,0.009434666898515489
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,2048,12288,0.029126223590638902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,768,6144,0.00832711077398724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,256,128,0.014415999253590902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,32,2048,0.006892444358931647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,4096,7168,0.030322667625215318
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,4096,512,0.027132444911532935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,768,2560,0.015976889265908133
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,65536,5120,0.05826933516396416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,5120,32,0.006230222268237009
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,6144,3072,0.02195555633968777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,7168,2048,0.021988444858127173
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3584,8192,0.015988444288571674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3072,768,0.003941333128346337
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,128,1024,0.012456888953844706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,512,32,0.009333333207501305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,16384,2048,0.03157866663402981
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,65536,6144,1.2875964906480577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,6144,8192,0.4321777820587158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,512,4096,0.013391999734772576
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,128,512,0.015486222174432544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,10240,10240,0.03524444500605265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,512,512,0.016746666696336534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,10240,128,0.010220444036854638
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,768,4096,0.019162666466501024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3584,256,0.005940444353553984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,2048,2048,0.014520888527234396
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1536,2048,0.006948444578382704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,7168,1024,0.007904888855086433
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1024,4096,0.007723555796676212
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,5120,4096,0.011702222128709158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,10240,512,0.018264888061417472
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,768,5120,0.008708444734414419
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,64,2560,0.006917333437336816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,65536,3072,0.051129778226216636
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3584,3072,0.007320000065697565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,256,6144,0.023942222197850544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,8192,65536,0.09489422374301487
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,5120,128,0.006585777633719974
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,512,128,0.0052177777720822235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,2048,4096,0.008295999632941352
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1024,512,0.005568888866239124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,16384,6144,0.022073777185546026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,1536,8192,0.03730222251680162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1024,6144,0.014378666877746582
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,12288,12288,0.09808355569839478
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,2560,5120,0.048735111951828
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,7168,16384,0.03678044345643785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,4096,4096,0.007998222278224098
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,6144,10240,0.03639199998643663
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,128,6144,0.005210666606823604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,4096,128,0.00554577757914861
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,2048,8192,0.01071999967098236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2560,7168,0.012293332980738746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,4096,2048,0.00693866651919153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,6144,4096,0.009330666727489894
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,7168,768,0.005604444278611078
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2048,65536,0.06307822465896606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,16384,256,0.006581333362393909
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,10240,128,0.005566222386227713
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2560,32,0.0052826665341854095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,512,65536,0.08871289094289143
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,4096,256,0.02167555524243249
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,256,12288,0.028839111328125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,65536,256,0.012896888785892062
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,16384,6144,0.04182666540145874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,16384,7168,0.027249778310457867
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,128,6144,0.036650667587916054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,64,6144,0.016184000505341422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,256,10240,0.012364444633324942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,32,1536,0.0034737777378824023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,4096,128,0.02882400155067444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,8192,512,0.009184888667530483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1536,512,0.005986666513813867
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3584,8192,0.01461155547036065
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,256,3072,0.007294221884674496
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,65536,12288,0.15569067001342773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,65536,5120,0.12980088922712538
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,128,3584,0.007515555454625024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,10240,768,0.0052177777720822235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,3584,2560,0.041922665304607816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,3584,65536,0.41949065526326496
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,768,256,0.0025155554629034465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,65536,768,0.022500443789694045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,768,65536,0.1872080034679837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,1536,4096,0.026696888936890498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,8192,256,0.0052800000541739995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,512,1536,0.006252444452709622
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,3584,3072,0.06078044573465983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,5120,6144,0.024864888853496973
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,1024,12288,0.01311911145846049
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,1024,12288,0.23156799210442436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,65536,6144,0.5476995574103461
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,7168,64,0.002513777878549364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,10240,4096,0.06937422355016072
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,3584,7168,0.010423999693658618
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,16384,1536,0.013271999855836233
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,10240,1024,0.006960888703664144
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,128,3072,0.007692444655630324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,16384,256,0.02500444485081567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,6144,8192,0.0425164434644911
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,65536,512,0.014813333749771118
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,32,32,0.0020879999630981022
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,5120,65536,0.22016089492373994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,768,768,0.005591111050711737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,32,7168,0.004211555752489301
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,6144,128,0.0026568888376156488
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,16384,256,0.0034737777378824023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,10240,7168,0.033391998873816595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,7168,8192,0.03628799981541104
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,5120,12288,0.03528977764977349
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,6144,16384,0.023399111297395494
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,12288,256,0.028147555059856836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,8192,6144,0.012848888834317526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,12288,1024,0.030673778719372217
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,4096,512,0.014309333430396186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,8192,4096,0.013408000270525614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,3584,256,0.01521511044767168
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,4096,5120,0.013367111484209696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2560,3072,0.011136000355084738
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,7168,2560,0.025408888856569927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,4096,256,0.005531555662552516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,256,128,0.006278222219811545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3584,32,0.005931555396980709
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,10240,65536,0.5415635638766819
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,64,512,0.008349333372380998
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,3584,3584,0.010412444670995077
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,768,5120,0.009705777797434065
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,5120,2048,0.006994666324721442
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,10240,3584,0.013820444544156393
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,10240,5120,0.02992266747686598
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,2560,8192,0.02369600037733714
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,8192,8192,0.023511999183230933
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2560,2048,0.007310222420427535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,7168,2048,0.010300444232092964
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,12288,8192,0.039905776580174766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,5120,128,0.0029297777348094513
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,10240,8192,0.05809511078728569
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,2048,12288,0.027760000692473516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,768,12288,0.01350311107105679
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,1536,1024,0.005949333310127258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,6144,2560,0.008421333299742805
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,7168,16384,0.02751911183198293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3584,1024,0.004954666727119022
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,768,5120,0.00850755555762185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,768,2048,0.006696888970004187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,5120,5120,0.012823999755912356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,3072,1024,0.005674666828579373
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,16384,256,0.017623111605644226
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,64,65536,0.15601600541008845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,5120,1536,0.00534400012758043
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,5120,2048,0.007027555671003129
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,256,12288,0.017695999807781644
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,6144,12288,0.05029333300060696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,128,10240,0.02060622142420875
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3072,512,0.005643555687533484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,128,256,0.005912000106440649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,7168,3072,0.021915555000305176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,6144,512,0.009007111191749573
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,8192,4096,0.026143999563323125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,5120,65536,0.07185510794321696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,8192,256,0.005924444645643234
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,16384,768,0.011170667078759937
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,64,1024,0.006953777538405524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,4096,65536,0.06668977604971991
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,4096,16384,0.03632622294955783
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2048,1024,0.0032293332947625052
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,128,1536,0.003642666671011183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,1536,768,0.00666844430896971
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1536,128,0.005609777652555042
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,4096,5120,0.024139554964171514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,1536,3584,0.015962666935390897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,128,1024,0.005969777703285217
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,65536,256,0.012235555383894177
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,10240,10240,0.04739822281731499
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,4096,16384,0.2698426776462131
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,768,7168,0.0191920002301534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2560,512,0.004182222402758068
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,128,12288,0.014001778430408902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,10240,8192,0.0347653329372406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,256,2560,0.006991110742092133
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1024,1536,0.003847111016511917
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,128,1536,0.0035093331502543557
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,7168,128,0.0034675554682811103
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,8192,8192,0.016137777103318107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,512,6144,0.018013333280881245
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,256,5120,0.009230222139093611
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,12288,32,0.00217333332531982
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,2560,128,0.0052471111218134565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,256,4096,0.06971466541290283
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,4096,8192,0.014888000157144336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,3072,1024,0.006631999793979857
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,512,768,0.005579555614127054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,2560,512,0.005962666538026597
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,64,1536,0.0036115555299652945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,65536,10240,0.14933066897922093
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,16384,4096,0.032299555010265775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,4096,64,0.005580444302823808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,1024,2560,0.007972444097201029
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,1024,16384,0.02444177700413598
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,64,64,0.006026666611433029
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,8192,2560,0.007647111184067196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,4096,512,0.003891555385457145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,65536,6144,0.06689155764049955
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,4096,512,0.007358222372002072
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,4096,5120,0.01589866644806332
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,64,1536,0.003972444269392225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,256,768,0.01423377792040507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,8192,256,0.020601777566803824
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,16384,256,0.060235553317599826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,8192,64,0.005598222215970357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,256,10240,0.08745421965916951
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,4096,32,0.006249777972698212
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,3072,768,0.00792266676823298
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,768,256,0.0024542222834295696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,12288,2048,0.027954667806625366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,4096,3072,0.008307555483447181
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1024,5120,0.016934222645229764
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,4096,12288,0.028852442900339764
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,32,64,0.0019155556543005838
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,2048,16384,0.11448000537024604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,2048,256,0.006639110959238476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,8192,6144,0.015648888217078317
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1536,768,0.006126222097211414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,16384,12288,0.06232177548938327
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,256,3584,0.00489066665371259
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,32,4096,0.007962666451931
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,5120,256,0.011744889120260874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1536,64,0.0018560000591807896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,256,1536,0.03490488727887472
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,65536,16384,0.16819822788238525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,2048,65536,0.2215111123190986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,2560,2048,0.029628445704778034
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3072,32,0.002492444382773505
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,65536,4096,0.26191732618543834
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,128,6144,0.020804444948832195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,12288,128,0.006245333287451003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,3072,7168,0.03397066725624932
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2560,6144,0.011189333266682096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,2560,3584,0.014021333720948962
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1024,32,0.005614222337802251
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2560,64,0.0028613333900769553
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,8192,7168,0.017130666308932833
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,2048,8192,0.04019200139575534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3072,256,0.0025111111915773815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3584,3072,0.0336133340994517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2048,6144,0.010788444843557147
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,768,2048,0.014442667365074158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,7168,8192,0.022648889157507155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,128,2048,0.006820444431569841
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,768,64,0.005229333208666907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,16384,4096,0.10573777887556289
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,768,6144,0.018041777941915724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,32,12288,0.014119111829333834
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,32,256,0.002185777864522404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2560,4096,0.029017776250839233
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,4096,8192,0.0342497792508867
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,1536,3584,0.03343288766013251
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,128,16384,0.005598222215970357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,256,12288,0.027279999521043565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,64,65536,0.008984888593355814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,4096,7168,0.011072888970375061
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,6144,768,0.02625688910484314
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3072,3584,0.008074666890833113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,256,2560,0.005239999956554837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,1536,3584,0.007622222105662028
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,8192,8192,0.10542844401465522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,10240,256,0.004896000027656555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,16384,2048,0.04693422383732266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,6144,256,0.005942222144868638
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,10240,5120,0.03304977880583869
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,768,7168,0.022262222237057153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,12288,1024,0.02162044412559933
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,3584,512,0.03566133313708835
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,8192,12288,0.023764444722069636
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,8192,5120,0.0712640020582411
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,16384,2048,0.013396444420019785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,3584,8192,0.012588444683286877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,6144,12288,0.03673600157101949
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,10240,512,0.00795111142926746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,12288,5120,0.04156000084347195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,12288,65536,0.7253262201944987
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,4096,16384,0.02129777769247691
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,16384,512,0.00408266650305854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,8192,1536,0.03695022066434225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,16384,3072,0.01443199978934394
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,6144,10240,0.03026311265097724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,6144,768,0.0069146665434042616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,32,10240,0.08324888679716322
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2560,128,0.005229333208666907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3072,32,0.005596444424655702
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,4096,65536,0.09086755249235366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1024,2048,0.01314933349688848
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,5120,512,0.005965333431959152
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,12288,32,0.008007110820876228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,128,2048,0.006656000183688269
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,7168,12288,0.04892266790072123
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3584,5120,0.01611199975013733
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,7168,5120,0.036815110180113055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,1024,12288,0.026865777042176988
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3584,7168,0.012141333685980903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,128,6144,0.009118222528033787
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3072,2560,0.06426666842566596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,64,2560,0.006721777634488211
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,7168,512,0.018445332845052082
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2560,64,0.005585777676767773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1536,16384,0.021128889587190416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,768,3072,0.03459199931886461
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,2560,16384,0.033583998680114746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,1024,4096,0.008785777621799046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1536,8192,0.011491555306646558
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2048,64,0.005868444426192178
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,65536,12288,0.1341760026084052
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,10240,256,0.009717333647939894
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,10240,512,0.030415111117892798
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2560,256,0.0028844444702068963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3584,5120,0.05352977911631266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,5120,8192,0.01591111057334476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3072,64,0.006284444282452266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,32,16384,0.010059555371602377
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,1024,768,0.017236444685194228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2048,3584,0.0074755557709270054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,12288,10240,0.04027377896838718
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,16384,3072,0.016993777619467843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,256,1536,0.014615111880832247
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3072,512,0.006979555719428592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,2048,256,0.014871110518773397
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,16384,16384,0.08582044310039944
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,3584,16384,0.021538666552967493
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,512,3584,0.011977777712874942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,256,4096,0.004801777915822135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,64,1024,0.009319999979601966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,64,32,0.0022604444788561925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,2560,10240,0.014136888914638095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2048,64,0.00590133335855272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,4096,3072,0.06362133555942111
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,1536,2048,0.018600889378123812
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1024,7168,0.011374221907721626
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,128,8192,0.01313777764638265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,10240,8192,0.021721776988771226
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,7168,7168,0.014190221826235453
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,3072,10240,0.026504889130592346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,12288,7168,0.020462221569485135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,6144,512,0.005614222337802251
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,1024,3584,0.015663999650213454
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,128,128,0.011776888536082374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,12288,2560,0.014078223043017916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,1024,3584,0.016567111015319824
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2048,3072,0.011880888707107969
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3072,512,0.005684444473849402
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,4096,8192,0.03375644485155741
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,10240,1024,0.007262222468852997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,64,1024,0.0031999999450312722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,256,3072,0.005583999885453119
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,768,5120,0.008840888738632202
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,64,512,0.005962666538026597
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,7168,768,0.029281778468026057
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2048,1024,0.003976888954639435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,768,7168,0.025427555044492085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1536,32,0.005596444424655702
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,12288,512,0.006591111007663939
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,65536,128,0.009697777529557547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3584,256,0.005299555758635203
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,5120,768,0.017228444417317707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,768,16384,0.045596443944507174
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1024,5120,0.012423111332787408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,4096,3584,0.007308444215191736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,1024,128,0.014185777968830533
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,512,768,0.007850666840871176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,4096,65536,0.07110489077038236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3072,64,0.005558222118351195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,1536,3072,0.015191111299726697
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,10240,65536,0.18733778264787462
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,4096,1536,0.0196426659822464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3584,10240,0.015084443820847405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,64,1536,0.003548444559176763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3584,64,0.0022106667359670005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,7168,10240,0.11395200093587239
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2560,1536,0.006608888920810487
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,10240,3584,0.03370577759212918
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,1024,512,0.005314666777849197
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,512,12288,0.026326222552193537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,10240,10240,0.14984356032477483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,3072,768,0.017238222890430026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,32,7168,0.007356444166766272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,12288,65536,0.19621866279178193
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,65536,32,0.022571555442280237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,10240,8192,0.41478223270840114
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,512,256,0.0032479998966058097
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,768,12288,0.011687111523416309
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,8192,2560,0.02966044346491496
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1536,32,0.004982222285535601
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,7168,5120,0.028279112444983587
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2560,1536,0.006672888994216919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,3072,3584,0.007694222033023834
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,6144,128,0.03458844290839301
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,16384,1536,0.010969777901967367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,512,2048,0.006672888994216919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,256,16384,0.00871022211180793
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,8192,8192,0.6536826557583278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,65536,2560,0.05992977486716377
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2048,2560,0.009376889301670922
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,7168,5120,0.17028978135850695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,3584,4096,0.036256889502207436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,128,256,0.003928889003064898
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,768,10240,0.017624000708262127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,512,12288,0.11126844088236491
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,3072,2048,0.02181600034236908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,3584,7168,0.02684533264901903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,768,2560,0.010007999837398529
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,128,6144,0.007682667010360294
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,65536,7168,0.07816711399290296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,5120,12288,0.018765333626005385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,2048,8192,0.012459555433856117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,65536,12288,0.24758309788174102
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,4096,1536,0.006360888895061281
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,8192,65536,0.13758488496144614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2048,3072,0.00793511089351442
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1536,32,0.0020888888587554297
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,12288,64,0.012106666962305704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,256,2048,0.017243555850452848
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,256,5120,0.004861333303981357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,4096,65536,2.5843902164035373
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,12288,128,0.009296000003814697
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,16384,1024,0.008068444828192392
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,16384,3072,0.01761155492729611
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1536,64,0.002591111179855135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,4096,7168,0.01681422193845113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,10240,16384,0.08503199948204888
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,12288,128,0.008377778033415476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,768,1536,0.007363555332024892
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,8192,2048,0.023644444015291
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,65536,7168,0.140940441025628
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,65536,3584,0.13137243853674993
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,7168,2560,0.010011555420027839
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,3072,1536,0.018259555101394653
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,2560,4096,0.025069332785076563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,32,65536,0.11679200331370036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,32,6144,0.009291555318567488
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,2048,1024,0.005987555616431766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,65536,12288,0.23266755210028756
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,12288,32,0.0021626665774318907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,7168,10240,0.05712178018358019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,6144,256,0.01685599982738495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,5120,2560,0.0069360000391801195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,8192,768,0.007624000310897827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,12288,2560,0.01329422245422999
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,768,3072,0.00563822231358952
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,2560,65536,0.0858408874935574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,10240,3584,0.020263999700546265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,5120,3072,0.01628888812330034
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,7168,4096,0.019201777047581144
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,6144,5120,0.03529689047071669
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3584,128,0.0028568889117903183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,7168,10240,0.017656889226701524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2560,2560,0.010140444669458602
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,7168,32,0.0022151112142536375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,3072,1536,0.006966222491529252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,65536,3584,0.03393155667516921
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3072,8192,0.01181333346499337
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,2560,5120,0.008373333348168267
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,512,3584,0.015576000014940897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,5120,12288,0.018040888839297824
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,4096,65536,0.08850755294164021
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,1024,128,0.014483556151390076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1536,32,0.006630222416586346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,12288,2048,0.011462222370836469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,3072,4096,0.007670222057236566
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,10240,256,0.005960000058015187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,256,32,0.0021991110924217436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,512,8192,0.07598755757013957
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,256,2048,0.006167999986145232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,4096,65536,0.06726933187908597
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3584,32,0.0022168887986077205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,512,128,0.0022604444788561925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,4096,5120,0.008378666308191087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,65536,768,0.030078222354253132
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,7168,5120,0.01108000013563368
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,8192,1536,0.027990221977233887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,65536,6144,0.06818311081992255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,5120,1024,0.00905866672595342
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,12288,7168,0.022325333621766832
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,32,64,0.0018373332503769135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,6144,32,0.0025600000388092464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1536,2048,0.006529777828190062
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,16384,32,0.009760888914267222
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,12288,512,0.03664177656173706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,65536,768,0.015858666764365304
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,768,6144,0.017678222722477384
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,65536,3072,0.3033244344923231
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,7168,12288,0.06071022484037611
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,1536,4096,0.007927999728255803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,3584,4096,0.008990222381220924
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,3584,5120,0.009674666656388177
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3584,8192,0.016025778320100572
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,4096,768,0.006636444479227066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3584,128,0.00249777775671747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2560,2048,0.04714577727847629
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,7168,768,0.00739822205570009
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,12288,32,0.005595555735958948
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,4096,1536,0.031015111340416804
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,128,4096,0.016559999850061204
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,256,2560,0.014139556222491793
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,32,2560,0.005008888741334279
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,1536,3584,0.007695111135641734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,1536,768,0.005635555419656966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3072,12288,0.020417778028382193
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,512,12288,0.02572622232966953
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3584,256,0.008618666893906063
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,3584,1024,0.015966221690177917
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,512,12288,0.034527109728919134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,16384,128,0.007984889050324758
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,8192,16384,1.1971261766221788
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,3584,7168,0.02279111080699497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,128,1024,0.005655111124118169
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,65536,128,0.007095110913117726
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,6144,1024,0.006102222121424145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,64,256,0.005256888767083486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1024,7168,0.008009778128729926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,65536,2560,0.5116426679823134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,8192,768,0.00868977771864997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,32,256,0.0029128889242808023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,12288,2048,0.09258044428295559
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,512,3072,0.016869333055284288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3584,2048,0.012440889245933957
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,65536,10240,0.08020622200436063
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3584,6144,0.016933333542611863
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,2048,2048,0.006599999964237213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,16384,10240,0.9000942442152234
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2560,1024,0.004557333472702238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,128,128,0.008024888734022776
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,12288,5120,0.024784889486100938
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,4096,768,0.00630666646692488
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,12288,12288,0.03743288914362589
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,512,6144,0.022647110952271357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,16384,1024,0.028156446086035833
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,12288,5120,0.15355643961164686
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,768,2048,0.008628444539176093
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,64,6144,0.009282666775915358
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,3584,3584,0.026404443714353774
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1024,512,0.0029102222373088202
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,16384,1024,0.011415111521879831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,6144,7168,0.30218399895562065
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,2560,1024,0.00564533347884814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,5120,7168,0.03858755694495307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,1024,6144,0.012808000048001608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,16384,16384,0.08374844657050239
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,6144,1536,0.005728000154097875
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,1024,7168,0.010168888502650792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,7168,1536,0.006296000132958095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1024,16384,0.03160266743765937
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,64,3072,0.007158222297827403
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,10240,65536,0.19322755601671007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,16384,512,0.02723377777470483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,10240,1536,0.02235288918018341
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,6144,1536,0.006598222172922558
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,128,768,0.0028728888266616394
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,7168,5120,0.025782222549120586
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,8192,5120,0.03035555614365472
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,128,8192,0.010759111079904767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,5120,1536,0.006342222293217977
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,3584,1536,0.019876443677478366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,65536,768,0.030356443590588037
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,128,4096,0.00665511108107037
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3584,32,0.005730666634109285
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2048,65536,0.04332888788647122
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,10240,2560,0.03606933355331421
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,256,3072,0.014160000615649752
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,16384,64,0.005955555372767978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,8192,16384,0.04851999878883362
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,768,7168,0.009697777529557547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,65536,128,0.005902222047249476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2048,1536,0.0036880001425743103
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,3584,7168,0.012056888805495368
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,10240,32,0.002190222135848469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,768,12288,0.13139822747972277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,16384,32,0.002143111079931259
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,7168,128,0.03692977627118429
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,16384,12288,0.060016890366872154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,4096,32,0.007630222373538547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2560,128,0.0035128887328836652
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,512,2560,0.00693866651919153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,8192,4096,0.01553244392077128
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,16384,7168,0.17983733283148873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,3584,2560,0.006917333437336816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,16384,768,0.020223110914230347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,12288,65536,0.15741599930657282
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,768,512,0.002489777695801523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,65536,3584,0.16993510723114014
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,1024,65536,0.3364080058203803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,8192,768,0.04187022315131294
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,3072,2560,0.02064533366097344
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,10240,256,0.0226239992512597
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,5120,128,0.02990844514634874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,3584,10240,0.01834488908449809
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,768,16384,0.020212444994184706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,256,12288,0.014141332772043018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,6144,1536,0.008653333617581261
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2560,8192,0.05086044470469157
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,1024,10240,0.011777777638700275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,8192,12288,0.03527911172972785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,12288,65536,0.14868799845377603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,6144,8192,0.01706133286158244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3584,1024,0.005943111247486538
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,1536,128,0.014827556080288358
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,256,2048,0.01313333296113544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,2048,3072,0.007666666474607255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,128,128,0.006640888750553131
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,3584,2048,0.02439733346303304
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1536,65536,0.1200453307893541
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,65536,512,0.012618666721714867
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,3072,3584,0.008364444805516137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,65536,128,0.011065777805116443
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3072,8192,0.015415110521846347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,1536,4096,0.008323555191357931
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2048,12288,0.017090666625234816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,2560,512,0.014175999495718213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,8192,12288,0.06492177645365398
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,256,8192,0.01039644413524204
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,5120,1024,0.003567111161020067
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,2048,128,0.005392000079154968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,512,5120,0.016573333077960543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,4096,256,0.005264889034960005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,12288,256,0.006224000205596288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,8192,8192,0.016573333077960543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,8192,256,0.006002666635645761
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,512,16384,0.016167999969588388
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,128,65536,0.0459537771013048
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,1536,1536,0.016063110695944894
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,10240,32,0.01668711172209846
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,10240,32,0.006346666564544042
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,768,16384,0.029636445972654555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,12288,65536,0.16705421606699625
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,512,512,0.012534221841229333
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,10240,12288,0.36938044759962296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,4096,6144,0.04436888959672716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,65536,768,0.014200000299347771
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,16384,2560,0.01475822263293796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,768,1536,0.0063235556913746735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,10240,6144,0.04446755515204536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,16384,768,0.013589333328935834
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,1536,3584,0.01687733332316081
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,2048,256,0.020376000139448378
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,64,5120,0.00452888881166776
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,6144,1536,0.008310221963458592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,2048,65536,0.06412444512049358
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,10240,128,0.006331555545330048
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,2048,8192,0.010808889236715106
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,256,128,0.012856889102194043
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,256,10240,0.012137778103351593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,16384,2048,0.011795555551846823
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,1024,2560,0.015549333559142219
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2048,8192,0.014852444330851236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3584,65536,0.6015422079298232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1024,768,0.018647111124462552
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,4096,768,0.009024889104896123
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1536,3584,0.008700444466537898
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,2560,8192,0.030415111117892798
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,4096,65536,0.08924000130759345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,768,1024,0.0032275555034478507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2048,256,0.002266666748457485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,4096,128,0.0022391111900409064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,65536,7168,0.06623288657930163
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,512,5120,0.006698666761318843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,1536,512,0.014502222339312235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1024,4096,0.007641777396202087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,512,4096,0.007692444655630324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,2560,65536,0.06430577569537692
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,8192,2560,0.009045333498054082
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,16384,16384,0.04951022068659464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1024,6144,0.006917333437336816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,6144,65536,0.1130595604578654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,5120,768,0.005630222045713001
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,7168,2048,0.016574222180578444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,32,7168,0.009745777481132084
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,768,3072,0.007045333584149678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,7168,7168,0.04190222091144986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,256,1536,0.0042444442709287005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1024,4096,0.007254222200976477
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,16384,65536,0.32191644774542916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,512,65536,0.08198844724231295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2048,16384,0.0387368897596995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,16384,7168,0.09325866566763984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,6144,2560,0.008575999902354347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,65536,16384,0.5995413462320963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,1024,10240,0.02548800077703264
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3072,5120,0.013800889253616333
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,2048,12288,0.014902222487661572
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,7168,6144,0.04348444607522753
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,64,8192,0.010439111126793755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,16384,1536,0.16922667291429308
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,8192,65536,0.34107467863294816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,5120,16384,0.024990222520298425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,4096,2560,0.007486222518814935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2048,2048,0.025246222813924152
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,7168,1536,0.03842666745185852
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,256,2048,0.004679999831649992
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,65536,5120,0.11123911539713542
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,8192,3584,0.03092355529467265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,256,768,0.014714666538768344
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3584,32,0.002630222174856398
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,128,16384,0.016269332832760282
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3584,7168,0.07635288768344455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,128,7168,0.009912889036867354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,128,5120,0.005084444251325396
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,65536,64,0.011200889117187925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1536,5120,0.01038755559258991
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,32,3584,0.004635555462704765
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,12288,4096,0.03148355417781406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,6144,4096,0.014578666951921252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,3072,4096,0.06973244746526082
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,65536,7168,0.06556444697909884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,5120,3584,0.01465777721669939
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,256,65536,0.015785778562227886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,4096,1024,0.008450667063395182
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,64,16384,0.004912888838185204
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,256,128,0.00565155554148886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,5120,512,0.015953777564896476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,10240,4096,0.0181688881582684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,4096,65536,0.4728044403923883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,6144,65536,0.07190399699740939
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,12288,4096,0.014602666099866232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3584,32,0.0028888889484935333
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,8192,64,0.005582222094138463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,2560,6144,0.013103110922707452
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,3584,256,0.015782222151756287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,1536,16384,0.029879109727011785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,6144,2048,0.009966221948464712
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,2560,128,0.0052622221410274506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,4096,512,0.003491555651028951
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,4096,512,0.008255999949243333
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,4096,1024,0.016187555260128446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,1536,2048,0.015536889433860779
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,2560,2048,0.0069724445541699724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,128,1536,0.012291555603345236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,8192,2048,0.020776000287797716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,8192,64,0.0022239999638663395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,12288,768,0.0069653333889113525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2048,1024,0.0062871111763848206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,8192,512,0.00590133335855272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,10240,256,0.0031902222997612427
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,65536,256,0.1063199970457289
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2560,128,0.010348444183667501
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,512,128,0.0052728888889153796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,6144,1536,0.02514933380815718
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,1024,1536,0.006310222049554189
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,3584,3072,0.018177777528762817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,768,65536,0.5499528778923882
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,16384,7168,0.05283733208974203
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,8192,65536,3.0581234825981984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,512,2048,0.009025777379671732
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3072,12288,0.2704257700178358
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,16384,8192,0.04432177874777052
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,1536,768,0.012735111018021902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,4096,7168,0.024872889121373493
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,65536,2560,0.04264266623391045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,65536,8192,0.09259200096130371
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,768,16384,0.028776887390348647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3584,12288,0.028175999720891316
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,10240,2048,0.016531555189026725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,6144,2560,0.025920889443821375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,12288,6144,0.018379555808173288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,6144,4096,0.014525334040323893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,10240,512,0.006252444452709622
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,512,4096,0.007977777885066139
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,4096,128,0.002452444492114915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2048,1024,0.005942222144868638
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,256,2048,0.004223111189073986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,10240,6144,0.05487377776039971
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,16384,16384,0.07938310835096571
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,16384,7168,0.022651554809676275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,10240,1024,0.019955555597941082
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,10240,3072,0.03158311049143473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,10240,12288,0.03891111082500882
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,768,256,0.0028524444335036804
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1024,4096,0.007787555456161499
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3584,2048,0.0693146652645535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,8192,3072,0.010535111029942831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,6144,1536,0.008991111483838823
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,10240,6144,0.031516446007622614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,6144,7168,0.01904977858066559
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3584,128,0.0021626665774318907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,256,128,0.005264889034960005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,5120,65536,0.1315822203954061
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,7168,16384,0.03715644280115763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,7168,8192,0.03208622336387634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,7168,2048,0.010454222559928894
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,256,10240,0.01182755579551061
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,12288,10240,0.08364355564117432
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,128,512,0.006241777704821692
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,1024,10240,0.017640888690948486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,5120,6144,0.10941955778333877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,256,768,0.004049777984619141
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3072,5120,0.11756800280676948
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,8192,3584,0.009777777724795872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1536,6144,0.009424000150627559
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,1024,16384,0.30580088827345103
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,2048,4096,0.025827555192841425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2048,12288,0.03561866614553664
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3072,2048,0.006706666615274217
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,7168,768,0.007306666837798224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,3584,2048,0.01790844400723775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,64,16384,0.02957688768704732
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,32,6144,0.004522666749027041
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,3072,12288,0.016484444340070088
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,10240,256,0.017245334055688646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,12288,2048,0.12033243974049886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,128,1024,0.005941333456171884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3584,768,0.0032106666929192017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,10240,768,0.009167110754383935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,6144,8192,0.015264888604482016
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,64,256,0.002398222270939085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,10240,1536,0.018486221631368
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,10240,3584,0.12786577807532415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,6144,7168,0.01501066651609209
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,64,10240,0.005671999934646819
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,8192,3584,0.02439466615517934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,6144,10240,0.033542222446865506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,4096,128,0.015156444576051501
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,64,10240,0.01240355521440506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,2560,2048,0.01717866626050737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,3584,3584,0.015079999963442484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,256,128,0.0021804444905784395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,256,6144,0.016051555673281353
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,16384,3072,0.037560890118281044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,768,2560,0.008796444369686974
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,3072,7168,0.040581332312689886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3584,3072,0.0074942219588491656
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,65536,8192,0.1280639966328939
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,16384,1024,0.04349244303173489
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,1024,128,0.00546488869521353
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,128,1536,0.0063688887490166565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,768,8192,0.011149333582984077
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,128,7168,0.0053288886944452924
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2560,256,0.01292711082432005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,64,64,0.0019377778387731975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,768,1024,0.007359111474619971
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,768,8192,0.019800000720553927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,10240,6144,0.023790222075250413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,768,3072,0.005629333357016246
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,1024,1536,0.01421688828203413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3072,4096,0.010904889139864178
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,32,12288,0.02665866745842828
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,256,10240,0.005657777604129579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,768,6144,0.011062222222487131
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,4096,8192,0.01521511044767168
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,2048,8192,0.031327111853493586
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,256,64,0.005608888963858287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,8192,8192,0.07035111056433783
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,12288,3072,0.027685332629415724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,4096,2560,0.006973333656787872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,512,7168,0.0162444445821974
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,512,512,0.01316177762216992
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,256,5120,0.017343110508388944
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1536,256,0.005764444255166584
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1024,65536,0.05675911241107517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,8192,3072,0.010716444088353051
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,2560,10240,0.04105244411362542
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3584,3584,0.007736889024575551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,768,2048,0.006585777633719974
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,7168,5120,0.07999289035797119
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,128,5120,0.018332444959216647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,12288,768,0.01960444450378418
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,256,6144,0.009381333159075843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,768,16384,0.037463999456829496
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,10240,12288,0.03372888763745626
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1536,256,0.009699555734793345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,8192,768,0.01960266629854838
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,512,768,0.015119110544522604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,10240,768,0.00665155549844106
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,7168,5120,0.01347644461525811
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,128,2560,0.013103110922707452
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,12288,12288,0.043455110655890576
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,512,8192,0.01108088923825158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,5120,1024,0.016015999846988253
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,1536,6144,0.02028622229894002
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,3584,16384,0.05172622203826904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,5120,65536,0.12922044595082602
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,16384,2560,0.25299912028842503
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,12288,5120,0.01663466625743442
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3072,65536,0.06083377864625719
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2048,4096,0.00830666638082928
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3072,10240,0.015806222955385845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,7168,65536,0.13873867193857828
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,128,10240,0.011776888536082374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,3584,2048,0.00664444433318244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,8192,2048,0.013730666703648038
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,8192,10240,0.5041146808200413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,128,7168,0.009385777844323052
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,12288,4096,0.04224888814820183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3584,1024,0.018944000204404194
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,65536,4096,0.03851111067665948
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,1024,1536,0.015471999843915304
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,512,7168,0.01202133297920227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3072,65536,0.07170489099290636
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,768,5120,0.01884533299340142
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3584,3072,0.010452444354693094
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,512,768,0.006232000059551663
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,768,1024,0.0042071110672420925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,64,2048,0.004646222210592694
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,256,128,0.0052800000541739995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,12288,10240,0.030713776747385662
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,768,256,0.010807111031479306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,512,7168,0.008167111211352879
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,5120,5120,0.01145244472556644
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,10240,16384,0.12959110736846924
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2560,4096,0.11024710867140029
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,256,1536,0.00360000009338061
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2048,5120,0.010315555665228102
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2048,4096,0.00979288915793101
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,2560,3584,0.027802666028340656
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,128,8192,0.02238311039076911
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,4096,2560,0.010814222196737925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,7168,128,0.006610666712125142
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,2560,10240,0.031876444816589355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,2048,1024,0.006274666637182236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,16384,5120,0.08143199814690484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,10240,8192,0.04450577828619215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,4096,256,0.0035617777870761026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,256,5120,0.01588888963063558
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,12288,6144,0.021545777718226116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,5120,1536,0.01145866678820716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,12288,2560,0.034640888373057045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1024,10240,0.03094844354523553
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1536,64,0.005618666609128316
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3072,6144,0.026294221480687458
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,32,7168,0.010045333041085137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1024,2048,0.007497777541478475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,65536,7168,0.23211110962761772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,4096,1536,0.006212444355090459
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,64,16384,0.016951999730534024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1024,2048,0.017927999297777813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,6144,2048,0.007055111229419708
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,1024,128,0.011069333387745751
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,7168,4096,0.012404444317022959
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,16384,12288,0.271941343943278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,5120,128,0.028914666838116113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,16384,3584,0.01829600003030565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,10240,4096,0.014420444766680399
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,256,5120,0.02000711030430264
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,8192,10240,0.08339644140667385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,1536,1536,0.009713778065310584
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,10240,768,0.01826400061448415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1024,64,0.002199999988079071
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,32,8192,0.011921777493423887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,768,3072,0.007679999702506595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,3072,3584,0.022392888863881428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,2048,65536,0.41660888989766437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,32,8192,0.0041866666740841335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,3584,256,0.015474667151769003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,6144,128,0.0021537778278191886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,512,6144,0.017255110873116385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,65536,1536,0.07056978013780382
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,6144,3584,0.010116444693671333
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,768,10240,0.011092444260915121
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,64,16384,0.0055440002017551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,256,65536,0.9294889238145617
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2048,2048,0.006976889239417181
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1024,64,0.005567111074924469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,65536,32,0.003900444342030419
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,8192,2560,0.007607111500369177
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,128,2048,0.007338666253619724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,256,8192,0.018910222583346896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,1024,8192,0.010648888846238455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,2560,1024,0.007139555282062954
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,8192,8192,0.01923377811908722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,8192,4096,0.02663822140958574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,4096,128,0.002144888871245914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,3072,10240,0.026103999879625108
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,12288,3072,0.03820888863669501
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,256,65536,0.0506897767384847
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,1024,2048,0.006296000132958095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,7168,10240,0.01625866691271464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,4096,2560,0.02232444451914893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,7168,256,0.002502222235004107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,512,5120,0.009007111191749573
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,2048,7168,0.009657777845859528
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,2048,1024,0.005923555543025334
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,4096,3584,0.019366222951147292
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,4096,12288,0.01925066610177358
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3072,12288,0.018072000808186002
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,6144,2048,0.007335111498832703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,16384,7168,0.023019555542204116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,256,7168,0.010037333601050908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,6144,1024,0.006212444355090459
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,128,16384,0.016827555166350473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1536,2560,0.004940444396601783
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,4096,4096,0.029184887806574505
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,10240,8192,0.022242666946517095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,3072,6144,0.07757777637905545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,10240,3072,0.021183111601405676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,5120,1536,0.006271111054552927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,5120,10240,0.0362391107612186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,8192,128,0.005903999838564131
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,1536,6144,0.018681777848137748
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3584,6144,0.012234666281276278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,6144,768,0.0188719994492001
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,3584,1024,0.005896889087226655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2560,16384,0.019911999503771465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,3072,7168,0.009707555174827576
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,512,2560,0.006919999917348226
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,4096,12288,0.02050133380624983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,7168,3072,0.008007999923494127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,5120,3072,0.02849422229660882
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,3584,3072,0.01351999988158544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,128,3072,0.007318221860461765
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,256,3072,0.00488800017370118
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2560,512,0.006928000185224745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,128,512,0.012373333175977072
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,1024,2048,0.015751111838552687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,128,10240,0.019527110788557265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,1536,768,0.005621333503060871
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3072,768,0.006249777972698212
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3072,5120,0.14989777406056723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1536,8192,0.009724444813198512
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,8192,3584,0.011111111276679568
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,768,3584,0.016843555702103507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,16384,16384,0.09830311271879409
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,65536,64,0.004187555362780889
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,64,256,0.005943111247486538
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3584,32,0.0052355556852287715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,768,4096,0.010399999717871347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,65536,12288,0.2603413263956706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2560,7168,0.013032889200581444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,256,2048,0.004209777961174647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,65536,512,0.014145778285132514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,6144,64,0.0028799999919202593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3584,512,0.006924444602595435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,256,16384,0.015509333875444202
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,7168,5120,0.13500711652967665
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,64,12288,0.014135999812020196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,2560,5120,0.008560889297061497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,256,4096,0.0087333329849773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,12288,10240,0.07661688990063138
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,5120,4096,0.010213333699438307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,128,32,0.006276444428496891
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,6144,32,0.002509333400262727
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,10240,64,0.0159582214223014
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,3072,1024,0.015507555670208402
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2048,16384,0.018840889135996502
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,512,6144,0.009019555317031013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,128,5120,0.008667555948098501
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,64,16384,0.005896889087226655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,8192,65536,0.17553510930803087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,3072,6144,0.009645333720578088
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1024,2560,0.007304888632562425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,5120,2048,0.007137777904669444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,3072,3072,0.01772177716096242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,64,256,0.0021751111166344746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,32,1024,0.005555555638339784
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,1024,16384,0.05167644553714328
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,512,3584,0.01923288901646932
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1536,1024,0.0035386664999855887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,512,3584,0.015104000767072042
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,65536,32,0.047367999951044716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,12288,7168,0.048335111803478666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,64,16384,0.007279111279381647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,128,2560,0.006923555499977536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,3584,65536,0.12619111273023817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,4096,3072,0.009687110781669617
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,512,512,0.022509333160188463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,768,5120,0.015275556180212231
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,65536,12288,0.12501422564188638
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,5120,10240,0.017860444055663217
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,6144,768,0.00328177772462368
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,64,7168,0.005264889034960005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,32,32,0.0018835556175973679
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,8192,1536,0.02167111138502757
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2560,10240,0.017058667209413316
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1024,1024,0.0035031110876136353
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,4096,512,0.005557333429654439
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3584,3072,0.01735911104414198
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,4096,16384,0.03836266530884637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,4096,12288,0.019647111495335896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2048,7168,0.012420444852775998
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,10240,2560,0.029501332177056208
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2048,2560,0.006968888971540663
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,8192,1024,0.019944889677895438
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2560,128,0.0024862223201327855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,768,3584,0.01685599982738495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,256,3072,0.005562666803598404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,12288,65536,0.27027911610073513
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,128,768,0.005616888817813661
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2048,12288,0.015789333317014907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,4096,768,0.016907556189431083
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,12288,768,0.009951999617947472
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,3584,3072,0.018146667215559218
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,6144,2560,0.00758755538198683
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,128,12288,0.025422222084469263
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,2560,10240,0.10778222481409709
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2048,256,0.0025351111673646504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,65536,8192,0.14613866806030273
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,3584,16384,0.4794658025105794
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2048,512,0.0033031110134389666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,128,5120,0.008660444782839881
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,12288,8192,0.030498666895760432
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,128,64,0.0019475554840432273
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1024,16384,0.013045333325862885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,16384,3072,0.1623093287150065
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,256,2048,0.006614222294754452
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,5120,2560,0.027243556247817144
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,32,3584,0.007620444728268518
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,3072,16384,0.09720444679260254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,256,7168,0.013066666821638743
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,4096,16384,0.022633777724372015
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1536,2048,0.006586666736337874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,16384,2560,0.14134132862091064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,32,2048,0.0042035554846127825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,65536,2048,0.10332710875405206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,2048,512,0.013588444226317935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,32,64,0.0052373334765434265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,6144,65536,0.07196444272994995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,3584,2048,0.016805332567956712
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1536,5120,0.08062844143973456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,768,1536,0.021253334151373968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,6144,64,0.002152888932161861
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3072,1536,0.006233777850866318
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,16384,128,0.028413332170910303
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,4096,512,0.005623111294375525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,64,65536,0.0155404441886478
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2048,2048,0.008100444244013892
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,2048,16384,0.018578666779730055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,12288,2048,0.011436444189813403
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3072,64,0.005618666609128316
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,128,2560,0.007266666326257918
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,8192,6144,0.018214222457673814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,8192,2560,0.015184000134468079
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2048,64,0.015807110402319167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,6144,256,0.016567111015319824
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3072,8192,0.02596355477968852
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,1536,65536,0.07540088891983032
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,768,3584,0.007606222397751278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,6144,4096,0.029855112234751385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,12288,512,0.007287999822033777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,7168,16384,0.02827289038234287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,12288,3584,0.01312800000111262
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,128,512,0.0025155554629034465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,5120,4096,0.01368711143732071
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3584,8192,0.03174755639500088
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,256,128,0.005234666582610872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,8192,12288,0.020592000749376085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,2560,256,0.015202666322390238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,8192,128,0.002805333377586471
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,16384,65536,2.596827612982856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,512,12288,0.014827556080288358
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2560,10240,0.14000622431437174
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,6144,1024,0.009304888546466827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,6144,3072,0.008008889026112026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2048,6144,0.011416888899273343
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,4096,12288,0.05741244554519653
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,2560,128,0.0053226666318045715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,8192,1024,0.006292444550328785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,768,768,0.0035413333939181436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,4096,16384,0.04633955491913689
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,6144,3584,0.023929778072569106
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,64,3072,0.007280888656775157
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,8192,64,0.010004444254769219
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,128,1024,0.006255110932721033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,512,6144,0.017871111631393433
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2048,12288,0.017807111144065857
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,128,1024,0.003019555575317807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,128,12288,0.022992889086405437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,1024,1536,0.006255110932721033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,32,3072,0.0045422220395671
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,65536,512,0.014496889379289416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,64,7168,0.03442222211096022
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,10240,10240,0.03851200143496195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,1024,12288,0.01346577786737018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,5120,512,0.005656888915432825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,5120,65536,0.07008177704281278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,768,3072,0.008061333662933773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,768,4096,0.01680266691578759
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,16384,3072,0.01309333327743742
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,16384,8192,0.9172586864895291
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,10240,256,0.019896000623703003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1024,2560,0.0170168893204795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,768,128,0.010013333625263637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2560,1024,0.0032355555643637977
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,3072,4096,0.04487377736303541
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,6144,7168,0.015172445111804538
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,768,128,0.005973333285914526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2560,12288,0.018808888064490426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,8192,768,0.01756888959142897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,12288,7168,0.02401777770784166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2560,2560,0.007256888680987888
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,768,8192,0.012162666353914471
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,7168,5120,0.012048000262843238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,65536,2560,0.4716613557603624
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,3072,128,0.005956444475385878
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,128,8192,0.010383110907342698
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,12288,1536,0.00795999997191959
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,512,1024,0.00591111100382275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,10240,16384,1.8271698421902125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3584,1024,0.007641777396202087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,12288,10240,0.027608889672491286
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,256,2048,0.004911111046870549
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,3584,16384,0.043303112188975014
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,4096,6144,0.012538666526476542
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,7168,2048,0.024466666910383437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,8192,5120,0.014003554979960123
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,12288,5120,0.021205332544114854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,512,10240,0.012422222230169507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,7168,3072,0.013209777573744455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,2560,12288,0.015484443969196744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,256,256,0.002144888871245914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,12288,3584,0.13442311022016737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,16384,6144,0.04418755571047465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,128,1536,0.006168889088763132
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,128,8192,0.010034666293197209
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,6144,7168,0.2705262237124973
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,10240,7168,0.2187031110127767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2560,8192,0.01529600057337019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,6144,3584,0.008613333106040955
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3584,4096,0.067958222495185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,16384,256,0.08860177463955349
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3584,1024,0.006039999839332368
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,6144,16384,0.02496444516711765
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,8192,6144,0.033490667740503945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,8192,1024,0.014523555835088095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,5120,2560,0.0074035558435651995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,16384,64,0.003951999876234266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,256,2560,0.0218213333023919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,12288,128,0.0121760004096561
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,512,65536,0.08563911252551609
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,2048,5120,0.017629333668284945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,8192,7168,0.030304888884226482
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,1024,65536,0.20785866843329537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1536,65536,0.035269333256615534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,2560,3584,0.007641777396202087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,7168,2048,0.006970666348934174
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,8192,12288,0.02161600026819441
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,12288,3584,0.27201777034335667
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,8192,768,0.018806222412321303
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,3072,4096,0.008663111262851292
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3072,8192,0.1304453346464369
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1024,7168,0.008228444390826756
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,1536,128,0.005255110975768831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,3584,768,0.0058959999846087555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1024,16384,0.011729777687125735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,65536,1024,0.06291377544403076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,10240,128,0.005961777849329843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,6144,16384,0.02312622302108341
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,12288,1536,0.009113777842786577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,32,64,0.006616888774765863
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1536,3584,0.007638221813572778
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,32,2560,0.004854222138722737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,768,3584,0.019742222295867074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,16384,65536,0.5264639854431152
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,256,10240,0.02167822255028619
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,5120,10240,0.3777626620398627
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,5120,1024,0.003269333392381668
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,1536,65536,0.0827466646830241
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,512,128,0.005592000153329637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3072,8192,0.014649778604507446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,12288,256,0.01774133410718706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,256,4096,0.007667555577225155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,512,12288,0.007329777710967594
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,7168,16384,0.054470221201578774
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,12288,3072,0.04500800040033129
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,65536,5120,0.10782488849427964
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,6144,3072,0.011220444407727985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3072,768,0.003379555626047982
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,8192,2048,0.008250666989220513
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,1024,10240,0.031786666976081
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3072,65536,0.06336088975270589
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,3584,512,0.005559111220969095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,6144,256,0.009406222237481011
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,1024,1024,0.015477332803938123
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,16384,2048,0.017285333739386666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,10240,65536,0.14258489343855116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,8192,8192,0.02309688925743103
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,8192,12288,0.020217777954207528
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,768,1536,0.006238222122192383
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,2048,65536,0.07612978087531196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,7168,32,0.010051555931568146
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,768,768,0.005934222290913264
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,8192,8192,0.022118222382333543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,64,1536,0.0035982223020659555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,10240,2560,0.029556443293889362
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,65536,16384,0.7854248682657877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,12288,8192,0.02331111166212294
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,768,8192,0.08628088898128933
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,7168,7168,0.11928444438510472
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,7168,3072,0.08106755548053317
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,12288,1024,0.02478666603565216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,12288,5120,0.019496000475353666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,65536,768,0.035419556829664446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,1024,10240,0.014542222023010254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,8192,4096,0.01957244508796268
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,10240,7168,0.07374222411049737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,8192,3584,0.01309333327743742
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,512,6144,0.0173653331067827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,1024,3584,0.07772266864776611
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,2560,2560,0.006991110742092133
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2048,32,0.006294222341643439
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,7168,1536,0.027492443720499676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1024,768,0.0032044444233179092
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,2560,768,0.005580444302823808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,1024,2048,0.00666933341158761
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,128,128,0.002151111140847206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,10240,3584,0.034408890538745456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3072,8192,0.020023110840055678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,1536,768,0.0062977779242727495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,3584,3072,0.013399110900031196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,12288,4096,0.014815110299322339
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,5120,65536,0.1158942249086168
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,16384,1024,0.00906488878859414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,768,12288,0.013441777891582914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,128,64,0.006625777731339137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,4096,8192,0.013217777841620974
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,65536,1024,0.015692444311247934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,64,5120,0.004943111290534337
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,6144,6144,0.03379111157523261
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,1024,3072,0.026146666871176824
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,2560,6144,0.009007111191749573
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,7168,7168,0.030235555436876085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,6144,2560,0.05854577488369412
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,64,10240,0.012200888660218982
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,8192,8192,0.03474311033884684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,3584,5120,0.031234665049446955
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,1024,12288,0.013515555196338229
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2048,64,0.0025555555605226094
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,768,5120,0.00833511104186376
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,1024,16384,0.016228444046444364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,768,768,0.0028808888875775864
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,128,32,0.0022035555707083806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,128,12288,0.005320888840489917
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,4096,3584,0.05908533599641588
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,7168,1536,0.0277804434299469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,5120,64,0.0022106667359670005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,5120,7168,0.035311109489864774
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,5120,12288,0.019080888893869188
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,1536,512,0.006000888844331105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,7168,512,0.004279110994603899
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,3072,3072,0.010646222366227044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,5120,7168,0.02763644523090786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,2048,768,0.009544000029563904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,4096,16384,0.022410665949185688
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,16384,10240,0.13663910494910345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,1024,2048,0.016194666425387066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,7168,1024,0.005943999936183293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,768,3584,0.008008889026112026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,2048,7168,0.021252445048756067
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,768,65536,0.08623733123143514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,6144,3584,0.10548355844285752
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,12288,768,0.006659555352396435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,128,4096,0.01313066648112403
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,128,3072,0.004604444321658877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,65536,768,0.02349511120054457
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,65536,256,0.012439111040698158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,32,3584,0.0076657773719893555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,1024,3072,0.017100445098347135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,4096,4096,0.011164444188276926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,65536,64,0.16417511304219565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3072,256,0.0033208889265855155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1536,12288,0.0177502218219969
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1024,2048,0.006634666687912411
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,5120,12288,0.10588533348507351
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,64,6144,0.009356444080670675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3072,10240,0.026292444931136236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,1024,2560,0.018505778577592637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3584,32,0.0022035555707083806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,2560,256,0.0052675555149714155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1024,3584,0.0068604445291890045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,768,2560,0.007058666812049017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,512,256,0.0053137776752312975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,3072,10240,0.014765333798196582
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,2048,12288,0.05167466733190748
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,16384,2560,0.014663111832406787
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,128,10240,0.03200711144341363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,10240,16384,0.05336622065967984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3072,32,0.005603555589914322
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,128,128,0.002209777840309673
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,7168,12288,0.02018222212791443
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,32,16384,0.0053671110007498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,12288,10240,0.027815110153622095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,16384,2048,0.03830488853984409
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,4096,8192,0.015481778317027621
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,512,128,0.013830222189426422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,12288,3584,0.015971556305885315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3072,512,0.005979555762476391
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2560,256,0.003212444484233856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,12288,1536,0.024346666203604803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,7168,5120,0.01352266636159685
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,7168,10240,0.0299253331290351
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2048,1024,0.003206222214632564
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1536,3072,0.007335999773608313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,10240,64,0.006680000159475539
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,16384,768,0.01313244468635983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,512,3584,0.00796622203456031
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,7168,256,0.017921777235137094
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,5120,3072,0.02254399988386366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,4096,768,0.016516443755891588
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,4096,3072,0.007518221934636434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,5120,10240,0.05312977896796333
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,7168,3072,0.022639110684394836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,64,2048,0.004932444542646408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,7168,16384,0.027102222045262653
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,256,512,0.0059573331640826324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3584,128,0.00221422231859631
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3584,10240,0.016340444485346477
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,16384,2560,0.014470222923490735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,6144,3584,0.02270488937695821
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,10240,8192,0.05372888843218485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,7168,768,0.006272888845867581
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,512,512,0.0069857777820693115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,128,1536,0.006009777800904379
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,1024,5120,0.016897777716318767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,128,1024,0.007970666719807519
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,65536,128,0.010130667024188572
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3072,768,0.0058871110280354815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,65536,512,0.03547911180390252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,65536,4096,0.051338666015201144
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,7168,16384,0.037308444579442344
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1024,3072,0.007340444458855524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,768,64,0.005254222287072076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,65536,3584,0.6060578028361002
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,16384,12288,0.08424088690016006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,256,1024,0.006640000061856375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,6144,512,0.0471111105548011
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,5120,7168,0.01976355579164293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,5120,12288,0.029902223083708022
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,3584,2560,0.0247706671555837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,65536,128,0.019650666250122916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3072,2048,0.0052639999323421055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,12288,65536,0.15647733211517334
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,2560,4096,0.019310222731696237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,512,3584,0.006308444258239534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2048,10240,0.015378667248619927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,64,10240,0.021529777182473078
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,128,512,0.002537777854336633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,12288,3584,0.028493334849675495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,512,64,0.005365333209435145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,5120,3584,0.023452444208992854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,16384,7168,0.02362488872475094
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3072,3584,0.007691555553012424
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1536,12288,0.018668444620238412
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,12288,1536,0.00907555553648207
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,3584,512,0.014716444744004143
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,10240,2048,0.044884443283081055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,512,6144,0.01033155537313885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,3584,8192,0.012390221986505719
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,7168,3072,0.00885511106914944
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,4096,16384,0.02305511136849721
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1024,16384,0.01351999988158544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,2560,128,0.014469333820872836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3584,10240,0.017645332548353408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,12288,128,0.01552977826860216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,128,4096,0.00831733312871721
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1024,8192,0.009944000177913243
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,1536,1536,0.014905777242448596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,5120,16384,0.029847111966874864
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,7168,7168,0.019555555449591745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,3072,768,0.018239999810854595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,2048,7168,0.009661333428488838
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,64,128,0.00629688882165485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,3072,7168,0.030573334958818223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1536,10240,0.012486221889654795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,65536,512,0.0470071103837755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,4096,32,0.0021848889688650766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,5120,65536,0.20747821860843232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,128,6144,0.00591111100382275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,5120,1536,0.0047688889834615915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,65536,128,0.009764444496896531
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,4096,5120,0.012069332930776807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2560,2560,0.005478222337034013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1024,64,0.005928888916969299
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,512,128,0.013778666655222574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,32,64,0.0018399999373488957
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,2048,10240,0.011722666521867117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,7168,5120,0.025773333178626165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,4096,128,0.002149333349532551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,128,512,0.01071111112833023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,5120,3072,0.009474666582213508
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,128,256,0.010112889111042023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2048,12288,0.018894222047593858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1536,2048,0.007225777539942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,7168,8192,0.3825715647803412
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,4096,2048,0.01581333412064446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,3072,3584,0.01685333251953125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,16384,12288,0.06514488988452487
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1536,4096,0.008051555189821456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,512,32,0.002504000026318762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,65536,32,0.016688000824716356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1536,2560,0.006981333096822103
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,16384,3584,0.021270222134060327
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,4096,10240,0.016927111479971144
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,16384,1024,0.007670222057236566
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3072,12288,0.019294222195943195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,16384,3584,0.02714755634466807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,12288,10240,0.045313778850767344
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,8192,16384,0.3261137803395589
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,16384,7168,0.02218222286966112
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,16384,65536,6.86495123969184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,512,6144,0.024886222349272832
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,256,8192,0.013272888958454132
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,16384,1536,0.029587556918462116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,16384,7168,0.021892444954978094
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,5120,2048,0.006268444574541516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2560,2560,0.0069671107663048645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,128,128,0.008029333419269985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,2048,2560,0.00741244438621733
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,7168,32,0.00378933341966735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,10240,32,0.0069226668112807805
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,512,10240,0.006624889042642381
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,256,2560,0.007261333366235097
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,16384,3584,0.020917332834667627
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,8192,512,0.006146666490369373
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,7168,768,0.009594666461149851
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,512,512,0.004031110968854693
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,128,4096,0.006752000086837345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,7168,3584,0.011415111521879831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1024,1024,0.006320000108745363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3072,16384,0.019091554813914828
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,768,256,0.014863111906581454
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,12288,3072,0.3398586644066705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,8192,1536,0.006050666587220297
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,8192,1536,0.02712000078625149
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,4096,512,0.0057297779454125305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,2560,5120,0.14339378145005968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,16384,8192,0.058541331026289195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,5120,256,0.01662133302953508
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,16384,128,0.02819733487235175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3584,5120,0.011703111231327057
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,64,2560,0.005361777626805835
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,65536,5120,0.12622311380174425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,2560,4096,0.011801777614487542
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2048,7168,0.0192275560564465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,4096,10240,0.033655110332700945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2048,1024,0.00629688882165485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3584,64,0.005331555588377847
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,8192,10240,0.045237332582473755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,64,256,0.00555644432703654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,6144,12288,0.03631110986073812
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,12288,65536,0.23882844712999132
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,3072,3584,0.019673777951134574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1536,2048,0.006607999818192587
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,1536,12288,0.018583999739752877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,12288,16384,0.06177688969506157
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,65536,5120,0.10096889071994358
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,1536,3584,0.007638221813572778
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,3072,2560,0.01684533390733931
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,64,1024,0.0028693332440323303
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,12288,1024,0.007623111208279927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,6144,5120,0.025085333320829604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,4096,10240,0.015255111787054272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1024,2560,0.033424887392256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,4096,3584,0.10482577482859294
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,16384,6144,0.0658079981803894
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,16384,10240,0.0702106687757704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,2048,4096,0.02683999968899621
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,8192,128,0.005933333188295364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1536,1536,0.0042275554604000514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1536,1024,0.004967111266321606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,16384,2048,0.20758667257097033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,768,65536,0.07693688737021552
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3072,2560,0.007274666594134436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,6144,2048,0.022639999787012737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,16384,3072,0.04202222161822849
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,64,65536,0.023843554986847773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,8192,3072,0.01068800025516086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,512,8192,0.07649244202507867
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,7168,4096,0.025635555386543274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2048,1536,0.006593777901596493
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,6144,3072,0.016639111770523917
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1024,256,0.006025777922736273
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,128,256,0.0021520000365045336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2048,5120,0.008627555436558193
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,7168,12288,0.2637217839558919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,16384,1024,0.0074728892909155945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,768,2560,0.007917332980367873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,10240,2048,0.011015111373530494
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2048,128,0.0021395554973019496
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2560,8192,0.01494311127397749
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,8192,5120,0.03657599952485826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2048,256,0.00239644447962443
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,128,3584,0.005929777605666055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,10240,7168,0.04873866505093045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,65536,256,0.011122666299343109
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,3584,65536,0.06815822256935967
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,6144,3584,0.013608888619475894
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,12288,16384,0.043269332912233144
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,768,3584,0.006014222072230444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3072,3072,0.006986666884687211
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,12288,512,0.007289778027269576
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,768,128,0.014683556225564746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,128,2560,0.004878222114510006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,64,12288,0.032925334241655134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,128,65536,0.0582853356997172
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,6144,32,0.003879111260175705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,12288,2048,0.009351111120647853
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,6144,8192,0.014478221535682678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,768,512,0.005714666512277391
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,2560,5120,0.01906399925549825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,256,3584,0.007721777591440413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,4096,4096,0.009387555221716562
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,7168,32,0.006059555543793573
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,4096,7168,0.011375111010339526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3584,10240,0.014923555983437432
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,6144,128,0.007264888948864407
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3072,4096,0.00870488915178511
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3584,2048,0.05963910950554741
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,4096,3584,0.0217902229891883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,256,2048,0.019673777951134574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,7168,3072,0.009112888740168678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,5120,16384,0.6085768805609809
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,10240,2560,0.09463555283016628
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,1536,16384,0.024688889582951862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,32,768,0.002895999906791581
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,32,768,0.00555377784702513
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,4096,1536,0.020290666156344943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,8192,8192,0.024676443801985845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,7168,16384,0.039911998642815485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,128,65536,0.04904177784919739
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,128,3072,0.007290666302045186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,3072,7168,0.027448889282014634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1536,3072,0.005631111148330901
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,4096,7168,0.036723554134368896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,1536,512,0.013092444174819522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3072,1024,0.01351111133893331
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,256,16384,0.016220443778567843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,10240,8192,0.03827288746833801
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,64,6144,0.0052737775776121355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3584,4096,0.013585777746306526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,768,768,0.002874666617976295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,12288,2048,0.17069244384765625
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,256,8192,0.010337777435779572
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,10240,10240,0.03577866819169786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2048,3072,0.007991111112965478
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,7168,16384,0.0451182226339976
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,12288,8192,0.04469866553942362
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,10240,7168,0.026941332552168105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3072,3072,0.009085333181752099
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,768,256,0.015113777584499784
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,512,64,0.0028417776856157514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2560,12288,0.014884443746672736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,7168,768,0.006699555450015598
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,5120,2048,0.006931555353932911
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,5120,16384,0.09994755850897895
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,4096,128,0.00591466658645206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,65536,128,0.010421333213647207
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2048,256,0.005882666756709416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,8192,10240,0.022368000613318548
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,3584,1024,0.005948444621430502
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,3584,1024,0.017529777354664274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3072,512,0.0058906666106647914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,512,2560,0.009999111294746399
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,256,4096,0.025954667064878676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,6144,256,0.00867911097076204
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,6144,256,0.0025235555238193935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1024,32,0.002181333386235767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,10240,16384,0.15138488345676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,768,32,0.0031902222997612427
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,2560,8192,0.012724444270133972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,256,10240,0.005290666802061929
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,12288,512,0.020560888780487906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,8192,16384,0.07519821988211738
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,256,65536,0.08506844441095988
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1024,10240,0.025795555777019922
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1024,12288,0.00998044427898195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,16384,12288,0.06151555644141304
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3584,128,0.0021475555582178966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,8192,6144,0.019326221611764696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,256,768,0.0028320000403457214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,2048,128,0.005534222142563925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,32,3584,0.01108088923825158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,5120,2560,0.009479111267460717
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,256,7168,0.00999022192425198
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,10240,1024,0.0169831116994222
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1024,3584,0.007657777931955125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,128,2048,0.006959111326270633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3072,512,0.005559999909665849
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,12288,32,0.010716444088353051
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,128,3584,0.014142221874660917
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3584,12288,0.01890222231547038
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,3584,65536,0.2644631067911784
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,256,64,0.005660444498062134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,7168,768,0.005937777873542573
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,32,3072,0.004859555512666702
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,7168,10240,0.03888355692227682
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,10240,10240,0.02932000160217285
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,64,16384,0.1345813274383545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1536,10240,0.01221866657336553
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,2048,2560,0.007385777930418651
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,3072,512,0.009199110998047723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,8192,3072,0.009770666559537252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,256,768,0.014859555496109856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,7168,256,0.0057502223385704895
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,65536,16384,0.15988177723354763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,128,768,0.002879111096262932
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,12288,8192,0.02341777748531765
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,768,16384,0.015483554866578845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,65536,128,0.012940444052219391
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,32,768,0.0074933336840735535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,512,1024,0.0034053333931499054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,8192,7168,0.19607910845014784
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,768,2560,0.015225778023401896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1024,2048,0.0069102222720781965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,512,1536,0.004229333251714706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,1536,3072,0.007294221884674496
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,8192,2048,0.025425778494940862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2560,10240,0.01550044450494978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,768,12288,0.009846222069528367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,256,3584,0.0073315559162033936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,7168,65536,0.39666133456759983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,16384,3584,0.0476817786693573
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,5120,12288,0.03725511166784499
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,1536,5120,0.009723555710580613
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,128,5120,0.00490133340160052
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1024,128,0.005573333137565189
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,10240,1536,0.02054222259256575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,6144,12288,0.22926667001512316
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,1024,256,0.0155102229780621
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,512,7168,0.009703111317422655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,64,1024,0.0031831111345026228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,4096,1536,0.029501332177056208
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,7168,10240,0.05427822139528063
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,16384,256,0.006299555715587404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,512,1024,0.0058853332367208265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,32,512,0.002551111082235972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,12288,768,0.019263111882739596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,10240,1536,0.008023111356629265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2560,4096,0.011031111081441244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,12288,10240,0.0359919998380873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,32,3072,0.007288888924651676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,3584,16384,0.025384889708624944
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,768,3584,0.007975111405054728
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,6144,10240,0.06203555398517185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2560,32,0.0053031109273433685
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,1536,6144,0.009332444104883406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,10240,64,0.0021315556433465746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,12288,16384,0.07994666364457872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,4096,7168,0.011858666936556498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,3584,5120,0.021357332666714985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,512,16384,0.008572444319725037
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,128,10240,0.019945777124828763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,512,6144,0.017960889471901786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,65536,2048,0.024001777172088623
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,512,3072,0.012055111428101858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,64,3584,0.008012444608741337
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,10240,12288,0.05600800116856893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,128,256,0.011772444678677453
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,3584,2560,0.007003555695215861
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,65536,10240,0.13108444213867188
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,64,2560,0.004249777644872665
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,768,256,0.006674666785531574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,128,256,0.005433777968088786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,64,32,0.0063493334584765965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,2048,10240,0.04600977897644043
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,256,512,0.005632888939645555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,256,12288,0.10214400291442871
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,128,1536,0.005992888990375731
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,256,5120,0.00846400029129452
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,1536,1536,0.02014755540423923
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,6144,3584,0.010258666343159145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,12288,128,0.02032800018787384
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,6144,10240,0.12166044447157119
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1024,256,0.0027226667023367355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,12288,4096,0.48313246832953555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,12288,64,0.033369776275422834
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,4096,1024,0.0063893331421746155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,2048,3072,0.015060444672902426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,32,2560,0.028196444114049275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,8192,12288,0.03197155396143595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,10240,1536,0.008653333617581261
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,10240,512,0.010878221856223213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,768,2560,0.008004444340864817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,6144,16384,0.04477599925465054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,7168,12288,0.03826755616399977
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,10240,65536,0.14745422204335532
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,4096,8192,0.014612444572978549
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,10240,1024,0.020917332834667627
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,768,65536,0.02384088933467865
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3584,512,0.0029644444584846497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,10240,3072,0.02570044497648875
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,128,3584,0.016177778442700703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,6144,512,0.00563822231358952
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,10240,768,0.006948444578382704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,6144,8192,0.019819556011093985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,256,512,0.011454222102959951
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,12288,512,0.043641779157850474
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1024,2560,0.005700444595681296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,512,512,0.007651555869314406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,3584,768,0.00565155554148886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,768,3584,0.017330666383107502
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,32,16384,0.01677600045998891
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,5120,1024,0.005994666781690385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,768,256,0.016572443975342643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3072,2560,0.02214488883813222
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,16384,16384,0.35059555371602374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,2048,10240,0.05493244528770447
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,2560,1024,0.005668444352017508
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,8192,2048,0.007635555333561367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,65536,3072,0.1498862240049574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2560,32,0.0021679999513758556
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,512,128,0.005948444621430502
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,128,512,0.005594666633341048
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,10240,2048,0.08057066467073229
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1536,256,0.0033164444482988785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,2048,128,0.005493333356248007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,32,4096,0.008007999923494127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,65536,3584,0.040983110666275024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,5120,3584,0.012121777567598553
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,6144,768,0.0034435554924938413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1024,1536,0.008439111212889353
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,16384,768,0.006435555716355641
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,768,128,0.004986666556861666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,4096,32,0.005325333525737126
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,64,3072,0.004603555632962121
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,2048,6144,0.018957333432303537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,16384,128,0.003522666792074839
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,2048,65536,0.09104622072643703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3072,6144,0.049508445792728
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,64,512,0.00564177789621883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,7168,3584,0.020044444335831534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,3584,65536,0.10284533103307088
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,64,64,0.0062977779242727495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,512,4096,0.01591911084122128
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2560,16384,0.017511111166742113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,2560,1024,0.006234666539563074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,8192,4096,0.010288000106811523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,7168,128,0.00563733321097162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3072,12288,0.01643822259373135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,512,6144,0.009305777649084726
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,512,256,0.014918223023414612
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,768,2048,0.01519733336236742
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,512,8192,0.04818844464090136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3072,8192,0.022701332966486614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,4096,5120,0.01072355525361167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,512,5120,0.016562667157914903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,4096,3072,0.030729777283138696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,128,10240,0.006031999985376994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,16384,128,0.015215999550289579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3072,128,0.003896000070704354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3584,2048,0.005209777918126848
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,65536,4096,4.856335110134549
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,10240,65536,0.15210755666097006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,12288,3584,0.03211911188231574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,256,256,0.002528000002106031
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,7168,768,0.0039022221333450745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,512,1536,0.00425155543618732
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,65536,1536,0.0722000002861023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1024,64,0.005587555468082428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,65536,2048,0.030418667528364394
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,3072,12288,0.01664177742269304
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,128,256,0.002141333288616604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,12288,7168,0.7899404631720649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,512,512,0.012400888734393649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,512,768,0.00554666668176651
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,12288,2048,0.010445333189434476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,4096,65536,0.0685911112361484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,64,1536,0.009992000129487779
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,64,2560,0.010110222631030613
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1024,65536,0.025417778227064345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,64,2048,0.003858666867017746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,7168,5120,0.03565866748491923
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,768,4096,0.007947555846638149
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,8192,10240,0.021714667479197185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,32,2048,0.006578666468461354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,512,6144,0.010390222072601318
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,3072,3072,0.043100443151262074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3072,128,0.0052728888889153796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,64,32,0.0018577778504954444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3584,4096,0.008877333667543199
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,65536,6144,0.12147022618187799
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,7168,512,0.01862133377128177
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,10240,65536,1.3468444612291124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,6144,2560,0.03336888882848952
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,512,2560,0.014873777826627096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,512,4096,0.012808889150619507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,2560,6144,0.00904266701804267
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3072,7168,0.05774222479926216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,128,512,0.002598222138153182
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3584,256,0.012087110843923358
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,512,65536,0.12666577763027614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,128,4096,0.00832444429397583
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,128,3072,0.007231111327807109
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,6144,16384,0.025776000486479864
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,4096,16384,0.022270222504933674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,2048,768,0.01866666641500261
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,768,32,0.002285333350300789
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,32,3072,0.01074577785200543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,5120,8192,0.07509777943293254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,64,16384,0.006762666834725275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3072,1024,0.02012888921631707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,128,6144,0.009344889058007134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1024,256,0.008350222474998897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,7168,2048,0.006979555719428592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,12288,5120,0.03821422325240241
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,768,1024,0.006259555617968242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,7168,8192,0.04866222209400601
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,7168,5120,0.012139555480745105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2560,3584,0.009526222116417354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,768,32,0.006263999889294307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,1024,16384,0.021942221456103857
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,8192,8192,0.04578311244646708
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,7168,8192,0.021910222040282354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,4096,6144,0.026534222894244727
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,768,3584,0.0057653333577844836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,6144,768,0.005932444499598608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,2560,768,0.005623111294375525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,3072,2048,0.007333333293596904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,6144,7168,0.0430391099717882
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,128,1024,0.008991111483838823
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2560,7168,0.017716444200939603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,2048,2560,0.020280000236299302
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,16384,256,0.006340444501903322
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,4096,10240,0.018429333964983623
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,12288,128,0.005619555711746216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,6144,512,0.004229333251714706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,65536,3072,1.9586088392469618
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,32,65536,0.02751022246148851
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,1024,16384,0.04048177931043837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,4096,768,0.00555644432703654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,12288,2560,0.03601510988341437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,3072,2048,0.015254222684436373
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,768,512,0.008623999853928884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,12288,8192,0.026280000805854797
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,12288,128,0.064410666624705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,64,8192,0.00545155546731419
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2048,32,0.003160888950030009
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,128,7168,0.01036177741156684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,16384,6144,0.02037244372897678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,256,10240,0.020636444290479023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,4096,12288,0.030446221431096394
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,32,768,0.0029155556112527847
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3072,1024,0.006676444576846228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,1024,3584,0.007651555869314406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2560,3584,0.008038221961922115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,5120,3072,0.007415999968846639
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,768,3072,0.01701955497264862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,10240,12288,0.0767351124021742
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3584,768,0.01568533314598931
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,128,128,0.01013422260681788
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2048,32,0.005393777870469623
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,65536,2560,0.084133333630032
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,5120,6144,0.026120889517996047
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,128,16384,0.01586933268441094
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,128,8192,0.011071999867757162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,128,8192,0.012552888856993781
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,2560,2048,0.006695111178689533
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,1024,2048,0.006673777682913675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,5120,8192,0.01551555593808492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,3072,65536,0.08560178014967178
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2560,2048,0.0070088886552386815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1024,7168,0.045273777511384755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,5120,12288,0.04943644338183933
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,1024,12288,0.025545777546034917
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3072,7168,0.017788444956143696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,3072,3584,0.007559110720952352
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,6144,10240,0.15594844023386636
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1536,2560,0.007680888805124495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,12288,3072,0.013826666606797112
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,7168,128,0.006996444529957241
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,5120,768,0.00795377790927887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,16384,65536,0.1683057811525133
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,512,32,0.002199999988079071
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,8192,5120,0.01183377785815133
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,5120,3072,0.008378666308191087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,768,2560,0.007385777930418651
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,12288,5120,0.01886311173439026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,64,128,0.005311111195219888
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,2048,256,0.0053226666318045715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,6144,8192,0.31997156143188477
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,6144,4096,0.01626666718059116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,128,3584,0.01555288831392924
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,3072,128,0.005615111026499007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,10240,512,0.008079110748238033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,8192,7168,0.01955200069480472
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3072,16384,0.03892711136076186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2560,768,0.005949333310127258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,16384,4096,0.036320000886917114
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,8192,10240,0.05319999986224704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1024,64,0.0021671110557185277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,7168,7168,0.17980089452531603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,512,12288,0.010057777994208867
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3072,10240,0.032687111033333674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,64,8192,0.01990133358372582
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,5120,512,0.004440000073777305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1024,5120,0.05585155884424845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,65536,3072,0.0651528901524014
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,128,256,0.0024862223201327855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,128,3072,0.0348693331082662
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,10240,7168,0.051798221137788564
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3072,6144,0.011151110960377587
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,2560,4096,0.014553777045673795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2048,2048,0.004518222063779831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,5120,10240,0.33117156558566624
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,3584,10240,0.028560890091790095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,1024,4096,0.016536000702116225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2048,6144,0.010488000180986194
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,64,64,0.001812444378932317
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,8192,1536,0.012073777616024017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,12288,5120,0.02606577840116289
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,768,12288,0.009930666950013902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,256,16384,0.016160888804329764
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,128,2560,0.00738844441043006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,768,65536,0.02256800068749322
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,512,7168,0.007059555914666917
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,256,1536,0.006321777900060018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2048,10240,0.010330666270520952
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,16384,3584,0.04585955540339152
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,3072,5120,0.06574133369657728
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,16384,12288,0.5655591222974989
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,1536,2048,0.015802666544914246
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,128,2560,0.004511111312442356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,768,32,0.006629333313968446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,10240,12288,0.047084444099002414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2560,2048,0.009789333575301701
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,16384,7168,0.04569955666859945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2048,16384,0.015481778317027621
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,768,7168,0.007360888852013483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3072,8192,0.012096000214417776
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,16384,7168,0.04151644309361776
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,64,2560,0.007265778051482306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2048,32,0.0025840000145965153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,8192,2048,0.008057778080304464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1024,32,0.002208000048995018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,16384,8192,0.09076533052656387
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,2560,256,0.005379555539952383
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,128,768,0.009327111144860586
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,1024,1024,0.005986666513813867
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,1024,2560,0.011094222466150919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,1024,2560,0.016978666186332703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,3072,256,0.018301332990328472
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3072,10240,0.0166631109184689
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,6144,64,0.0029013332807355454
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1024,12288,0.015132443772421943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,4096,8192,0.013763555222087435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,64,512,0.002520888836847411
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,4096,8192,0.021002666817771062
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,16384,12288,3.341915554470486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3584,1024,0.0036595554815398324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,128,768,0.010046222143703038
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3072,3072,0.014876445134480795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,4096,32,0.014858666393491956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,8192,65536,0.22156266371409097
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,32,32,0.0022115556316243275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,65536,16384,0.49875466028849286
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,768,3584,0.006973333656787872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,16384,7168,2.254121780395508
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,256,12288,0.05350311266051399
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2048,8192,0.019133332702848647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1536,256,0.003930666794379552
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,32,6144,0.00812177773978975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2560,65536,0.061331556902991406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,3072,128,0.005466666486528184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,256,128,0.0022577777918842104
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2048,32,0.005592888842026393
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,512,7168,0.012240000069141388
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,1024,512,0.013078221844302284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,768,1536,0.006329777754015393
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,10240,128,0.00258577780591117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,4096,10240,0.01780533293883006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3072,1536,0.006375999914275275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,12288,1536,0.07189689079920451
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,32,64,0.006258666515350342
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,32,256,0.008005333443482717
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,12288,768,0.024068444967269897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,6144,256,0.005590222362014983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,256,4096,0.008281778130266402
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,4096,65536,0.07754577530754937
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,3072,128,0.005614222337802251
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,8192,1024,0.006281777802440856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,1024,3584,0.01588622232278188
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3072,1536,0.003960888832807541
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,5120,64,0.01276177747382058
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,7168,2048,0.02143200072977278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,10240,10240,0.025641777449183997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1536,16384,0.013761777844693927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,32,128,0.006601777755551868
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,256,128,0.0021555556191338436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,6144,768,0.006287999865081575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,7168,8192,0.018256000346607633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,7168,6144,0.012977778083748288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,12288,512,0.004282666577233208
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,2048,4096,0.01720266706413693
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,256,8192,0.011415111521879831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,512,5120,0.00832622249921163
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,5120,1536,0.006586666736337874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,4096,1536,0.0069351109365622205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,1536,256,0.0052382221652401825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,512,2048,0.006622222148709827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,65536,6144,0.14148710833655462
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3072,32,0.0021608887861172357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2048,1024,0.003495111233658261
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,64,3584,0.01612711118327247
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1536,64,0.0022159999029503926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,5120,768,0.005945777727497949
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,16384,7168,0.03595911131964789
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1024,512,0.006004444426960415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,16384,3072,0.018719999326599967
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,65536,5120,0.10929599735471938
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,12288,65536,0.14762400256262884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,1024,6144,0.01108444482088089
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,4096,3072,0.023453333311610754
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,7168,5120,0.015359110302395292
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,32,128,0.006259555617968242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,16384,65536,0.21370932790968153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,10240,4096,0.016350222958458793
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,5120,8192,0.021909332937664453
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,7168,5120,0.01109155515829722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3072,3584,0.010149333212110732
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,128,8192,0.010775999890433418
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,10240,128,0.006264888991912206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,2560,1536,0.015192889504962497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,4096,6144,0.012088889049159156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,65536,12288,1.2642044491238065
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,12288,256,0.006686222222116258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3584,5120,0.012121777567598553
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,8192,8192,0.019283556275897555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,512,256,0.00556355549229516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3072,10240,0.2284462187025282
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,1536,1024,0.026562665899594624
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,8192,1024,0.006666666517655055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,65536,768,0.03135910961363051
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1536,6144,0.009935110807418823
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,6144,8192,0.03956177830696106
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,1536,8192,0.010773333410422007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,6144,128,0.019160888261265226
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,7168,256,0.005602666487296422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,65536,2048,0.08915911118189494
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,3072,8192,0.014183999763594734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,6144,65536,0.23066843880547416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,4096,4096,0.06740978029039171
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,7168,3072,0.03210044569439358
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,256,256,0.0022328889204396144
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2048,512,0.00258577780591117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,3072,3584,0.007418666448858049
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2560,16384,0.035982221364974976
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,3584,3072,0.007239110767841339
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,1024,2048,0.0063635553750726916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,4096,2560,0.013540444274743399
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,1024,12288,0.08013421959347196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,6144,16384,0.07971822553210788
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,12288,4096,0.047183109654320605
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,1536,512,0.005576000031497743
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,8192,5120,0.012504888905419244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,16384,4096,0.027835554546780054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,8192,3072,0.010188444621033138
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1536,32,0.005640000104904175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,7168,512,0.005948444621430502
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,64,3072,0.004592888885074192
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,8192,10240,0.04042933384577433
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,7168,5120,0.03152799937460158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,8192,2048,0.009876444107956356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,65536,1024,0.10632622241973877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,5120,16384,0.022039999564488728
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2560,4096,0.08412088950475057
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1536,12288,0.054121779070960156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,1536,3072,0.008732444710201686
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2048,32,0.002144888871245914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,128,1024,0.003149333306484752
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,6144,256,0.005592888842026393
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,3584,128,0.014849777022997538
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,128,768,0.006077333456940121
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,512,256,0.005991999887757831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,3072,1024,0.006243555496136348
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,10240,128,0.007610666255156199
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,3584,5120,0.00959644466638565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,65536,4096,0.11615555816226536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,5120,5120,0.02586488922437032
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,32,1024,0.005965333431959152
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3584,768,0.004264889078007804
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,16384,768,0.008447111480765873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,128,10240,0.011534222298198275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,12288,65536,0.43637956513298887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,10240,512,0.026495999760097925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,4096,16384,0.035407112704383
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,1536,128,0.005695111221737332
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,1024,4096,0.025664889150195654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,768,64,0.001855111163523462
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,7168,1536,0.0069617778062820435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2048,64,0.0038942222793896994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,256,512,0.006061333335108227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,2048,1024,0.009015111459626092
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,5120,1536,0.005982222242487802
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,3072,7168,0.06650400161743164
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2560,16384,0.022522666388087805
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,7168,512,0.0063368889192740125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,7168,12288,0.0214720004134708
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,256,256,0.005598222215970357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,128,65536,0.4838764402601454
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,3584,3584,0.030580444468392268
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,3584,768,0.016355555918481614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,5120,32,0.009391110804345872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,3072,1536,0.0063368889192740125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,16384,512,0.003929777691761653
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,2048,3584,0.017234666479958426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,1536,512,0.0059573331640826324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,2048,4096,0.016597333881590102
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,5120,65536,0.06942577494515313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,3584,768,0.023774221539497375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,256,3072,0.007364444434642792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,8192,1024,0.00630666646692488
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,5120,1024,0.0036524443162812125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3072,16384,0.06609155734380086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,128,7168,0.010078222387366826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,7168,7168,0.019150222341219585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,65536,1536,0.0321688883834415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,256,3072,0.004912000149488449
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,4096,4096,0.08004444175296359
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,10240,5120,0.0387368897596995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2560,3584,0.008901333643330468
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,128,7168,0.017267554998397827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,64,4096,0.004214222232500712
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,10240,4096,0.016044444508022733
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,256,1024,0.006577777779764599
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,10240,512,0.06571466392940946
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,12288,256,0.006292444550328785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,5120,2048,0.006568888823191325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3584,1536,0.005946666830115848
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,64,1536,0.006324444380071428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,3072,3072,0.007333333293596904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,256,12288,0.013167111410035027
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,7168,128,0.015196444259749519
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,12288,128,0.002485333424475458
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2560,128,0.006252444452709622
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2048,1024,0.005954666684071223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,2048,768,0.0063164445261160536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,65536,2560,0.08073066340552436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,768,128,0.014274666706720987
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,32,7168,0.005378666851255629
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,10240,4096,0.014198222094111972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,16384,2048,0.012090666426552666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,6144,16384,0.024251555403073628
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,7168,128,0.002879111096262932
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,32,32,0.005603555589914322
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,5120,64,0.00299466670387321
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,1536,1024,0.006268444574541516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,8192,768,0.003923555629120933
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,3584,1536,0.006260444306664997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,65536,768,0.10346311330795288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,4096,768,0.0058711109062035876
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,6144,256,0.004882666799757215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,8192,64,0.005347555710209741
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,1024,128,0.016582222448454965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,5120,1536,0.017515555024147034
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,7168,10240,0.04754399922158983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1024,8192,0.015541333291265698
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,7168,3584,0.00794488853878445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,8192,3584,0.018260444204012554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,65536,8192,0.15854756037394205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,8192,7168,0.020249777370029025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2560,2560,0.007164444360468123
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,128,16384,0.016823111308945548
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,64,8192,0.01032266683048672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,6144,8192,0.03456977672047085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,64,12288,0.004896000027656555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,16384,65536,0.2101608912150065
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,7168,128,0.005214222189452913
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,4096,3072,0.02409244411521488
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,3072,65536,0.06774755318959554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,16384,128,0.016189333465364244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,64,7168,0.010442666709423065
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2048,3584,0.021534222695562575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3072,128,0.005902222047249476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,3072,256,0.0052977779673205475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,65536,256,0.018964444597562153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,1024,1536,0.007303111255168915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,10240,3584,0.016214221715927124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,512,1536,0.0038746665749284956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,5120,1024,0.011782222323947482
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,768,5120,0.008371555142932469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1536,6144,0.009112888740168678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,3584,6144,0.010057777994208867
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,8192,7168,0.017657778329319425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,6144,5120,0.023181334137916565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,16384,768,0.019223110543357003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,128,2048,0.01723377737734053
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1024,768,0.006423111177153057
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2560,64,0.005949333310127258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,128,1536,0.006031111296680238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,65536,1024,0.03933866818745931
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,4096,3072,0.007317333585686154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,768,1536,0.0038924444880750445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2048,256,0.002593777866827117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,5120,7168,0.018838221828142803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,65536,512,0.013478221992651621
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,6144,16384,0.03260622090763516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,6144,2560,0.013228444589508904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,1536,1536,0.01683822274208069
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,2048,2048,0.01924444403913286
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,512,16384,0.05383822321891785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,16384,4096,0.01663555536005232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,16384,65536,14.7759518093533
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2048,1024,0.007033777733643849
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,5120,65536,1.626556396484375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1024,512,0.01649777756796943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2560,256,0.006654222392373615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3072,1536,0.007359111474619971
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1024,768,0.003600888782077365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,1536,2560,0.016349333855840895
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1536,128,0.0022133334229389825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,768,768,0.0028728888266616394
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,256,1536,0.0069822221994400024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1536,32,0.006315555423498154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,6144,65536,0.32178489367167157
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,65536,16384,0.3019759919908312
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,256,4096,0.008029333419269985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,1024,1536,0.0075804442167282104
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,256,32,0.0021475555582178966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,1024,256,0.0053280000057485365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,16384,65536,0.30028798845079213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,10240,10240,0.08816533618503147
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,7168,5120,0.02861066659291585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,2048,1536,0.01887022289964888
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,256,16384,0.01607466737429301
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,6144,128,0.014856888188256158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,5120,4096,0.024995555480321247
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,128,1024,0.02755733331044515
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,16384,5120,0.018603555030292936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2048,6144,0.03201155530081855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,3072,16384,0.1319813330968221
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,10240,1024,0.03016266557905409
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,8192,8192,0.015463999576038785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,10240,512,0.06491021977530585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,16384,12288,0.05564710828993055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,2048,10240,0.011784888803958893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,256,4096,0.008364444805516137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3584,32,0.006284444282452266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3072,128,0.005616888817813661
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,3584,8192,0.025593777497609455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,5120,16384,0.08265599939558241
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2560,64,0.007947555846638149
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,6144,10240,0.05409155620468987
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,256,5120,0.008975110948085785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2560,1536,0.006623999940024481
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3072,16384,0.019549333386951022
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,12288,512,0.04792266752984789
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,16384,16384,0.06972266568077935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2560,3072,0.06501955456203885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,32,7168,0.0052391112678580815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,512,5120,0.016506666938463848
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,512,10240,0.015792000624868605
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1024,32,0.002918222298224767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,7168,256,0.0025759999536805684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,5120,2048,0.01721333298418257
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3072,3072,0.049329777558644615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1024,768,0.005648889061477449
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,10240,5120,0.021215111017227173
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2560,256,0.0028702221396896574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,10240,10240,0.605638239118788
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,128,10240,0.007809778054555257
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,1024,3584,0.008660444782839881
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,128,5120,0.02334755493534936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,12288,2048,0.011145778000354767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,128,8192,0.010718222293588849
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3072,256,0.0029128889242808023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,10240,2048,0.012289777398109436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3584,256,0.0022728888110982045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2048,64,0.005657777604129579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,512,4096,0.010429333481523724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,2560,12288,0.016576000385814242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,16384,4096,0.090865777598487
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,8192,5120,0.015095111396577625
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,5120,16384,0.05610489183002048
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1536,3584,0.006091555373536215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,3072,128,0.0053173332578606075
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1024,8192,0.028187556399239436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,7168,5120,0.1198622253206041
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,3072,256,0.014834667245546976
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,1536,65536,0.14145067003038195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,5120,128,0.014494222071435718
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,128,4096,0.004561777744028303
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,12288,768,0.0069822221994400024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,5120,1024,0.004791999856630961
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,768,7168,0.010384889112578498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,2560,768,0.005653333332803514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,10240,3584,0.015395555231306287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,3584,16384,0.05036977926890055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,16384,2560,0.013629333012633853
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,64,1024,0.007239110767841339
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,512,7168,0.010088000032636855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,7168,5120,0.011969777444998423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,1536,65536,0.1161679956648085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1024,5120,0.06238933404286703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,16384,2048,0.015241778559154935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,1024,3584,0.047127998537487455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,256,3584,0.02367199957370758
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2560,128,0.005244444641802046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,512,1024,0.0031582222630580268
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,4096,2048,0.021010667085647583
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,4096,1536,0.015474667151769003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,256,2560,0.013800000150998434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,7168,1536,0.09072888559765285
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,768,3072,0.04061244593726264
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,3584,16384,0.0207360006040997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,32,16384,0.005201777650250329
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2560,2048,0.0070364442136552595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1024,16384,0.017486222916179232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,3584,512,0.016935111747847665
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,65536,3584,0.03303644392225478
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,7168,7168,0.016387555334303115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,5120,3584,0.009128889275921715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,1536,7168,0.009639110830095079
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,512,16384,0.018010665973027546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3072,256,0.006039999839332368
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,16384,2560,0.03524177935388353
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,256,10240,0.01201244443655014
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,6144,3584,0.02849422229660882
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,10240,768,0.009264888862768808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,1536,6144,0.07606044742796156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,2560,3072,0.010283555421564314
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,6144,1536,0.006544000158707301
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,10240,2560,0.01146755533085929
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,65536,3584,0.03691644469896952
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,3072,256,0.014855111638704935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,768,768,0.0069288888739215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,16384,128,0.00905511114332411
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,4096,4096,0.043506665362252124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,2048,3072,0.023745778534147475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,128,10240,0.011749332977665795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,10240,256,0.01754133403301239
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,12288,256,0.00867911097076204
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,128,32,0.002536000063021978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,5120,3584,0.025733333494928148
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,7168,65536,0.07924533552593656
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,6144,4096,0.02475111186504364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2048,1536,0.005897777775923411
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,7168,256,0.0028391112056043413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,32,2560,0.007134222322040134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2560,2560,0.006759111252095964
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,32,12288,0.005538666827811136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2560,8192,0.011081777513027191
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,12288,7168,0.024151111642519634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3072,7168,0.013168888787428537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,512,2560,0.015879111157523263
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,128,2048,0.006945778098371293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,4096,3072,0.017847999930381775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,5120,2048,0.0069679998689227635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3584,3584,0.05388977792527941
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,65536,768,0.023256889647907678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,4096,4096,0.12068977620866563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1536,3584,0.007553777760929531
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,3072,1024,0.023711999257405598
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,8192,2560,0.008805333740181392
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,16384,6144,0.04098577631844415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3072,2560,0.007034666836261749
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1536,16384,0.01720622181892395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,768,7168,0.013831111292044321
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,768,10240,0.012154666913880242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,1024,1024,0.03355288836691115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,1024,1024,0.0058959999846087555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,4096,65536,0.07166577710045709
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,5120,65536,3.1407430436876087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1536,65536,0.07656888829337226
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,7168,1536,0.008077333370844523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,512,10240,0.013578666581047906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,512,3584,0.007629333270920648
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3584,128,0.006412444429265128
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,10240,2560,0.011094222466150919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,128,256,0.005208000126812193
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,64,1024,0.006249777972698212
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,256,12288,0.004951111144489712
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,5120,6144,0.02852177619934082
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2048,32,0.0021582223061058256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,7168,5120,0.013762666947311826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,6144,3072,0.06780533658133613
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,768,10240,0.016343111793200176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,10240,5120,0.016229333149062265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,5120,5120,0.012450666891203986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,5120,16384,0.0665244460105896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,4096,16384,0.034322665797339544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,4096,4096,0.008026666939258575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,128,2560,0.00692622239391009
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,3584,3584,0.027063111464182537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,10240,6144,0.020223110914230347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1024,65536,0.3268062273661296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,768,6144,0.006252444452709622
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,7168,16384,0.062387552526262074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,1536,3584,0.050832887490590416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,128,2560,0.009391110804345872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,128,128,0.0025191110455327565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,2048,512,0.014793778459231058
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,128,1024,0.014197332991494073
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,1024,2560,0.01479644411140018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,6144,768,0.006166222194830577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,4096,2560,0.01719555589887831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,16384,64,0.0028080000645584534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,512,256,0.00591111100382275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,1536,5120,0.018545778261290658
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,12288,2048,0.019010666343900893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,1536,3072,0.007687110867765214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1536,10240,0.08024888568454318
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,10240,768,0.004922666483455234
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,64,32,0.0018728888697094387
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,6144,1536,0.006963555183675554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,128,64,0.002003555496533712
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,512,3072,0.009378666679064432
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,1536,10240,0.04381422201792399
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3584,6144,0.012143111063374413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,128,256,0.012462221913867526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,7168,4096,0.02542488939232296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,512,128,0.009347555538018545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,1024,1536,0.006212444355090459
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,10240,256,0.006973333656787872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,12288,512,0.020648888415760465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,512,16384,0.018382221460342407
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,16384,256,0.017108443710539076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3584,3072,0.010500444306267632
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,10240,2048,0.0177520000272327
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,7168,8192,0.016036444240146212
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,128,3584,0.00756444450881746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,8192,3584,0.18552622530195448
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,4096,10240,0.028712888558705647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,128,4096,0.008002666963471307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2048,32,0.006631999793979857
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,5120,65536,0.0999786655108134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2560,16384,0.17349421977996826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2560,1024,0.006621333460013072
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1536,12288,0.017795556121402316
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,12288,128,0.006260444306664997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,16384,2560,0.014969777729776172
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,65536,768,0.015126221709781222
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,12288,4096,0.05753688679801094
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,3072,2048,0.018555555078718398
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,128,6144,0.09814577632480198
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,12288,1536,0.025407110651334126
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,4096,3072,0.016943999462657504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,12288,32,0.006980444822046492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1024,2560,0.010016000105275048
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,32,8192,0.005902222047249476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,256,2560,0.013832888669437833
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,10240,512,0.01568622224860721
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,2560,512,0.006015111174848344
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,12288,6144,0.03050044510099623
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,5120,256,0.005296000176005893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,512,2048,0.01779911087618934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,4096,5120,0.02784444557295905
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3072,16384,0.04119911127620273
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,6144,32,0.005374222166008419
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2048,64,0.0063279999627007385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,5120,1536,0.021066667305098638
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,128,7168,0.009384888741705153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,256,32,0.006631111105283101
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2560,3072,0.021584888299306233
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,8192,3072,0.010813333094120026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,16384,8192,0.05448266532686022
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,10240,5120,0.03866844375928243
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,768,2560,0.006969778074158563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,6144,3584,0.010428444378905825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,12288,4096,0.019263111882739596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,10240,768,0.0076453329788313965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,7168,16384,0.025235555238193933
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,3072,3584,0.008032889001899296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,8192,1536,0.006931555353932911
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,3072,4096,0.007992000215583378
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,256,1024,0.0032088889016045462
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,10240,6144,0.022277333670192297
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,32,7168,0.05736177497439914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,256,65536,0.016558221644825406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,12288,128,0.006279110908508301
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,16384,12288,0.033160001039505005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,768,256,0.005631999837027655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,4096,256,0.006639110959238476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,6144,12288,0.08826844559775458
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,256,128,0.005664888769388199
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,16384,12288,0.15062577194637722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,4096,8192,0.015301333533393012
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,768,6144,0.009553777674833933
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,5120,8192,0.022080888350804646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,10240,6144,0.020104888412687514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,3072,10240,0.034095999267366194
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2560,10240,0.012979555461141797
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,32,64,0.00702400008837382
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,4096,4096,0.00797333319981893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,64,768,0.005578666511509154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,3072,16384,0.03519644339879354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,8192,4096,0.03005689051416185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,6144,4096,0.011807999677128263
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1024,1536,0.006633777585294511
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,512,512,0.00562755556570159
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1024,64,0.006320888797442119
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1536,1024,0.006266666783226862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,2048,7168,0.009377777576446533
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,16384,8192,0.051114665137396924
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,768,6144,0.04349866509437561
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,16384,65536,0.2871742248535156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,7168,8192,0.016275554895401
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,256,7168,0.010032888915803697
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,32,12288,0.013399110900031196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,16384,3072,0.1632631089952257
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,16384,128,0.015501333607567681
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,512,10240,0.025174222058720056
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,65536,3584,0.0694942209455702
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,10240,7168,0.01962488889694214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,3584,1536,0.015973332855436537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,3584,2560,0.0068808889223469635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,8192,3072,0.008382221890820397
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,3584,2560,0.02627022233274248
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,128,2560,0.01516266663869222
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,1024,128,0.014480888843536377
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,32,128,0.005676444619894028
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,256,2048,0.005072888814740711
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1024,1024,0.0032302221904198327
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,3072,2560,0.0196115556690428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,768,4096,0.008353778057628209
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,6144,32,0.002543111021320025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,8192,7168,0.5404382281833225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2048,64,0.010407999985747868
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,4096,8192,0.0760515530904134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,64,32,0.003961777935425441
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,512,128,0.007648000286685095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1536,2048,0.0064088888466358185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2560,3072,0.006762666834725275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,65536,6144,0.11047999726401435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,8192,1536,0.0069466663731469055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,256,65536,0.04868622289763557
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,3072,1024,0.02293955617480808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,768,10240,0.022191999687088862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,7168,64,0.006241777704821692
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,6144,2048,0.0075253330998950535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1536,1536,0.004563555535342959
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,3072,16384,0.025302223033375208
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,6144,4096,0.013328889177905189
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,7168,256,0.006127111199829314
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,16384,128,0.061774220731523305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,4096,2560,0.037512001064088606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2048,256,0.00258933338854048
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3072,7168,0.013791110780504016
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,512,3584,0.010703110860453712
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,65536,8192,0.23740977711147734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,768,3072,0.020679111282030743
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2560,3584,0.008655110994974772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,6144,3072,0.009730666875839233
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,1024,6144,0.043856001562542386
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,65536,256,0.07013510995441012
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,3584,256,0.01625955601533254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,3072,16384,0.03380711211098565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,512,1536,0.006606222026877933
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,32,3072,0.004242666479614046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,6144,512,0.005655999812814925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,1536,12288,0.015178667174445258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,4096,6144,0.010674667027261523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,7168,2560,0.008206222620275285
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3072,5120,0.01000533335738712
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,64,128,0.0021786666992637846
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,5120,2048,0.008538666698667739
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,6144,128,0.005347555710209741
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,16384,32,0.014831999937693277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,16384,10240,0.09885155492358738
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,256,2048,0.004929777648713854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,1536,1536,0.017955554856194388
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,16384,2048,0.036759111616346575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1536,4096,0.007694222033023834
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,768,512,0.00574933323595259
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,65536,16384,0.14446843994988337
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,8192,65536,0.18103999561733672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,32,16384,0.0120000003112687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,3072,65536,0.4204711119333903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,128,8192,0.01146755533085929
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,10240,256,0.002973333415057924
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,1536,768,0.017609778377744887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,768,12288,0.040933334165149264
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1024,12288,0.018135999639829
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,2560,16384,0.04397955536842346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,768,12288,0.01939022209909227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,65536,256,0.06287288665771484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,256,512,0.005939555664857228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,4096,3072,0.017260443833139207
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,8192,3584,0.01088355564408832
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,512,8192,0.010415111151006486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,6144,32,0.008622222476535374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3072,65536,0.07220533159044054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,512,1024,0.007027555671003129
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,512,6144,0.009406222237481011
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,5120,5120,0.06480622291564941
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,7168,8192,0.023057777020666335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2560,2560,0.07104977634218004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,4096,65536,0.06747821966807048
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2560,1536,0.012950222525331708
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,12288,10240,0.12038400438096787
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,256,3072,0.014863999353514777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,65536,256,0.007152888509962294
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,32,1536,0.019302222463819716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,8192,7168,0.02275200022591485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,128,512,0.002599999929467837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,7168,7168,0.012647111382749347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,65536,128,0.016530666086408827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,128,512,0.011415111521879831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,1024,6144,0.008746667040718926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,6144,128,0.005963555640644497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,3072,7168,0.04166222280926175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,12288,128,0.002558222247494592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,8192,10240,0.05473777651786804
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3584,2048,0.0069982219073507525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,64,768,0.0028693332440323303
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,10240,7168,0.022117333279715642
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,7168,16384,0.04774844315316942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,64,8192,0.0046453335218959385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,4096,7168,0.01312355531586541
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,5120,2560,0.014899555179807874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,65536,12288,0.43988800048828125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,2560,65536,0.9170506795247396
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,256,4096,0.0096124443742964
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,65536,3072,0.07833777533637153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,3584,10240,0.16559022002749973
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,7168,16384,0.04337866769896614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,128,7168,0.018616888258192275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,16384,2560,0.015235554840829639
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,256,6144,0.009122666385438707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,7168,1536,0.00904088881280687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,32,32,0.0020417778028382198
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,2048,512,0.013175999952687157
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,12288,4096,0.016509332590632968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3584,8192,0.013551111022631327
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,10240,1024,0.012521777715947894
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,4096,65536,0.10043911139170329
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,10240,7168,0.09816977712843154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1024,16384,0.017657778329319425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,65536,1024,0.0643706652853224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,768,5120,0.009399111072222391
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,12288,5120,0.02404533326625824
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,512,8192,0.010594666832023196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,3584,7168,0.011436444189813403
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,10240,32,0.006018666757477655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,65536,4096,0.04037066631846958
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,12288,4096,0.020655999581019085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1024,65536,0.7917440202501086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,768,1536,0.00830844458606508
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,512,1024,0.016923555069499545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,256,256,0.005699555493063397
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,768,2048,0.007997333175606197
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,512,7168,0.01035822182893753
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,512,4096,0.006618666566080517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,8192,16384,0.05011111166742113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,1536,768,0.005570666657553778
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,128,2048,0.004580444345871608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1024,256,0.005931555396980709
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,16384,1024,0.022849778334299724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,65536,512,0.05715644359588623
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,12288,7168,0.4995386865403917
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,4096,5120,0.01495733360449473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,1536,4096,0.02292799949645996
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,16384,1024,0.033702221181657575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,12288,128,0.028006222512986925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,256,10240,0.006631999793979857
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,128,128,0.011094222466150919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,6144,3584,0.014122666584120857
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,4096,7168,0.012693333129088083
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,5120,6144,0.016927111479971144
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1536,256,0.002568888788421949
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1024,8192,0.08840266863505046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,2560,3584,0.007666666474607255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,5120,16384,0.030305776331159804
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,65536,768,0.04008000095685323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,8192,2048,0.046760890218946666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,12288,1536,0.02347911066479153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1024,32,0.0052933332820733385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,1536,16384,0.05299733413590325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,128,128,0.012133333418104382
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1024,1024,0.005314666777849197
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,8192,4096,0.01128533316983117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,64,6144,0.0052426668504873914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,16384,32,0.010074666804737514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,512,512,0.005695111221737332
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,1536,512,0.006549333532651265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,4096,3584,0.008652444514963362
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,5120,6144,0.010683555569913654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1024,256,0.005561777700980504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1024,32,0.0069475554757648045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3072,3584,0.010118222071064843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3584,8192,0.015810666812790763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2048,10240,0.01032266683048672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,768,4096,0.00868622213602066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,768,1536,0.006249777972698212
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,32,768,0.005623999983072281
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,7168,1536,0.007327111230956183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,8192,768,0.006583111153708563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,8192,16384,0.04840533269776238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2560,768,0.004190222256713443
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1024,12288,0.013824000126785703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,1024,5120,0.020037333170572918
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,10240,6144,0.01828799976242913
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,768,32,0.0058764442801475525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2048,1536,0.006303999986913469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,6144,3584,0.015989333391189575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,12288,64,0.010144000252087912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,6144,4096,0.011950222154458364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,512,12288,0.061108443472120494
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,128,7168,0.02962755494647556
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,16384,3584,0.014564444621404013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2560,2560,0.019716444942686293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,6144,3072,0.008107555409272512
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,512,12288,0.01092888911565145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,5120,7168,0.030496888690524634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,8192,10240,0.02207999924818675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,1024,3584,0.01703999936580658
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,8192,16384,0.026612444056404963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,1536,16384,0.12448710865444607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,1024,1024,0.014614222778214348
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,768,3584,0.0306328899330563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,5120,6144,0.01145866678820716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,128,128,0.007050666544172499
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,8192,2560,0.01038044442733129
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3584,32,0.006167111297448476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,1536,2560,0.00701155596309238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3072,6144,0.1403448846605089
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,2048,6144,0.035304887427224055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1024,2560,0.005867555737495422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,7168,16384,0.1177342202928331
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,2560,1024,0.006073777874310811
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,6144,3072,0.01235466698805491
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,32,8192,0.02687377731005351
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,8192,2560,0.019976889093716938
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3584,1536,0.004362666772471534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2048,768,0.006362666686375936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,4096,5120,0.013208889298968844
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,16384,128,0.0028880000528362063
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,512,2048,0.013805333111021252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,5120,16384,0.04551911022928026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,6144,1536,0.015253333581818474
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,6144,32,0.002463999928699599
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,64,768,0.0064328888224230866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,768,5120,0.016910221841600206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,5120,16384,0.030918220678965252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,768,65536,0.12774754895104304
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,32,8192,0.005630222045713001
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,12288,256,0.00666133314371109
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,10240,512,0.006680888848172293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,128,6144,0.02578488820128971
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3072,7168,0.013095999757448832
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,3072,3072,0.0069724445541699724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,8192,16384,0.04515555500984192
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,128,1536,0.0036142224238978494
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,6144,3072,0.12609244717492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,5120,1024,0.008961777720186446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,2560,3072,0.017252445220947266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,16384,3584,0.03205688794453939
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,12288,256,0.01723022262255351
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,5120,7168,0.09487199783325195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,64,5120,0.015372445185979208
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,4096,2048,0.005408000200986862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,3072,8192,0.013777777552604675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1536,768,0.002921777880854077
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2048,16384,0.07027822070651583
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,6144,6144,0.02648177742958069
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1536,2560,0.0069884442620807225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2560,1536,0.0038773334688610504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,10240,10240,0.06016800138685438
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,5120,65536,0.254967106713189
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,65536,768,0.03393422232733832
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2560,65536,0.06283111042446561
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,1024,4096,0.00798044436507755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,12288,512,0.0031831111345026228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,128,12288,0.00526577772365676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,7168,6144,0.02032977839310964
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,5120,10240,0.024823110964563157
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,768,3584,0.0075937774446275495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,10240,16384,0.0957146684328715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,10240,6144,0.04409866531689962
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1024,65536,0.024018666810459558
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2048,16384,0.020981333321995203
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,12288,768,0.020971554848882887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,256,3584,0.008002666963471307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1536,3584,0.009874666730562845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,64,128,0.002218666589922375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2048,64,0.0052844443255000645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,7168,3072,0.00869955536392
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,256,5120,0.008364444805516137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,6144,2048,0.029350221157073975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,7168,512,0.003644444462325838
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,128,10240,0.08391910791397095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,16384,7168,0.2638382116953532
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2560,3584,0.009954666925801171
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,12288,4096,0.041013333532545306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,7168,16384,0.055693335003323026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2048,1536,0.0038968887594011095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,3584,128,0.005209777918126848
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,8192,1536,0.01776622235774994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2560,1024,0.028042667441897925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,1536,10240,0.03085333440038893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3072,8192,0.02034311162100898
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,5120,512,0.015458666616015963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,10240,7168,0.0196782218085395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,256,2560,0.0069617778062820435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,12288,6144,0.03590844406021966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,64,6144,0.010016000105275048
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,6144,3072,0.011094222466150919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,256,4096,0.015474667151769003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1024,2560,0.004926222066084544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3072,64,0.006706666615274217
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,6144,64,0.0022311111291249595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,12288,3584,0.013767110804716745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,16384,768,0.006280000011126201
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,256,512,0.005598222215970357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,16384,8192,0.37147198783026797
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1024,64,0.0022640000614855024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,4096,10240,0.015024888846609326
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3072,16384,0.020184000333150227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,768,6144,0.022445332672860887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,8192,768,0.023099554909600153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,8192,8192,0.033560888634787664
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3072,3584,0.00832444429397583
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,1024,768,0.005708444449636672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1536,8192,0.014560888210932413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,12288,4096,0.016407110624843173
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,3072,256,0.015497777197096082
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,12288,16384,0.0878560013241238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2560,3584,0.014397333065668741
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,12288,768,0.009751111268997192
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,7168,768,0.006027555714050929
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,65536,512,0.09280444516075982
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,10240,2560,0.010633777413103314
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3584,3072,0.013311111264758639
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,32,4096,0.0042453333735466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,128,2048,0.004585777719815572
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,6144,8192,0.050808889998330005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,5120,12288,0.034952001439200506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,5120,4096,0.011936888926559024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,6144,128,0.014862222803963555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,10240,12288,0.16756888230641684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3072,3584,0.00814666681819492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,768,64,0.005879999862776862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,3584,5120,0.009020444419648912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,7168,2048,0.00795555528667238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,768,256,0.005623111294375525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,512,1536,0.00833511104186376
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3072,10240,0.027882665395736694
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,7168,1536,0.019034667147530448
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,32,128,0.0021804444905784395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3072,2048,0.006162666612201267
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,1536,3072,0.007329777710967594
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,8192,1024,0.00508266646001074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,6144,3584,0.00962222201956643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,1536,256,0.015217777755525378
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,4096,4096,0.020682666036817763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,12288,64,0.005274666680230035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,1536,3584,0.021021333005693223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,32,5120,0.008979555633332994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2048,1536,0.006686222222116258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2048,32,0.010372444159454769
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,5120,3072,0.020828444096777175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3584,128,0.013785777820481194
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,64,1024,0.004924444274769889
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2048,8192,0.009384888741705153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,256,256,0.01388177772363027
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,32,1536,0.00673333348499404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1024,3072,0.01901155544651879
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,32,10240,0.005051555732885997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,12288,768,0.004935111022657818
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,512,2048,0.006598222172922558
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,65536,8192,2.1551946004231772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,6144,256,0.004549333204825719
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,256,64,0.01111288865407308
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,256,2560,0.0492524438434177
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,64,16384,0.005295111073387994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2048,5120,0.016902221573723685
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,256,7168,0.027081777652104694
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,12288,4096,0.014767999450365702
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3072,1536,0.006066666709052191
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,512,2048,0.004657777647177379
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,64,128,0.0022168887986077205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,768,768,0.01865688959757487
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,10240,7168,0.06604088677300347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,10240,6144,0.027162666122118633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,65536,10240,0.9073919720119901
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,3584,10240,0.014731556177139282
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,7168,256,0.03403466608789232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,256,1536,0.013515555196338229
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,12288,16384,0.8222817844814725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,8192,5120,0.03463110989994473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,12288,7168,0.06861244307623969
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,2560,2560,0.046316444873809814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,512,5120,0.025071110990312364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,512,3072,0.019882667395803664
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,16384,512,0.025260445144441392
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3584,128,0.023173333870040044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,4096,2048,0.005936000082227919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2048,12288,0.011797332929240333
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,5120,1536,0.006618666566080517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,16384,16384,0.08505600028567845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,12288,512,0.012671999633312225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,65536,256,0.055551999145083956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,8192,10240,0.021312000022994146
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,256,6144,0.00904177791542477
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,128,12288,0.023794665932655334
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,768,12288,0.013087111214796701
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,8192,2560,0.02260977857642704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,65536,1536,0.05381688806745741
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,7168,768,0.015971556305885315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3584,6144,0.013751111096805997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,768,5120,0.027830223242441814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,64,128,0.0021342221233579847
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,8192,2048,0.007174222005738153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,768,6144,0.008986666798591614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,16384,7168,0.02751111156410641
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,7168,7168,0.03537155522240533
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,65536,3584,0.1045866674847073
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,128,3072,0.005408000200986862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,256,10240,0.012058667010731168
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,65536,2048,0.023615999354256526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,2048,1536,0.02532177832391527
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,768,16384,0.0359226663907369
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,12288,512,0.01994044416480594
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,768,256,0.005249777601824866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,65536,2560,0.05029511120584276
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,128,128,0.010816888676749336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2048,1024,0.0032417778339650896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,12288,8192,0.027514666318893433
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2560,4096,0.008413333031866286
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,7168,4096,0.010986666712496014
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,16384,512,0.02233422299226125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,128,3584,0.007408888803588019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2560,1536,0.005569777968857024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2560,6144,0.009847111172146266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,1024,512,0.012777778009573618
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1536,2048,0.006599999964237213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,3072,2560,0.0083137775460879
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,16384,768,0.007206222249401941
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,7168,512,0.017511111166742113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,8192,1024,0.017587555779351127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1024,256,0.012103111379676394
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,128,5120,0.040652443965276085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,7168,1024,0.004383999854326248
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,768,16384,0.010038221875826517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1024,3584,0.04485244552294413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,6144,512,0.005978666659858491
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,4096,64,0.006621333460013072
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1024,3584,0.007733333441946242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,5120,3584,0.0240000006225374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,64,64,0.0018435555199782054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,10240,65536,0.29742489920722115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,256,3584,0.007662222617202335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,3072,768,0.005576888720194499
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,65536,8192,0.07419466972351074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,6144,16384,0.024736889534526404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,256,2560,0.014163555370436775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,16384,8192,0.03834222091568841
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,65536,128,0.006626666833957036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2560,256,0.012103111379676394
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,512,1024,0.005981333139869902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,128,128,0.006052444378534953
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,512,1536,0.014149333039919535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,512,10240,0.015480000111791821
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,16384,10240,0.23172177208794487
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,2048,2560,0.0069297779765393995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,1536,8192,0.011326221956147088
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,2560,128,0.01463822192615933
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,128,512,0.005655111124118169
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1536,2560,0.024329778220918443
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,4096,2560,0.006579555571079254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,1024,8192,0.036559998989105225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,2560,3072,0.021560889151361253
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,10240,3072,0.025895110434956018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,1536,6144,0.023059555225902136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,5120,6144,0.014179555906189812
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,16384,2560,0.016729777057965595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,7168,4096,0.043113778034845986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,7168,6144,0.07695733176337348
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3072,768,0.007269333634111617
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,128,4096,0.007797333101431529
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,5120,2048,0.018959111637539335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,32,4096,0.008007999923494127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,32,2560,0.007319110963079665
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,128,10240,0.012368000215954252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,256,768,0.006272000157170826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,512,768,0.005797333187527127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,8192,1024,0.033915556139416166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,8192,3584,0.024754666619830664
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,512,1536,0.013719999955760108
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,3584,8192,0.03526666760444641
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,10240,5120,0.03920533259709676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,3584,1024,0.005904888941182031
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,6144,512,0.003915555361244413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,256,65536,0.05849244197209676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,7168,6144,0.015881778465376962
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,12288,8192,0.031255112753974065
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,512,64,0.0021280000607172647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2048,128,0.002234666711754269
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,256,1024,0.0028906667398081887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2560,7168,0.14038399855295816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,3072,256,0.015779554843902588
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,6144,512,0.0059573331640826324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2560,4096,0.008357333640257517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,2560,16384,0.02108977735042572
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,8192,4096,0.01816444430086348
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3584,4096,0.009698666632175446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3584,512,0.0035173334181308746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,65536,128,0.19797866874270967
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,6144,256,0.0188773340649075
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,2560,128,0.014463111758232117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1536,4096,0.007281777759393056
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,256,512,0.002513777878549364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,128,3584,0.004908444566859139
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1024,16384,0.01829866733815935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2560,1024,0.004056888735956616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,7168,2560,0.02697066631582048
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,1536,4096,0.007702222300900354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3072,3072,0.011838222543398539
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2560,7168,0.013028444515334235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,12288,768,0.026086222794320848
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1024,4096,0.007632888853549957
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3072,5120,0.07655733161502414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1024,64,0.0053013331360287136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,6144,65536,0.07776000102361043
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,8192,2048,0.023266666465335425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,768,7168,0.006592000110281839
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,10240,256,0.009332444104883406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,1536,2048,0.017657778329319425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,7168,12288,0.030696001317765977
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,16384,16384,0.0517502228418986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,6144,1536,0.01767200065983666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,2048,5120,0.00866844422287411
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2560,3584,0.012496888637542725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2048,2560,0.010462221999963125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3584,65536,0.06924355692333646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,6144,2048,0.09385510947969224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,64,2048,0.006565333240562015
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,2560,3584,0.007894222107198503
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1536,768,0.0028826666788922418
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,128,4096,0.015136889285511441
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,12288,768,0.007319110963079665
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2048,3072,0.007332444190979004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,16384,128,0.01551999979548984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,5120,3072,0.020624000165197585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,12288,16384,0.06868088907665677
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,8192,10240,0.24370755089653862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,10240,768,0.006671999891599019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2048,7168,0.011963555382357704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,4096,768,0.005952888892756567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,32,3072,0.009723555710580613
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,12288,256,0.028374221589830186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,6144,10240,0.03451733456717597
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1024,10240,0.021469333105617102
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,12288,10240,0.05855377515157064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2560,12288,0.01462933255566491
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3584,32,0.022340445054901972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,7168,4096,0.010299555957317352
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,4096,512,0.002891555635465516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,4096,3072,0.017875555488798354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,6144,8192,0.04591288831498888
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,5120,10240,0.03677422139379714
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2048,1024,0.008691555923885768
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,8192,512,0.005611555443869696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3072,1536,0.0039057777159743835
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,2048,7168,0.02068000038464864
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,512,768,0.0029111111329661477
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,16384,65536,9.566524929470486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,768,3072,0.007268444531493717
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,65536,12288,0.26736177338494194
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,512,256,0.005304000029961268
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,16384,1536,0.009850666754775578
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,5120,6144,0.022878222995334204
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3072,6144,0.010741333166758219
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,7168,3072,0.01718577742576599
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,2048,3072,0.008672888908121321
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,12288,768,0.007622222105662028
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,65536,256,0.026885333988401625
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,5120,1536,0.02329066726896498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,512,1536,0.0062773335311147906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,6144,10240,0.055220444997151695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,65536,64,0.010034666293197209
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,12288,16384,0.04159644577238295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,1024,12288,0.013455111119482251
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,64,65536,0.04537333382500542
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,65536,1536,0.04376888937420315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,768,5120,0.008656000097592672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,2560,16384,0.04260266820589701
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,64,1536,0.006321777900060018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1536,1536,0.004936888813972473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,512,12288,0.013440888788965015
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,5120,512,0.003584889074166616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,8192,3584,0.009019555317031013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,512,2560,0.007306666837798224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,512,2560,0.007038222418891058
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2048,2560,0.02829688787460327
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,65536,2048,0.14560088846418592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,7168,32,0.006248000181383557
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,8192,5120,0.2744062211778429
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,65536,16384,0.40493332015143496
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,4096,128,0.0022053333620230355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,5120,16384,0.05171644356515673
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,7168,7168,0.02940355406867133
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2560,8192,0.011633777783976661
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3584,1024,0.006249777972698212
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,512,10240,0.012038222617573209
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,5120,1536,0.01958488921324412
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,16384,256,0.021033777130974665
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,8192,3072,0.010001777774757808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3584,6144,0.10105688704384698
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,65536,1024,0.08471022049585979
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,8192,128,0.005934222290913264
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,256,512,0.0032302221904198327
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1536,65536,0.05804355276955498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,6144,128,0.005615111026499007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,32,2048,0.0069715554515520734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,512,1536,0.006162666612201267
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,768,8192,0.010695999695195092
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,12288,5120,0.018213333355055917
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,7168,8192,0.018654222289721172
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2048,5120,0.010784889260927836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,7168,2048,0.0070186663005087115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,128,16384,0.0063475556671619415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,5120,7168,0.016712889075279236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,4096,7168,0.10599022441440159
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,128,1536,0.005939555664857228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,3072,128,0.014862222803963555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,6144,256,0.005640000104904175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,6144,256,0.0052737775776121355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,512,32,0.002136888810329967
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,12288,2560,0.012134222520722283
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,64,128,0.005616000129116907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,5120,512,0.0032248888164758682
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1536,64,0.002566222308410539
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,4096,65536,0.06993244753943549
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,10240,3072,0.014808000789748298
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,6144,5120,0.07954933245976765
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,768,64,0.002275555498070187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,65536,256,0.02511111067401038
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,7168,512,0.017198221551047433
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,10240,1536,0.008019555773999956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,65536,128,0.05675644344753689
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,768,128,0.00665866666369968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,16384,7168,0.10672177871068318
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1536,6144,0.00964355551534229
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,6144,6144,0.018631110588709515
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,6144,12288,0.019525332583321463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,4096,1024,0.003593777616818746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,10240,32,0.0021946666141351065
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,1024,6144,0.031127111779318914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,1024,768,0.016311999824311998
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3072,64,0.002895111011134254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,64,4096,0.011316444310877057
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,8192,2560,0.027331555883089702
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,3584,2048,0.006599111275540457
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,10240,768,0.018631110588709515
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3072,128,0.01070666644308302
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,7168,128,0.013171555267439948
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3584,256,0.004537777768241035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1536,8192,0.00941066692272822
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,1536,1536,0.0069271110826068455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,7168,256,0.003216888962520493
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,128,128,0.011790222591824003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,768,1024,0.005956444475385878
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,16384,10240,0.0446648895740509
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,12288,2048,0.009728888670603434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,768,512,0.004849777867396672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,3584,7168,0.011870221959220039
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,2560,4096,0.008013333711359236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,4096,3072,0.008377778033415476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,128,3584,0.012140444583363004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,8192,3584,0.010968888799349466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3072,4096,0.011719110939237805
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,128,5120,0.01054488867521286
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,32,2560,0.006956444846259222
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,7168,65536,0.13518044683668348
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,8192,512,0.005631111148330901
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,768,4096,0.01642666591538323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1536,12288,0.016556445095274184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,5120,2048,0.026352889007992212
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,128,2560,0.007297777467303806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2560,512,0.0035964445107513
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,512,2048,0.01443911095460256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,3584,2560,0.02211644417709774
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,10240,5120,0.019720888800091214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,12288,12288,0.03676088982158237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2048,12288,0.054177777634726636
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,512,16384,0.027467555469936792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,6144,12288,0.049351112710105054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3584,3584,0.012122666670216454
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,6144,2048,0.006986666884687211
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,6144,5120,0.025088888075616624
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,12288,512,0.007416889071464539
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2560,12288,0.015076445208655464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,4096,65536,0.08905955817964341
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,10240,1024,0.007983999947706858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,6144,5120,0.013069333301650153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,10240,2560,0.032152000400755144
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3584,1024,0.021519111262427434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,8192,512,0.005941333456171884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,65536,16384,9.515641106499567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,256,16384,0.02036177780893114
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,128,1024,0.018548443913459778
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,2048,1024,0.01553422212600708
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,4096,7168,0.028167112006081477
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,512,2560,0.0069617778062820435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,32,3072,0.004595555365085602
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,5120,2048,0.007288888924651676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,2560,7168,0.1050106684366862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,32,4096,0.0321084459622701
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,16384,7168,0.07658044497172038
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,65536,8192,4.148821512858073
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,768,2560,0.010026666853162978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,256,10240,0.007009777757856581
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,1536,8192,0.010777778095669217
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,8192,512,0.006052444378534953
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,10240,10240,0.02753066685464647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,4096,1536,0.006314666734801398
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1024,12288,0.013802666631009845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,1536,16384,0.030619555049472388
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,8192,256,0.005608888963858287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2048,5120,0.007347555624114142
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,128,3072,0.004967999955018361
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,3584,3072,0.007335999773608313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,6144,3584,0.023717333873112995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,8192,12288,0.021125333176718816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,8192,12288,0.06278400288687812
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,768,6144,0.04056266612476773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1024,1024,0.005884444548024072
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,5120,128,0.005245333330498801
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,32,512,0.005615111026499007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,1536,12288,0.04372444417741564
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,256,65536,0.016526222229003906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,16384,65536,0.21598577499389648
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,256,1536,0.003580444388919406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,7168,2048,0.05543555484877693
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,6144,12288,0.3643431133694119
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,65536,256,0.025314667158656653
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,3584,3072,0.007315555380450354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,768,4096,0.008580444587601556
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,16384,65536,0.17130843798319498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,8192,3072,0.02575911084810893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,2560,7168,0.014172444740931192
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,256,1024,0.012734221915404001
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,32,1024,0.005616000129116907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,512,3072,0.005286222116814719
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,5120,512,0.0035546666218174826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,8192,256,0.017176000608338248
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,2560,128,0.006333333336644703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,4096,128,0.0052737775776121355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,768,6144,0.018184888694021437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1024,3584,0.007649777664078607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,768,3072,0.00795555528667238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,10240,65536,0.1906355619430542
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1536,12288,0.016746666696336534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,16384,8192,0.02528000043498145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,10240,4096,0.03509777784347534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,1024,2048,0.05068711108631558
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1536,65536,0.06124355395634969
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,2048,8192,0.02404800057411194
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1024,3072,0.009962666365835402
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,128,8192,0.004882666799757215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,5120,32,0.005980444451173146
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,256,768,0.006052444378534953
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,768,128,0.005567111074924469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,65536,256,0.00979200005531311
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,8192,3584,0.019832889238993328
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,10240,128,0.015462223026487561
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3072,64,0.005938666562239329
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,256,512,0.0052266667286554975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3584,1536,0.019134221805466544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,8192,3584,0.030574222405751545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,16384,128,0.015199111567603217
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,256,10240,0.02130577796035343
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,32,3072,0.005565333283609814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,6144,8192,0.014917333920796713
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,5120,2048,0.09792621930440266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,7168,1024,0.005945777727497949
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2560,128,0.0021973333011070886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,512,512,0.0052915554907586836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,12288,512,0.00655644428398874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,768,3072,0.006360000206364526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,10240,65536,0.14393333594004312
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,2560,10240,0.04875911275545756
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,5120,1536,0.016537777251667447
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,512,512,0.005582222094138463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,768,256,0.006637333167923822
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,65536,5120,0.0659102201461792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3584,1536,0.004657777647177379
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,768,3072,0.017065778374671936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,32,5120,0.008752889103359645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,10240,128,0.006629333313968446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,10240,7168,0.07775288820266724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,3072,65536,0.10002399815453424
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,6144,10240,0.05443911088837517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,16384,8192,0.030115554730097454
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1024,6144,0.010082667072614035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,1024,3584,0.007635555333561367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1024,3584,0.01055555542310079
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,512,256,0.005629333357016246
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,256,256,0.005252444495757421
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,32,128,0.0019084444890419643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,512,1536,0.013455111119482251
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,768,2048,0.014855111638704935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,128,2560,0.009364444348547194
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,12288,8192,0.02404800057411194
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,1536,1024,0.005640000104904175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,768,128,0.014135110709402295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,64,3072,0.025395555628670588
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,32,64,0.005617777920431561
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,3072,1024,0.016340444485346477
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,128,5120,0.0090524446633127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,7168,5120,0.11666756206088597
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,512,12288,0.044032888280020825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,768,768,0.0035342222286595237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,512,1536,0.019073777728610568
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2560,32,0.00463733325401942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2560,8192,0.012774222426944308
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,768,16384,0.033985777033699885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,8192,1024,0.020431111256281536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,128,8192,0.005548444473081165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,8192,7168,0.03564088874393039
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,768,3072,0.007487110793590546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,256,65536,0.0781208872795105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3072,2560,0.006786666810512543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1024,12288,0.01070666644308302
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,2048,2048,0.015475556254386902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,65536,12288,0.7238977750142416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,1024,7168,0.03362933463520474
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,5120,2560,0.006948444578382704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,10240,768,0.006621333460013072
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,768,4096,0.007967111137178209
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,1024,1024,0.017975999249352347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,768,32,0.005156444592608346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,8192,2048,0.007344889144102733
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,32,6144,0.0063075555695427795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3584,5120,0.020184000333150227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,4096,16384,0.04511555698182848
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1024,512,0.005998222364319696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,256,128,0.002187555655837059
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1024,3584,0.005611555443869696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,2560,6144,0.00907733374171787
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,16384,12288,0.03974399964014689
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3072,4096,0.008563555777072906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,3584,1024,0.00628355559375551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1024,32,0.0022453332526816265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,128,768,0.006280000011126201
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,2560,2048,0.006656888872385025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2048,2560,0.007631999750932057
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,10240,6144,0.18850933180914986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,3072,3072,0.027303111222055223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,7168,2560,0.026945778065257605
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,512,256,0.010124444133705562
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,16384,4096,0.03755377729733785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,7168,10240,0.9063128365410699
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1536,512,0.011767110890812345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,1024,3072,0.022291556000709534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,2048,3072,0.016948444975747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,4096,3072,0.007381333245171442
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,65536,6144,0.14593244923485651
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,6144,768,0.01724444495307075
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,8192,3584,0.009481777747472128
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,4096,10240,0.03215555681122674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,10240,4096,0.01699199941423204
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,12288,4096,0.07928088638517591
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,6144,1024,0.03374400072627597
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,16384,8192,0.026497777965333726
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3072,5120,0.00962755580743154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,256,64,0.002187555655837059
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,16384,4096,0.015276443627145557
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,2048,256,0.005294222384691238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,7168,512,0.02241777711444431
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,2048,128,0.005477333234416113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,256,32,0.005967111223273807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,768,3072,0.02797599964671665
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,4096,16384,0.024107555548350017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,128,6144,0.005372444374693765
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,512,768,0.005961777849329843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1024,2560,0.007391999993059371
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,768,3072,0.01789688898457421
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,16384,6144,0.029314666986465454
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,5120,1536,0.010654222634103564
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,512,6144,0.017948443690935772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,7168,10240,0.05584177705976698
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,1536,8192,0.01074399964676963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,10240,1536,0.008420444197124904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,4096,768,0.004634666774008009
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,16384,10240,0.02958400050799052
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3072,65536,0.1370897822909885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,7168,65536,0.39571287896898055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,1536,768,0.00795199970404307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,4096,65536,0.12475732962290446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,768,5120,0.017251556118329365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,7168,5120,0.22766666942172578
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,65536,1024,0.026168000366952684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,3072,512,0.005670222143332164
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,10240,2048,0.02404622236887614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,5120,512,0.010103111465771994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2560,4096,0.008447111480765873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1536,5120,0.008741333252853816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3072,2560,0.0075786668393347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,128,128,0.0052577778697013855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,2048,12288,0.028382221857706707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1024,768,0.005972444597217772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,65536,4096,0.1102035575442844
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,7168,512,0.0035982223020659555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,3072,512,0.016543999314308167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,3072,3072,0.02455288834042019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,128,768,0.013798221945762634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,8192,2048,0.008640888664457533
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,10240,1024,0.00628355559375551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,12288,128,0.009715555442704095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,6144,8192,0.0148026661740409
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,10240,32,0.005967111223273807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,2048,12288,0.014819555812411837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,128,256,0.0024480000138282776
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,8192,512,0.018216000662909616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,768,2560,0.005431111074156231
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,512,10240,0.020679111282030743
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,256,65536,0.016241777274343703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,128,3072,0.015480889214409722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1536,2560,0.006962666908899943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,1024,3584,0.007320889168315464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3584,128,0.00592533333433999
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,7168,3072,0.00905955582857132
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,4096,16384,0.02566844390498267
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,256,3072,0.0129466669427024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3584,128,0.0053075556125905775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,256,10240,0.00563466673096021
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,256,16384,0.016174222032229107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,65536,5120,0.10251555840174358
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,7168,6144,0.01866044435236189
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,768,7168,0.019819556011093985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,32,10240,0.031868444548712835
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,6144,3584,0.00960888879166709
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,64,64,0.005254222287072076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,16384,10240,0.047930667797724404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,1536,4096,0.018631999691327412
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,512,3584,0.006478222293986215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,64,2048,0.004951111144489712
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,8192,7168,0.016167999969588388
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,128,1024,0.007284444239404466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,7168,768,0.010103111465771994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,4096,16384,0.5185955365498861
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,256,512,0.003946666502290302
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,7168,8192,0.03184444374508328
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,16384,768,0.00795555528667238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,7168,8192,0.02077866593996684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,768,768,0.0029013332807355454
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3072,1024,0.006285333385070165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,128,256,0.0021582223061058256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2048,6144,0.0676106678114997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,512,3584,0.007599111232492659
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1024,64,0.002149333349532551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,768,65536,0.06204977962705824
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,128,1024,0.006312888943486744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,64,32,0.0021706666383478376
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,8192,2560,0.008620444271299574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,3584,12288,0.01681422193845113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,6144,12288,0.021212443709373474
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3584,4096,0.009699555734793345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,10240,12288,0.04601600103908115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,7168,5120,0.014468444718254937
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,4096,3072,0.007315555380450354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,256,4096,0.017624000708262127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,32,3072,0.007340444458855524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,256,16384,0.039862222141689725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,10240,4096,0.013907555076811047
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,3072,512,0.005534222142563925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,12288,512,0.006754666566848755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,6144,5120,0.010716444088353051
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,128,32,0.0018435555199782054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,768,8192,0.013861333330472311
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,3584,1536,0.006292444550328785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,1536,128,0.005574222240183089
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,32,512,0.0025066667132907445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,32,256,0.005612444546487596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,5120,3584,0.010323555933104621
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,10240,7168,0.019663110375404358
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,4096,64,0.001839111041691568
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,6144,512,0.002557333351837264
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,7168,65536,0.13501066631740993
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,3072,5120,0.03239999877081977
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,5120,3584,0.009359111388524374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,8192,3072,0.01091466678513421
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1536,6144,0.008718222379684448
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,7168,65536,0.28876177469889325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,4096,1536,0.01791288952032725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,2560,16384,0.16285599602593318
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,3072,1024,0.005581333405441708
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,12288,256,0.0028702221396896574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,16384,1024,0.15482400523291692
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,65536,2048,0.02348533272743225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,768,4096,0.00795999997191959
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,768,6144,0.007287110719415877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,3584,10240,0.03405333227581448
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,16384,7168,0.11935999658372666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1536,128,0.00590844452381134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1024,7168,0.009460444251696268
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,5120,65536,0.18971733252207437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,3584,6144,0.009305777649084726
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,65536,2048,0.4342631234063043
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,32,65536,0.007629333270920648
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,16384,16384,0.06961599985758464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,256,5120,0.02100533412562476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,64,12288,0.006301333506902059
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,1536,1024,0.014169777433077494
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1024,768,0.007648000286685095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,32,512,0.0025537777692079544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,65536,6144,0.1267697811126709
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,3072,65536,0.06644355588489108
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,7168,1024,0.019627556204795837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,8192,1024,0.006678222368160884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1536,512,0.002569777684079276
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,65536,3072,0.09777600235409206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,12288,1024,0.00794400026400884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,512,1024,0.006574222197135289
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,5120,1024,0.016576000385814242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,4096,8192,0.014160888062583076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,3584,4096,0.00794400026400884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,10240,6144,0.03587822119394938
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,512,2048,0.022831110490692988
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,512,768,0.014926221635606555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,768,4096,0.005960888746711943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,128,3584,0.005250666704442766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,8192,128,0.005616888817813661
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,7168,4096,0.014456000592973499
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,32,128,0.0021439999755885866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,4096,7168,0.012404444317022959
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,10240,4096,0.12697688738505045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,10240,7168,0.02184711065557268
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,128,128,0.0052728888889153796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2560,65536,0.07155377997292413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,2048,2560,0.0069395556218094295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,16384,6144,0.2777173254224989
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,8192,2560,0.026367111338509455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,2560,3072,0.007269333634111617
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,512,3584,0.006000888844331105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,16384,768,0.008107555409272512
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,256,3072,0.007418666448858049
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,5120,65536,2.022722668117947
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,6144,2048,0.05306844578848945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,3584,6144,0.04543733265664843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,3584,5120,0.03173333406448364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,10240,4096,0.030605332718955144
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,256,1024,0.005648889061477449
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,6144,7168,0.05627999703089396
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3584,8192,0.20507644282446968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,1024,512,0.012811555630630918
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,64,32,0.005530666559934616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,12288,5120,0.03089511063363817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,768,1024,0.012200000385443369
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,7168,10240,0.49532000223795575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,128,5120,0.022455111145973206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,256,10240,0.03356800145573086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,8192,10240,0.020968889196713764
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,7168,16384,0.04436444573932224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,4096,768,0.03180444571706984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3584,1024,0.004543999830881755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,10240,12288,0.06765866941875882
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,6144,2560,0.014891554911931356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,32,2048,0.006970666348934174
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2048,512,0.007302222152551015
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,7168,6144,0.011815110842386881
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,6144,1024,0.012141333685980903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,5120,16384,0.042703999413384326
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,8192,2048,0.013991999957296582
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,16384,16384,0.09006577730178833
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1024,10240,0.00979466653532452
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,16384,2560,0.034289777278900146
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,256,1024,0.005963555640644497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,7168,10240,0.07117066780726115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,8192,1024,0.019303111566437613
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3072,4096,0.00869422240389718
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,64,128,0.005572444448868434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,256,2560,0.0069919998447100324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,5120,4096,0.02660622199376424
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,12288,16384,0.17893155415852866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,8192,10240,0.031672000885009766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,4096,3584,0.008716444174448649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,512,2048,0.004274666723277834
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1024,6144,0.006712889091836081
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3584,512,0.0037333332002162933
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,7168,512,0.007083555890454187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,5120,5120,0.012453333371215396
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,128,768,0.01370933320787218
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,10240,128,0.015588444140222339
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2560,64,0.0019742221468024785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3072,256,0.002918222298224767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,16384,64,0.08008178075154622
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,4096,8192,0.01593866613176134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,32,32,0.00564177789621883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,16384,6144,0.14948888619740805
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,4096,4096,0.009617778162161509
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,2560,12288,0.049399998452928334
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,5120,10240,0.09239111344019572
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,6144,4096,0.025639111797014873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2048,32,0.002238222294383579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,3072,1024,0.005959110955397288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,4096,16384,0.02166755497455597
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,3584,16384,0.023734221855799358
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,768,10240,0.011063999599880643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,65536,6144,0.08477333519193862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,5120,32,0.007666666474607255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,5120,65536,0.12569244702657065
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,10240,3584,0.2200240029229058
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,64,5120,0.00869955536392
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,4096,1024,0.009974222216341231
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,768,2560,0.005576888720194499
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,64,7168,0.005581333405441708
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2560,4096,0.008289777570300633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,512,10240,0.09638044569227432
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,16384,256,0.006673777682913675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3072,4096,0.062421335114373096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,256,8192,0.018466666340827942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2048,12288,0.016979555288950603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,10240,4096,0.01865866614712609
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,512,10240,0.012477333347002665
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1536,1536,0.004005333201752769
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,3584,6144,0.010024000373151567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,5120,65536,0.1265244483947754
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,16384,4096,0.01555377741654714
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,65536,1536,0.0878417756822374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,5120,3584,0.17223910490671793
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3584,7168,0.014730667074521383
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2048,3584,0.007655999726719326
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,10240,7168,0.039303110705481634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,65536,512,0.01706044375896454
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,65536,1536,0.19352266523573133
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,16384,768,0.007688889073001013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,7168,2048,0.006918222126033571
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2048,16384,0.030545777744717065
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,16384,12288,0.042373332712385386
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,256,65536,0.060384001996782094
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,65536,16384,0.14113067256079778
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3584,8192,0.016452444924248587
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,5120,16384,0.26586755116780597
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,2048,12288,0.015612444943851896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,512,32,0.005552000055710475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,10240,65536,0.1393066644668579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,768,2560,0.02552355494764116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,6144,2048,0.05937866369883219
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,65536,3584,0.10849777857462566
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,8192,5120,0.03695199886957804
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,12288,3072,0.015357333752844067
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,128,32,0.001875555556681421
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1536,768,0.00627377794848548
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,65536,7168,4.688032786051433
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,256,256,0.0038702223036024305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,4096,256,0.02809777855873108
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,768,512,0.012804444465372296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3584,4096,0.020358221398459542
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,8192,12288,0.06253244479497273
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,16384,5120,0.0366284449895223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,2560,7168,0.03321244319279989
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,10240,1024,0.021627555290857952
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1536,10240,0.01038755559258991
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,128,65536,0.06127999888526069
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,12288,16384,0.04018400112787882
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1536,16384,0.019849777221679688
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2560,2048,0.006907555378145642
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,4096,512,0.01514933341079288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,7168,6144,0.018716444571812946
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,64,3584,0.004941333499219683
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,65536,768,0.03439999951256646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,512,512,0.005223111146026188
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,2048,7168,0.02373955481582218
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,7168,12288,0.020071110791630216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,4096,64,0.002926222152180142
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,12288,2048,0.03779733180999756
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,7168,256,0.01701955497264862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,10240,10240,0.053709334797329374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,7168,65536,0.0843671096695794
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,5120,1024,0.0059262220230367445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,4096,768,0.006298666612969504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,768,256,0.00527022240890397
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,128,512,0.011407111254003314
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,128,3072,0.004619555754794015
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3072,4096,0.008017777568764156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,12288,256,0.058627552456325956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,2560,3072,0.007319110963079665
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,65536,128,0.10444088776906331
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,128,10240,0.011757333245542316
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,2560,3584,0.00810933361450831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,3584,768,0.017222222354676988
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,256,5120,0.008323555191357931
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,1024,6144,0.008919999831252629
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3584,512,0.0027315556589100095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,7168,4096,0.012621333201726278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,768,2048,0.004293333325121137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2560,8192,0.10675733619266087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,10240,768,0.0052284445199701525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,128,10240,0.012108444339699216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,64,12288,0.013789333403110504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,4096,3072,0.05128533310360379
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2560,1536,0.008422222402360704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2048,2048,0.006933333145247565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,12288,8192,0.02388977838887109
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,10240,3072,0.013643555343151093
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,10240,256,0.04096800088882446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2048,128,0.008255999949243333
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3072,768,0.0031884445084465873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2560,2560,0.005938666562239329
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,5120,6144,0.07579911417431302
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,32,64,0.007711999946170383
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,5120,12288,0.021261332763565913
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,7168,1024,0.006296000132958095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,5120,8192,0.014150222142537435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,256,32,0.0069226668112807805
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,5120,128,0.0053022222386466126
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,768,1024,0.005967999911970562
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,1536,512,0.013252444565296173
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1536,1024,0.003167111012670729
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,16384,16384,1.0260675218370225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,32,3072,0.01200355589389801
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,256,16384,0.030914667579862807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,3584,4096,0.03416088885731167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,12288,65536,0.15314488940768772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,2560,512,0.03090311090151469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1024,4096,0.008278222547637092
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,4096,1024,0.006311999840868845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,7168,32,0.006155555446942647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,512,3584,0.00740444411834081
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,128,65536,0.08708088927798802
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3584,16384,0.1510897742377387
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,768,8192,0.012647999657524956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,768,3072,0.007432888779375289
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,5120,3072,0.007349333001507654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,16384,768,0.0069884442620807225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2048,512,0.006046222315894232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,1024,3584,0.007393777370452881
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,256,1024,0.012256888879670037
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,3072,1536,0.006670222100284364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,256,256,0.006311999840868845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,768,768,0.005525333185990651
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,12288,3072,0.027456889549891155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,8192,4096,0.026392888691690233
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,65536,3584,0.03478844298256768
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2048,5120,0.008934222161769867
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,7168,2560,0.08468800120883518
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,768,16384,0.017218665944205392
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,3072,12288,0.028259555498758953
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,5120,768,0.008601778083377415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,5120,128,0.004192000048028098
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1024,768,0.0038977778620190094
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,10240,6144,0.035218665997187294
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,6144,2560,0.02677511175473531
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,10240,4096,0.016865778300497267
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,7168,512,0.008716444174448649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,7168,3072,0.009320889082219865
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,6144,65536,0.11947377522786458
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2560,5120,0.015300444430775113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,512,128,0.005395555661784277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,6144,8192,0.017834666702482436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,1536,10240,0.06640355454550849
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,8192,12288,0.050831112596723765
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,16384,1536,0.01184622198343277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,3072,3072,0.007700444095664554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,128,2048,0.006315555423498154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,3584,4096,0.008028444316652086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2048,12288,0.025405334101782903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,65536,6144,0.15045421653323704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,16384,7168,0.05076000094413757
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2048,12288,0.0135057775510682
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,4096,5120,0.010195555786291758
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,5120,7168,0.043311112456851535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,10240,65536,1.0866355895996094
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,256,2560,0.03442755672666762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,7168,512,0.004975111120276981
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,2560,1536,0.008069333102968004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,128,512,0.0026088888860411118
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,256,7168,0.005643555687533484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3072,256,0.0053102220926019884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,5120,65536,0.07002311282687716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,1536,256,0.008678221868144141
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,512,5120,0.008981333010726506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,256,64,0.0021991110924217436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,256,1024,0.003181333343187968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,3072,1536,0.056503110461764865
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,8192,1024,0.0070159998204973005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,65536,4096,0.0403733319706387
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,256,7168,0.009703999592198266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1536,65536,0.036143110858069524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,3072,128,0.01479555500878228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,3584,8192,0.057300441794925265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,6144,8192,0.12525422043270534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3584,8192,0.013047999805874295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,256,2560,0.007304888632562425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,5120,256,0.00790666706032223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,32,5120,0.00903911143541336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,512,2560,0.007312888900438945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,10240,3072,0.025451555848121643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,128,3584,0.004950222041871813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,12288,7168,0.08654755353927612
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3072,16384,0.02134399943881565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,4096,8192,0.015227556228637695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,1024,1024,0.005962666538026597
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,128,768,0.002883555574549569
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,768,16384,0.033703111939960055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,7168,16384,0.18898933463626436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,7168,1536,0.007068444457319047
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2048,3584,0.07736888859007093
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,12288,1024,0.019075555933846366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,4096,512,0.00602311102880372
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,1536,8192,0.010471111370457543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3584,3072,0.011067555182509951
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,65536,64,0.006568000134494569
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,16384,6144,0.019490665859646268
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,3584,256,0.016899555921554565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,2048,1536,0.018210666047202218
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,8192,3584,0.014739554789331226
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,7168,7168,0.02032711108525594
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,6144,2048,0.0069671107663048645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,6144,4096,0.009322666459613377
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,512,256,0.014165333575672574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,8192,7168,0.020615999897321064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,1536,5120,0.00868711123863856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,3072,7168,0.035938666926489934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,256,3584,0.0073315559162033936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,64,3072,0.0053137776752312975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,1024,7168,0.019146665930747986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,3584,3072,0.016864000095261466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3072,256,0.0055440002017551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3072,10240,0.017196445001496207
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,8192,2560,0.052629331747690834
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,2048,256,0.0052666668262746595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,128,1024,0.002924444360865487
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,768,16384,0.042655110359191895
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,12288,1536,0.010608889162540436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,128,16384,0.02846222122510274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,16384,3072,0.013304000099500021
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,65536,6144,0.14138044251335993
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,256,6144,0.009014222357008193
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,3584,2560,0.023784000012609694
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,5120,7168,0.02955377764172024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,6144,12288,0.06987288925382826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,6144,4096,0.10080177916420831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,256,10240,0.01204711116022534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,10240,2048,0.030536890029907227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,1536,768,0.02047288914521535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,12288,3072,0.013276444541083442
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,512,768,0.015461333923869662
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,12288,128,0.012390221986505719
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,3584,1024,0.01534222231970893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,5120,4096,0.020296888218985665
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,4096,16384,0.7689048979017469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,16384,10240,0.16945689254336885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,6144,512,0.016750221451123554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,4096,4096,0.011711999773979187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,12288,10240,0.05504799882570902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,6144,7168,0.014917333920796713
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,512,2048,0.006622222148709827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,512,3584,0.016031111280123394
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,256,3072,0.01480977733929952
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,768,3072,0.015463999576038785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,5120,512,0.00591555568906996
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,512,2560,0.019529778096410964
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2560,6144,0.12129067050086127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,128,6144,0.01035377797153261
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,12288,128,0.0032720000793536506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,5120,5120,0.012137778103351593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,128,12288,0.006608888920810487
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3072,1024,0.006601777755551868
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,5120,768,0.013597333596812354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,8192,16384,0.21273510985904268
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,128,65536,0.29985067579481334
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,7168,3584,0.011120888921949597
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,6144,8192,0.026166222161716882
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,128,12288,0.01313066648112403
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2048,128,0.0107360002067354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,65536,16384,0.30786310301886666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,4096,2048,0.006606222026877933
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,5120,5120,0.009355555805895064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,128,6144,0.0173688895172543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2560,32,0.0022400000856982335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,64,10240,0.020025778147909377
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,6144,2048,0.01943911115328471
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,16384,3584,0.016025778320100572
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,5120,1024,0.004011555678314633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,7168,12288,0.032997333341174655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,65536,6144,0.2884097894032796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,64,32,0.005311111195219888
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,16384,65536,0.16647199789683023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2560,2048,0.005375111268626318
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,1536,10240,0.047015110651652016
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,512,7168,0.010095111197895473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3584,1536,0.006721777634488211
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,5120,768,0.018384888768196106
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,128,5120,0.01755644381046295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2560,256,0.005962666538026597
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,12288,7168,0.020554666717847187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,6144,1536,0.006592888798978593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,2560,8192,0.010764444039927589
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,8192,3072,0.01293777757220798
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,2048,8192,0.10727733373641968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,8192,4096,0.01329422245422999
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,1024,1024,0.005922666854328579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,12288,2048,0.04996888836224874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2560,2048,0.013402666482660504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2560,2048,0.00868266655339135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,512,10240,0.03393244412210252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,1536,5120,0.02102311121092902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,1536,8192,0.021736888421906367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,10240,512,0.018546667363908555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,10240,4096,0.2445920043521457
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,2560,10240,0.0127688886390792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,8192,16384,0.07921955320570204
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,6144,5120,0.11988977591196696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,2560,6144,0.1686462163925171
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,16384,1536,0.03106133474244012
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,5120,4096,0.023991111252042983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,8192,512,0.01791111131509145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,10240,3584,0.014633778068754407
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,1536,6144,0.008981333010726506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,16384,4096,0.019129777948061626
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3584,65536,0.1642693281173706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,512,16384,0.028470221492979262
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1536,64,0.00526311124364535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,16384,65536,1.0730159547593858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,12288,10240,0.12411555978986953
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,12288,16384,0.040644443697399564
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2048,2560,0.004918222212129169
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,3584,10240,0.025126222107145522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,4096,512,0.005621333503060871
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,4096,2048,0.010407999985747868
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,2048,2560,0.0151555554734336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,6144,8192,0.02433066732353634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,4096,768,0.02257688840230306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,10240,65536,0.2902951240539551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3072,7168,0.15997244252098933
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,16384,3584,0.3399653434753418
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,128,6144,0.013087111214796701
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,7168,3584,0.01738844480779436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,10240,128,0.04556711183653938
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,64,4096,0.007910221815109253
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3584,4096,0.01089244418674045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,3584,8192,0.01644622286160787
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2048,16384,0.020959999826219346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,8192,256,0.017147555947303772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,16384,3072,0.013654222091039022
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,3072,6144,0.009024000002278222
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,768,32,0.0021822222818930945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,32,2560,0.004931555440028508
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,512,64,0.0019866666860050624
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1536,32,0.006254222244024277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,512,16384,0.015545777148670621
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,32,65536,0.02754488918516371
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,512,8192,0.01960355540116628
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1536,32,0.002144888871245914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,256,1536,0.0038337777886125776
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,6144,5120,0.2963164382510715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,128,2048,0.014464888307783337
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,768,16384,0.01462311049302419
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,5120,5120,0.02004888819323646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,5120,7168,0.019554666346973844
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,256,4096,0.015817777978049386
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,6144,3584,0.029860443539089624
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1024,16384,0.018754666050275166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3584,8192,0.014546665880415173
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,12288,2048,0.030384888251622517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,65536,10240,0.19563021924760607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,512,3584,0.007259555160999298
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1536,65536,0.048152890470292836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,32,512,0.005621333503060871
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,1024,256,0.00527555536892679
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,1024,6144,0.023018666439586218
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,7168,512,0.03011288907792833
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,32,3072,0.011039999624093374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,12288,512,0.006337777607970768
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,3584,5120,0.021474666065639917
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,5120,6144,0.023192889160580103
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,4096,1024,0.003257777748836411
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,12288,2048,0.011076444553004371
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,6144,65536,0.2535173363155789
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,1536,12288,0.0277173337009218
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,1536,3584,0.02273333403799269
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,2560,5120,0.009391110804345872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,65536,4096,0.11574578285217285
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,256,4096,0.004604444321658877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,12288,8192,0.03643111056751675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,6144,16384,0.023904000719388325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,3584,65536,0.0658515559302436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,12288,1536,0.008008889026112026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1024,512,0.0029102222373088202
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,12288,32,0.0032871110985676446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,16384,2048,0.01164533363448249
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,256,4096,0.015662221444977652
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,12288,2048,0.028582221931881372
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,5120,512,0.02278577784697215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,10240,65536,1.326117303636339
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,6144,4096,0.009007111191749573
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,128,4096,0.02658044464058346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,1536,2048,0.019744000501102872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1024,256,0.005936000082227919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,512,512,0.0025920000755124623
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,6144,2048,0.007002666592597961
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,4096,1536,0.006692444284756978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,512,16384,0.011704000333944956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,3584,6144,0.011047110789351992
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,6144,16384,0.06726488802168104
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,3584,128,0.014575999644067554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,10240,768,0.07753333118226793
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,4096,12288,0.1805208921432495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,10240,65536,0.1554186609056261
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,12288,65536,0.25985601213243276
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3072,16384,0.021232000655598108
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,768,8192,0.010416000253624385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,8192,3584,0.02036177780893114
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,2048,5120,0.028179556131362915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2048,5120,0.016083555089102853
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,7168,8192,0.015185778339703878
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,10240,5120,0.1302808920542399
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,7168,768,0.010243555737866296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,6144,1536,0.01962933275434706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,4096,8192,0.01514488955338796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,4096,64,0.0025048889219760895
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,768,1536,0.01073066641887029
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,5120,2048,0.015995555453830294
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,768,16384,0.284679995642768
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,768,128,0.005215999980767568
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,7168,256,0.00564177789621883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,128,3584,0.008054222497675154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1536,8192,0.011065777805116443
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,256,128,0.0052800000541739995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,64,128,0.0022115556316243275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,512,1024,0.013394667042626275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,512,6144,0.006650666809744305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1024,2560,0.004939555707905027
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,768,3072,0.005636444522274866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,4096,1536,0.015840000576443143
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,3584,2560,0.006934222247865465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,1024,16384,0.02838488750987583
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,5120,5120,0.01643111142847273
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3584,7168,0.23241066932678223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,8192,6144,0.06485866838031344
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,1536,10240,0.03269333309597439
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,12288,1536,0.008455110920800103
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,256,3584,0.0052560000783867305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,16384,768,0.007082666787836287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,4096,512,0.006953777538405524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,1024,2048,0.034134222401512995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,12288,7168,0.04835377799140083
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,65536,6144,0.1482808854844835
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,2560,768,0.01685333251953125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,2560,768,0.020358221398459542
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,3072,65536,0.06568088796403673
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,7168,16384,0.20844177405039468
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2560,512,0.006280000011126201
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,12288,512,0.07824444770812988
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,7168,5120,0.01219733307758967
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,1024,5120,0.009755555954244401
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,768,128,0.0023235556566052968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,5120,7168,0.018427555759747822
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,5120,7168,0.01480977733929952
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,128,3584,0.004937777916590373
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,64,64,0.0063075555695427795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3072,6144,0.01391377713945177
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,32,32,0.0018853334089120228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,2560,3584,0.01999200052685208
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,6144,768,0.004228444563017951
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3584,512,0.005943999936183293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,1024,512,0.005608888963858287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2560,7168,0.010304888917340172
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,16384,64,0.0022808888720141514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1536,16384,0.01937333411640591
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,1024,4096,0.007986666427718269
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,64,4096,0.00795733349190818
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,6144,7168,0.016920889417330425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,3584,3072,0.007668444679843054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1536,5120,0.00963466697269016
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,12288,7168,0.023471999499532912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2560,65536,0.06068711148367988
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,12288,3584,0.026155556241671245
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2048,128,0.005633777628342311
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,128,65536,0.06933244731691149
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,256,16384,0.01093777765830358
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,2048,768,0.005609777652555042
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,256,256,0.0053324442770746015
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,768,2048,0.00498933345079422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,768,512,0.0028924445311228433
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,5120,768,0.02033155494266086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,2560,2048,0.018799111247062683
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2048,2048,0.006145777801672618
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,128,768,0.01349955548842748
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1536,8192,0.014624888698259989
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,32,1536,0.003779555360476176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,8192,3584,0.011071999867757162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,6144,65536,0.11983111169603135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,128,6144,0.005968889014588461
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,256,8192,0.010779555473062726
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,16384,3584,0.024725332856178284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2560,768,0.005934222290913264
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,6144,768,0.006240889016124938
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,3072,8192,0.01187999960449007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2560,1024,0.007314666277832455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,1536,12288,0.02663822140958574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,12288,3584,0.01349511080318027
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,512,10240,0.011711999773979187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,6144,12288,0.030441777573691473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3584,3584,0.009699555734793345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,8192,256,0.0025066667132907445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,65536,3072,0.05646577808592054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,64,16384,0.018350222044520907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,7168,512,0.002579555536309878
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3584,16384,0.057796445157792836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,5120,768,0.005660444498062134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,4096,2048,0.006974221931563483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,16384,65536,0.1618142260445489
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,256,2560,0.007669332954618666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,12288,3584,0.03388888968361749
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,6144,3072,0.008368000388145447
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,256,768,0.02029244436158074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2560,32,0.01182755579551061
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3072,8192,0.014446222119861178
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,2048,65536,0.07685333490371704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,16384,8192,0.05040977729691399
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,32,8192,0.005274666680230035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,5120,64,0.01672266589270698
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,10240,512,0.003947555604908201
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,7168,8192,0.02258577777279748
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,12288,128,0.04870666729079353
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,3584,4096,0.00794933322403166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,8192,768,0.023903111616770428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,256,3072,0.03604711095492045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,256,65536,0.01683822274208069
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,768,2560,0.00692622239391009
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,16384,16384,0.07387466563118829
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3584,5120,0.1301813390519884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3072,5120,0.008979555633332994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,3072,6144,0.026335999369621277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,65536,16384,0.13999644915262857
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,7168,6144,0.016898666818936665
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,4096,512,0.00526577772365676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2560,65536,0.06114221943749321
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,768,10240,0.01238666640387641
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,12288,65536,0.36001155111524796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,512,3072,0.010395555860466426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,65536,64,0.011624000138706632
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,2048,6144,0.03532800078392029
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,1536,4096,0.02749066717094845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,65536,16384,0.18726578023698595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2560,16384,0.02051555613676707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,10240,2048,0.03012977706061469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,768,8192,0.009951111343171861
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,7168,5120,0.016172443826993305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,128,256,0.0021706666383478376
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,8192,32,0.002163555473089218
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,16384,1536,0.009300444689061906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,3584,4096,0.007991111112965478
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,128,32,0.0025066667132907445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,6144,3072,0.0804799993832906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,12288,3584,0.017112000121010672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,3584,1024,0.015490666031837463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,128,5120,0.015465777781274585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1536,5120,0.008818666968080733
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,64,4096,0.004608889006906086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,1024,5120,0.009024000002278222
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3072,3072,0.0074551113777690465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,512,3072,0.007669332954618666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,4096,5120,0.011424000064531961
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,2048,1536,0.016193777322769165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3584,3072,0.011162666810883416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,1024,12288,0.013456000222100152
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,8192,16384,0.050343109501732715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,7168,3072,0.14041866196526423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2560,1024,0.003536888708670934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,32,10240,0.012784889174832238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,5120,16384,0.05461600091722277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,3072,3584,0.014177777700954013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3072,256,0.008271111382378472
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,3584,2048,0.006907555378145642
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,8192,7168,0.014313777287801107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,65536,2048,0.30828799141777885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,6144,3072,0.05027733246485392
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,5120,3072,0.008873778084913889
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,16384,256,0.017502221796247695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,7168,3072,0.027263111538357202
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,2048,512,0.00555377784702513
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,65536,7168,0.08810310893588597
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,8192,2048,0.045569777488708496
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,2560,768,0.005583999885453119
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,7168,6144,0.14041866196526423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,6144,3584,0.04270222120814853
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,65536,512,0.014709333578745524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,5120,10240,0.0404151115152571
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,8192,2560,0.010151999692122141
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,7168,3584,0.22720710436503092
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,6144,8192,0.015455999308162265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,1024,256,0.007984889050324758
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,128,3584,0.013759999639458127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3584,7168,0.016143111719025504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,128,7168,0.017863111363516915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,6144,7168,0.20754044585757783
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,512,3072,0.014836443795098199
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,768,10240,0.013304000099500021
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,128,256,0.0021724444296624926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,65536,128,0.010328888893127441
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,5120,16384,0.029901332325405542
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,32,3072,0.007989333735571967
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,4096,128,0.005254222287072076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3584,2560,0.006149333384301927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1536,256,0.0059368887709246735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,256,32,0.0018782222436534034
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3072,65536,0.19062222374810112
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1024,65536,0.029170665476057265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,5120,4096,0.014142221874660917
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2560,7168,0.07837955819235908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,512,8192,0.0196106665664249
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,256,12288,0.019479999939600628
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,5120,3584,0.06871555248896281
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,2048,6144,0.008990222381220924
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,64,3584,0.007263999846246507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,768,16384,0.01516711049609714
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,5120,12288,0.01773244473669264
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,1536,5120,0.008424888882372115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1024,512,0.005583111196756363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,12288,65536,0.42837778727213544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,7168,12288,0.019879110985332064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,32,7168,0.005538666827811136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,64,32,0.0018542222678661346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,1024,16384,0.07940977811813354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1024,512,0.014373333917723762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,512,1024,0.006271111054552927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,12288,3584,0.025387555360794067
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,65536,3072,0.09008355273140801
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3072,2560,0.013159111142158508
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,7168,64,0.006637333167923822
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1024,32,0.0018728888697094387
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,256,768,0.013468444347381592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,128,1536,0.00795199970404307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,10240,16384,0.4844951099819607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,65536,16384,0.22646133104960123
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,12288,16384,0.10965066485934788
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,6144,128,0.014774221513006421
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,1536,8192,0.02681244413057963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,128,12288,0.02875822120242649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,16384,10240,0.05691466728846232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3072,1024,0.0038542221817705366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,65536,7168,0.1375946601231893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,256,7168,0.01202222208182017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,1024,6144,0.015271999769740634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,4096,10240,0.016550223032633465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,4096,3072,0.009761778016885122
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,8192,32,0.005592888842026393
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,6144,1024,0.005896889087226655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,64,1024,0.005987555616431766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1536,768,0.0032266666077905228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,8192,128,0.005404444618357553
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,3584,2048,0.0758204460144043
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,4096,5120,0.029193778832753498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,32,1024,0.003003555453485913
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,8192,16384,0.04534044530656603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,64,2560,0.004975999808973736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,65536,128,0.010327111515733931
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,3072,256,0.005609777652555042
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,6144,256,0.017104888955752056
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,12288,128,0.006320888797442119
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,12288,6144,0.04398577743106418
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,6144,6144,0.06654844681421916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,4096,65536,0.08912355369991726
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2048,1024,0.00425511101881663
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,1536,7168,0.009703999592198266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2048,768,0.007497777541478475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,5120,128,0.005255110975768831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,10240,768,0.023822221491071913
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,10240,6144,0.08592444658279419
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,512,65536,0.05563555823432075
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,128,2560,0.006744888921578725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,3072,65536,1.9478933546278212
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1536,1536,0.00830577810605367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,6144,4096,0.023943111300468445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,3072,512,0.019985778464211356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3584,768,0.005960000058015187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,2048,7168,0.037943111525641546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3072,512,0.024783111280865137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,4096,3584,0.009199110998047723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,7168,10240,0.025527111358112756
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,128,256,0.012764444781674279
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,7168,12288,0.13506844308641222
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,3072,65536,0.06617866622077094
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1024,256,0.002542222125662698
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,2048,256,0.005588444570700328
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3584,12288,0.11525866720411514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,512,65536,0.02013066742155287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,64,512,0.00564177789621883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2048,256,0.0035457776652442086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,768,8192,0.02718577782313029
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,65536,1024,0.06864444414774577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,5120,8192,0.015139554937680563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,10240,10240,0.3164409001668294
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,512,4096,0.008038221961922115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,7168,512,0.008309333688682979
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1024,3072,0.005584888988071018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,8192,12288,0.03663466705216302
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,512,16384,0.02180533276663886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,6144,8192,0.06589955753750272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,7168,7168,0.017439110411538016
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,12288,65536,0.23142488797505698
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,16384,768,0.019118222925398085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,4096,8192,0.01644622286160787
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,6144,6144,0.01646133263905843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,128,65536,0.08578399817148845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,65536,1536,0.06942755646175809
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,8192,16384,0.04021688964631822
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,16384,1536,0.22499023543463814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,8192,8192,0.04024799995952182
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,512,6144,0.10757866832945083
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2048,128,0.0021537778278191886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,128,128,0.011744889120260874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,10240,2048,0.009939555492666032
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,10240,768,0.07717511388990614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,128,12288,0.02200355629126231
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,256,3584,0.007502222226725684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,12288,16384,0.061576002173953585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3584,10240,0.01997422178586324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1536,16384,0.25083200136820477
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,32,256,0.0021964444054497615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,128,3072,0.013803555733627744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,2560,1536,0.015104000767072042
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,2560,128,0.014501333236694336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,32,1536,0.0037573331760035623
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,2560,6144,0.024078221784697637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1024,128,0.0027182222240500977
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,512,3584,0.007792889244026608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,768,64,0.006594666590293248
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,256,2560,0.013798221945762634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,7168,10240,0.020557334025700886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,4096,5120,0.011979555918110741
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1024,32,0.008236444658703273
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,65536,4096,0.0995840032895406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,256,12288,0.00907466643386417
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2560,1536,0.005591111050711737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,5120,2560,0.00720355576939053
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,3584,768,0.005643555687533484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,32,64,0.0022400000856982335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,768,1536,0.004193777839342753
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,2048,10240,0.012272000312805176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,64,768,0.0028533333291610083
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,512,2048,0.00480355570713679
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,32,256,0.005848000033034219
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,65536,3072,0.07114489210976495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,768,5120,0.01958400011062622
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,256,4096,0.009204444785912832
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,10240,256,0.006616000086069107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,12288,768,0.022355554832352534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3584,5120,0.042576889197031655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1536,12288,0.012106666962305704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,256,3072,0.005581333405441708
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1536,2560,0.00490133340160052
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2560,256,0.002520888836847411
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,6144,12288,0.1403599977493286
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,256,7168,0.005240889059172736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1024,1024,0.007967111137178209
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,10240,1024,0.006966222491529252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1024,128,0.006602666858169768
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,65536,32,0.016604445046848722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,65536,4096,0.08278044727113512
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,2560,2560,0.007305777735180325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,2048,2048,0.01475288967291514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2560,65536,0.05395288930998909
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,12288,6144,0.024801777468787298
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,5120,1024,0.00628355559375551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,6144,16384,0.03782044516669379
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1024,8192,0.00926399976015091
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,10240,512,0.006298666612969504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,3584,4096,0.01664888858795166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,12288,4096,0.017078222499953378
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,10240,768,0.013937777943081327
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3584,8192,0.08737778001361424
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,5120,2560,0.007204444044166141
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,5120,3584,0.00849955528974533
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,6144,1024,0.005967111223273807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,10240,64,0.00254133323000537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,12288,6144,0.021505778034528095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1024,256,0.0025315555847353404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,65536,512,0.18305333455403647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2560,65536,0.6597368982103136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,1024,2048,0.009680888719028896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,16384,512,0.006680888848172293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,6144,4096,0.011578666667143503
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,7168,5120,0.012486221889654795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,5120,16384,0.02369511127471924
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,12288,5120,0.01865600049495697
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,12288,16384,0.05903644694222344
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3072,12288,0.14528977870941162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,128,3584,0.016805332567956712
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,256,6144,0.004981333182917701
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,128,512,0.0052560000783867305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,7168,128,0.0028880000528362063
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,65536,2560,0.031665778822369046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,8192,5120,0.01651733285850949
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3584,768,0.006014222072230444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,2560,4096,0.009778666827413771
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,128,128,0.0026515554636716843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,12288,5120,0.01865244408448537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,8192,768,0.017664889494578045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,256,8192,0.012091555529170565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,5120,2560,0.008641777767075432
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,12288,128,0.015191111299726697
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,2048,7168,0.029260443316565618
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,7168,4096,0.02978044417169359
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,256,128,0.011408000356621213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,8192,3072,0.010815111299355825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,16384,16384,0.07849778069390191
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,10240,16384,0.06083999739752876
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,65536,8192,0.2740088833702935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,6144,3072,0.011056889262464313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,1536,16384,0.02086044516828325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,5120,16384,0.023572444915771484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,12288,512,0.0032497776879204642
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,8192,10240,0.03579466541608175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,128,512,0.00286666676402092
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,7168,10240,0.0507315562831031
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1024,3584,0.01072799993885888
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3072,7168,0.012772444221708508
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,6144,65536,0.07686933543947008
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,65536,1536,0.07135377989874946
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1536,256,0.0025128889828920364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,6144,7168,0.018068444397714403
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2560,5120,0.057370669311947294
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,10240,6144,0.021317332983016968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,7168,1024,0.025778666138648987
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,128,3072,0.005927111125654644
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,3072,3584,0.029312888781229656
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,65536,1024,0.03496977686882019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,10240,768,0.006633777585294511
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1024,10240,0.009748444788985783
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,12288,7168,0.022508444057570562
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,10240,10240,0.10396977927949692
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,4096,3072,0.008867555194430882
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,1536,5120,0.00906755526860555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,12288,3584,0.016786666380034555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,65536,16384,0.1381395525402493
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,6144,16384,0.026100445124838088
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3072,128,0.005647111270162795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,32,3584,0.005960888746711943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,1536,128,0.005557333429654439
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,7168,3584,0.013620444469981723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,3072,128,0.014851555228233337
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1536,7168,0.06386044290330675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,128,128,0.005616888817813661
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,512,3072,0.005611555443869696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,1536,1024,0.01681066718366411
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,10240,4096,0.016012445092201233
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,128,6144,0.005264889034960005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,1024,65536,0.13331911298963758
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1536,1536,0.00666133314371109
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,8192,3584,0.054139554500579834
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,8192,1024,0.024145777026812237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,3072,12288,0.017236444685194228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,2048,2048,0.017637333936161466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,4096,4096,0.15267911222245958
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,65536,128,0.01958488921324412
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,3072,5120,0.02028622229894002
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,10240,7168,0.03991377684805129
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,1024,5120,0.02477333280775282
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,7168,10240,0.0344782206747267
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,10240,10240,0.029044446018007066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3584,10240,0.023457777169015672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,10240,3072,0.028757333755493164
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,10240,6144,0.020599999361568026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,2048,128,0.005207111024194294
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,7168,32,0.006855999843941794
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1536,768,0.005899555567238066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,4096,8192,0.01309511148267322
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,5120,2560,0.0070977773931291364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,512,2560,0.0237137774626414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,512,1536,0.006633777585294511
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,7168,1536,0.017851556340853374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,512,7168,0.009653333160612319
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,3072,512,0.014286221729384528
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2560,512,0.019337778290112812
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,512,3072,0.01536622146765391
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,3072,2560,0.01090400003724628
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3584,768,0.003850666599141227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,128,256,0.0058613332609335584
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,10240,2560,0.012551999754375882
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3584,1536,0.04761688907941183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,4096,768,0.020095111595259774
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,65536,4096,0.380312893125746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,768,2560,0.0053680001033677
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,4096,8192,0.015430221954981485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2048,768,0.003567999849716822
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2560,1536,0.00666933341158761
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,512,3072,0.01533511115445031
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,2048,128,0.005325333525737126
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,7168,2048,0.013385777672131857
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3072,65536,0.06022933456632826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,65536,7168,0.14781867133246526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2560,128,0.005620444400442972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,64,2560,0.005006222261322869
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,3584,128,0.00720355576939053
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,1024,6144,0.009530666801664565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,6144,128,0.01216533366176817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,256,256,0.0023911111056804657
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,512,1536,0.005995555470387141
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,512,768,0.008973333570692275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,512,2048,0.0069360000391801195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3072,4096,0.09472088681326972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,3072,1536,0.01885422236389584
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,3072,512,0.014760888285107084
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,16384,1024,0.026351999905374315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,512,768,0.015153777268197803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,128,768,0.002806222273243798
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,3072,1536,0.006261333409282897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,64,6144,0.009631111390060848
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3584,768,0.00628977765639623
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,6144,6144,0.02641600039270189
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,12288,3072,0.17857333024342856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,5120,3072,0.012459555433856117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,8192,65536,0.1538942257563273
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,10240,12288,0.03269688950644599
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,256,12288,0.02751733362674713
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,5120,4096,0.01109333336353302
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3584,256,0.0024871110088295406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,768,256,0.005534222142563925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1024,3584,0.00526311124364535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,12288,1024,0.020918221937285528
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,768,1536,0.013822221921549903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,512,3072,0.007265778051482306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,10240,6144,0.058238221539391406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,5120,768,0.03444000085194906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,768,8192,0.010677333507272931
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,256,16384,0.01629333363638984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,512,16384,0.021381333470344543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,65536,5120,0.40909422768486875
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,10240,512,0.022568888134426538
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,128,5120,0.008444444172912175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,8192,65536,0.08006933662626478
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1536,6144,0.04850577645831638
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1024,512,0.0032035555276605818
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3584,7168,0.01204177737236023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,4096,6144,0.09544799725214641
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2560,768,0.005920889063013925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,6144,6144,0.028773334291246202
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2048,2048,0.006592000110281839
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1024,2560,0.004860444615284602
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,128,1024,0.0028364445186323593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,10240,32,0.0028568889117903183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,768,5120,0.007651555869314406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,128,256,0.005895111295912001
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,6144,6144,0.013148444394270578
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,5120,2048,0.006311111152172089
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,3072,10240,0.0443848901324802
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,6144,1024,0.00416088890698221
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,65536,32,0.004559999952713649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,8192,7168,0.0991555584801568
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,4096,10240,0.016566221912701923
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,1536,5120,0.029475556479560003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,2048,8192,0.025713778204388086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,768,5120,0.01886755559179518
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,768,65536,0.029368887344996136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,8192,10240,0.02749066717094845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,256,3584,0.008039111064540016
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,128,8192,0.021670222282409668
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,64,1536,0.019109333554903667
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,5120,2048,0.00675644435816341
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,2048,65536,0.07079288694593641
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,3072,3072,0.0073626670572492815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,8192,12288,0.05105333195792305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,12288,2560,0.02639466689692603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,768,256,0.015531554818153381
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,512,10240,0.022993778189023335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3584,1024,0.007726222276687622
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,10240,5120,0.0984462234708998
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,65536,7168,1.9819422827826605
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,768,512,0.009329777624871995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,16384,128,0.006321777900060018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,3584,3584,0.008988444175985124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,16384,12288,0.076873779296875
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,64,768,0.014463111758232117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,32,64,0.0019680000841617584
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,4096,256,0.005260444349712796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,2560,1024,0.029203553994496662
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,4096,4096,0.009256000320116678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,65536,10240,0.4087653425004747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,7168,3072,0.043825778696272105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,12288,6144,0.02019733356104957
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1536,3072,0.007341333561473423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,512,768,0.005608888963858287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,3584,3584,0.027468444572554693
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,65536,6144,3.5030506981743703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,256,2048,0.023730667101012334
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,128,768,0.006001777946949005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,32,12288,0.01019377758105596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,256,768,0.005960000058015187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2560,768,0.005967111223273807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,4096,10240,0.025388444463411968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,65536,128,0.011052444577217102
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,16384,12288,0.08490488926569621
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2048,3584,0.00926755534278022
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,65536,2048,0.02755733331044515
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,1024,8192,0.01072177787621816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,16384,3584,0.03986400034692552
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,4096,2560,0.008551999926567078
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,10240,6144,0.022311111291249592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,256,3072,0.007334222396214803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,2048,4096,0.017550221747822232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,768,64,0.010415111151006486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,8192,6144,0.03344088792800903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,6144,1536,0.01127199994193183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,768,2560,0.00510044437315729
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,128,3584,0.007739555504586961
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,256,12288,0.007317333585686154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,2048,7168,0.03768444392416213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,16384,768,0.11421244674258763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,12288,32,0.0023973333752817577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,7168,256,0.009083555804358589
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,768,6144,0.01854222185081906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2048,64,0.005922666854328579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,3072,5120,0.02872266703181797
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,128,5120,0.009069333473841349
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,32,256,0.0021973333011070886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3584,16384,0.021640888518757288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,1024,3072,0.042796442906061806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,768,16384,0.06865777571996053
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,128,65536,0.1618222263124254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,512,3584,0.005665777872006099
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,4096,16384,0.03142222099834018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,16384,7168,0.036186665296554565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,12288,128,0.006231110956933763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2560,7168,0.016823111308945548
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,256,512,0.005371555685997009
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,5120,6144,0.032644444041781955
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,16384,6144,0.04861688945028517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,10240,3072,0.012882666455374824
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,5120,128,0.009408888717492422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,1536,12288,0.03083555565940009
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,512,12288,0.013061333033773633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,256,256,0.012343111137549082
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,256,3584,0.004875555634498596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,256,1024,0.006225777582989798
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,8192,4096,0.011024889018800525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,7168,16384,0.09713866975572373
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3072,512,0.003286222202910317
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2560,6144,0.012749333348539142
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2560,128,0.004618666652176115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,8192,2560,0.01550222271018558
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,768,2048,0.006705777926577463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,5120,128,0.005595555735958948
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,3072,512,0.006360888895061281
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,5120,10240,0.0354693333307902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,6144,3072,0.012235555383894177
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,10240,128,0.024897777371936377
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,6144,4096,0.020996444755130347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,256,4096,0.01608533329433865
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,8192,3072,0.009377777576446533
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,6144,2560,0.06217155853907267
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,5120,4096,0.011448889142937131
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,32,16384,0.046560890144772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,8192,256,0.005564444594913059
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,8192,1536,0.006693333387374878
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,768,3072,0.007350222104125553
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,12288,16384,0.06326666805479261
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,3072,1536,0.03744355506367154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,4096,1024,0.00620266670982043
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1024,1024,0.004015999949640698
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,65536,4096,0.08976888656616211
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,256,1536,0.01295555548535453
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,5120,512,0.015243555108706156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,2048,10240,0.04528355598449707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1024,65536,0.036510222487979464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,5120,2048,0.021015110943052504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,128,2560,0.010120000276300643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,65536,3072,0.03714311122894287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,8192,768,0.006623111251327727
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3072,4096,0.008674666285514832
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1536,6144,0.009384888741705153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,2048,768,0.005576000031497743
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,3584,256,0.005407111098368962
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,16384,16384,0.17037067148420545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1024,32,0.004947555561860402
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2560,65536,0.05492088860935635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,12288,6144,0.018909333480728995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,2560,10240,0.02576711111598545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,8192,1536,0.019702222612169053
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,4096,65536,0.12837244404686823
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,256,768,0.0028844444702068963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,4096,3584,0.015328889091809591
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,512,512,0.0025537777692079544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,16384,128,0.015507555670208402
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,1536,128,0.006215999937719769
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,512,2560,0.007294221884674496
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,16384,1536,0.09675911400053237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,7168,10240,0.02657777733272976
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,1536,256,0.00563733321097162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,3072,128,0.0148062225845125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,10240,3584,0.013344888885815939
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,768,12288,0.013188444077968597
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,8192,3072,0.023595554961098567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,4096,5120,0.01053422192732493
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,16384,5120,0.03752622339460585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,256,3072,0.01443733274936676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,7168,1024,0.03828622235192193
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,768,7168,0.01013777818944719
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,128,12288,0.01910044418440925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,4096,32,0.005872888697518243
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,3584,65536,0.07044088840484619
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,10240,32,0.002491555487116178
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,256,16384,0.0076453329788313965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,12288,16384,0.07373422384262085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,768,6144,0.018562666243977018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1536,65536,0.03535999854405721
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,768,12288,0.07969066831800672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,2048,16384,0.14249244001176622
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,16384,3072,0.014826666977670459
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,3584,2560,0.0899591114785936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,6144,2048,0.008310221963458592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,512,6144,0.006586666736337874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,6144,12288,0.01796711153454251
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,6144,5120,0.01237511138121287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,2560,6144,0.03562666641341315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,2048,10240,0.09395111269421047
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,12288,5120,0.021589333812395733
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,512,768,0.01517777807182736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,65536,12288,0.24438312318589953
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,12288,256,0.006607111129495833
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,256,65536,0.05686044692993164
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,6144,128,0.00222044438123703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,128,128,0.00535377777285046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,2048,8192,0.01072622256146537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,32,7168,0.009933333430025313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,768,768,0.006247111078765657
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3072,128,0.008004444340864817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,32,6144,0.009345778160625035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,2560,7168,0.02238133384121789
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,768,1024,0.009640889035330879
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,10240,4096,0.01901955571439531
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,65536,4096,0.04134933484925164
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,768,65536,0.1193457841873169
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3072,12288,0.01832266648610433
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1024,128,0.005594666633341048
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1536,64,0.0021973333011070886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3584,512,0.0026586666289303037
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,256,16384,0.01623911162217458
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,10240,768,0.010785777535703448
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,7168,5120,0.0387368897596995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1024,5120,0.008663111262851292
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,3584,5120,0.07315022415584989
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2560,3072,0.006622222148709827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,12288,5120,0.01870577699608273
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,512,256,0.0029031110720502008
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,12288,768,0.006301333506902059
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,2048,10240,0.024280000064108107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,16384,1536,0.050869332419501416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,16384,3584,0.03286400106218126
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,12288,6144,0.09761689106623332
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3072,1536,0.00563822231358952
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,64,65536,0.04539822207556831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,7168,256,0.015511110424995422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1536,128,0.0027297778675953546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,65536,3584,0.22885245747036406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1536,2048,0.004527111020353106
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,128,12288,0.013434666726324292
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,5120,768,0.00816888858874639
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,7168,768,0.03716088996993171
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,5120,16384,0.036943111154768206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,3072,10240,0.12027111318376328
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,64,256,0.005567111074924469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,2048,768,0.01832444469134013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,6144,3584,0.04442933201789856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3584,1536,0.00666133314371109
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1024,1536,0.006716444260544247
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,65536,3072,0.5440995428297255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,7168,256,0.02772355576356252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,768,7168,0.02139733400609758
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,5120,6144,0.01794488893614875
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,512,512,0.005934222290913264
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,3072,65536,0.20425332917107475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,65536,10240,0.10824977689319187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,256,12288,0.006551110910044775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,4096,512,0.0028719999310043124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,3584,2048,0.010879110958841113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,12288,65536,0.3004968961079915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,10240,32,0.017903111047214933
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,1024,3584,0.02572622232966953
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1536,3584,0.011720889144473605
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2048,3584,0.010270222193664974
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,3584,2048,0.02679466704527537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1536,5120,0.013417777915795645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,7168,16384,0.0336408880021837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,512,128,0.005595555735958948
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,128,768,0.0059164443777667145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2048,6144,0.008700444466537898
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1536,64,0.002165333264403873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,1536,12288,0.013457777599493662
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,256,8192,0.01960711181163788
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,4096,1536,0.006278222219811545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,2048,10240,0.015272888872358533
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2560,64,0.005943999936183293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1536,128,0.0021848889688650766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,128,16384,0.01578044394652049
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,256,16384,0.13436799579196507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,1024,128,0.00776711106300354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,64,64,0.005662222289376789
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,5120,128,0.012875555290116204
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,65536,2048,0.023748444186316595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,32,1536,0.006265777680608962
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,8192,512,0.01719288859102461
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,1024,3584,0.01850488947497474
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,5120,128,0.014826666977670459
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,8192,6144,0.05925866630342272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,2048,7168,0.01036088913679123
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,256,512,0.012395555774370829
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,32,2560,0.00490399988161193
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,5120,6144,0.03541333476702372
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,512,65536,0.059030221568213574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,10240,64,0.027934223413467407
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,64,128,0.0024604443460702896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,8192,2048,0.010985777609878115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,1024,6144,0.0237128883600235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,6144,7168,0.013192888763215808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,128,1024,0.0031795555518733132
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,3072,4096,0.018052443861961365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,256,12288,0.04353777898682488
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,7168,2560,0.00959288908375634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,10240,256,0.006194666855865055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,7168,12288,0.02918755677011278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,10240,65536,0.142140441470676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,65536,1536,0.07129333416620891
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3584,16384,0.023528888821601868
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,3584,3584,0.02638133366902669
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,5120,256,0.013082666529549493
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1536,65536,0.05843200286229452
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3072,8192,0.013821333646774292
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,2048,1024,0.0148062225845125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,6144,128,0.005437333136796951
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,5120,65536,5.835370805528428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,10240,7168,0.03197955422931247
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,10240,2560,0.011221333510345884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,10240,10240,0.05920355849795871
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,16384,5120,0.0831164452764723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,512,512,0.006040888941950268
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2048,8192,0.02299555473857456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,512,16384,0.02660711109638214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,1536,7168,0.02018666598531935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,128,65536,0.04944355620278252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,256,7168,0.018650665879249573
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,512,64,0.0021644443687465456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3072,6144,0.18425867292616102
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,512,1024,0.00434044458799892
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,10240,4096,0.019931554794311523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,16384,12288,0.04045866595374213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,128,2048,0.005023111071851519
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,12288,65536,5.359575907389323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,768,6144,0.024657777614063684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,2048,1536,0.03063911199569702
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,2560,12288,0.31742223103841144
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,4096,7168,0.05736977524227566
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,32,512,0.00254844439526399
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,768,65536,0.02100533412562476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,16384,6144,0.023204444183243647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,4096,1024,0.006755555669466655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,3072,256,0.015138667490747241
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,16384,1024,0.013622221847375235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,12288,8192,0.08068711227840848
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,1536,12288,0.03052622079849243
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,16384,2048,0.015086222026083203
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,1024,128,0.014631110760900708
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,128,3072,0.014192000031471252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2560,3072,0.02456088860829671
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,128,64,0.00565244464410676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,6144,10240,0.026299556096394856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3584,4096,0.009072889056470659
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3584,16384,0.044863998889923096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,256,7168,0.010076444182131026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2048,768,0.005864000154866113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,6144,5120,0.012121777567598553
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,64,5120,0.01277599980433782
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,5120,128,0.005636444522274866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,10240,12288,0.11762311723497178
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2048,4096,0.07459733221266004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,768,6144,0.007634667058785756
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,16384,10240,0.0302746660179562
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,3072,16384,0.03636533353063796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,2560,16384,0.2213644451565213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,768,6144,0.00905866672595342
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,1536,5120,0.009015111459626092
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,10240,2560,0.012433778080675336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2048,3584,0.02180266711446974
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,12288,2560,0.014140443669425117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,3072,768,0.017196445001496207
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2560,2048,0.004212444441186057
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,2560,7168,0.010623111493057676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,65536,128,0.01037688884470198
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,256,2560,0.008263111114501953
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,16384,1024,0.028082665469911363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,6144,5120,0.012011555333932241
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,128,16384,0.005988444305128521
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,4096,128,0.005660444498062134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,512,4096,0.00793599999613232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,128,6144,0.009726222190592024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1536,6144,0.009857777920034196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,5120,6144,0.30109866460164386
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3584,3584,0.008008889026112026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,10240,1024,0.031062222189373438
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2560,512,0.0035137778355015647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,2560,5120,0.018623999423450895
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,10240,1024,0.01756266587310367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,16384,3584,0.10582844416300456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3072,3584,0.009301332963837517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,256,2048,0.006593777901596493
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,64,64,0.0018355554590622585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,7168,65536,0.12004266844855414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,512,10240,0.024013333850436743
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,2560,1536,0.01584533353646596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,768,7168,0.009675555759006077
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,256,10240,0.02400355537732442
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,1024,256,0.005320000151793162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,2560,3072,0.022709333234363135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3072,2048,0.013367111484209696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,128,768,0.005586666779385672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,8192,5120,0.02352622151374817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1536,768,0.007000000112586551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1024,7168,0.007642666498819987
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,10240,2048,0.008649778034951951
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,6144,128,0.007290666302045186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,128,7168,0.009323555562231276
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,6144,3072,0.024192000428835552
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,64,512,0.0025315555847353404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,16384,5120,0.02495199938615163
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,6144,3072,0.008184000021881526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,512,64,0.0021679999513758556
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,5120,512,0.019271110494931538
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,8192,8192,0.03353155652681986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1024,512,0.005528888768619961
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1024,8192,0.012673777838548025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,3072,3584,0.02743822170628442
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,768,512,0.003185777821474605
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,1024,6144,0.009183111290136972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,6144,5120,0.013417777915795645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,1024,65536,0.120596448580424
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,1024,1024,0.00665155549844106
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,6144,7168,0.014793778459231058
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,8192,12288,0.03400889039039612
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2560,64,0.0022044444663657085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,256,3072,0.0069751110341813825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3072,64,0.010062221851613786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,2048,10240,0.013073777986897362
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1536,2048,0.007052444749408298
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3072,768,0.0031928889867332247
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,8192,65536,0.0878373318248325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,3584,65536,0.06595289044910006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,5120,2560,0.015556444724400839
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,128,12288,0.05937600135803223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3584,3072,0.00721066693464915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,65536,128,0.006065777606434292
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,512,7168,0.010043555663691627
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,10240,2560,0.010278222461541494
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,7168,6144,0.02950577934583028
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1536,3072,0.007265778051482306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,10240,1536,0.007260444263617198
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,10240,512,0.024653333756658766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,256,5120,0.01683288812637329
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,8192,2048,0.008642666869693333
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,768,8192,0.013259555730554791
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,6144,12288,0.021565333008766174
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,6144,512,0.01682133310370975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3584,6144,0.012634666429625617
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3584,10240,0.016248888439602323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,10240,3584,0.015357333752844067
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,6144,8192,0.04839377933078342
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,5120,3584,0.009467555416954888
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,16384,7168,0.03604888916015625
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,1024,256,0.017489777670966256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,32,32,0.005559999909665849
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,5120,64,0.008303999900817871
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,512,3072,0.01484266585773892
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,6144,16384,0.04181155562400818
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,128,256,0.002165333264403873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,65536,4096,0.09734133217069839
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,6144,768,0.006233777850866318
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,5120,7168,0.01687111126052009
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,7168,256,0.004214222232500712
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,2560,8192,0.024688889582951862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,128,2560,0.007257777783605788
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,128,7168,0.021027555068333943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,512,2560,0.014858666393491956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,8192,65536,0.4392000039418538
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,512,65536,0.0796711113717821
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,512,512,0.0025271111064487034
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,12288,12288,0.035038222869237266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1536,4096,0.03710577885309855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,10240,16384,0.08406755659315322
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,512,4096,0.02357599967055851
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,5120,768,0.002893333426780171
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,1536,3584,0.02401688860522376
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,16384,16384,3.4377547370062937
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,512,4096,0.031104889180925157
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,6144,512,0.02163733376397027
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,5120,16384,0.049384001228544444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,10240,6144,0.04807466599676344
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,7168,256,0.0052737775776121355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,32,256,0.005223999834722943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2048,512,0.0059004442559348206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,10240,8192,0.022583999567561682
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,7168,65536,1.7313138114081488
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,8192,2560,0.02481866710715824
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,8192,256,0.010416888528399997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2560,6144,0.01888444523016612
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,10240,8192,0.022543110781245764
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,1536,1536,0.006246222390068903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,768,8192,0.017664000391960144
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,12288,128,0.006634666687912411
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,128,3072,0.004878222114510006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,10240,128,0.0022026666750510535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,2048,5120,0.008700444466537898
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,10240,256,0.006263111200597551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,1024,8192,0.01333866682317522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,16384,8192,0.048469334840774536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,16384,16384,0.0754222207599216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,768,16384,0.03137600090768602
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,1536,1024,0.014105778601434497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3072,64,0.0059004442559348206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,128,32,0.0018808889306253856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1024,3072,0.007423111134105259
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,16384,2048,0.1197928852505154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,64,65536,0.009344889058007134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,10240,128,0.005984000033802456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,768,16384,0.027868443065219458
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,128,65536,0.09640622138977051
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,7168,4096,0.015572445260153877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,8192,2560,0.010187555518415239
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,5120,64,0.005672889037264719
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,8192,64,0.004067555483844546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1024,512,0.002969777832428614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,7168,12288,0.06471111377080281
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,12288,6144,0.02759733299414317
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1536,5120,0.04430311255984836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,7168,3584,0.013613333304723104
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,5120,768,0.005640000104904175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,256,10240,0.020241777102152508
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,4096,7168,0.1470817724863688
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,8192,12288,0.04010933306482103
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,2560,2560,0.006647999915811751
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,65536,2048,0.02923111120859782
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,768,1536,0.0038684445122877755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2048,32,0.0028320000403457214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,5120,2560,0.02328088879585266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,64,12288,0.014495111174053617
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,2560,5120,0.025106666816605464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,3072,6144,0.010514666636784872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,7168,6144,0.011865778101815118
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,65536,8192,0.14261333147684732
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3584,16384,0.023344000180562336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,32,32,0.006671999891599019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,512,256,0.0052897776994440295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,32,10240,0.010782221953074137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,128,6144,0.0053377776510185665
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,12288,64,0.006329777754015393
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2048,128,0.0021537778278191886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,8192,10240,0.3931404484642877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,6144,512,0.0032879999942249725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,8192,2560,0.012638222012254926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,4096,768,0.009352888498041365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,10240,2560,0.011830222275522021
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,1024,3584,0.020241777102152508
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,8192,3584,0.01127822200457255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,10240,768,0.006583999842405319
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,128,5120,0.008307555483447181
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,65536,10240,0.09001421928405762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1024,2560,0.00702400008837382
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,6144,1536,0.006286222073766921
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2560,32,0.0022337778160969415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,6144,32,0.002891555635465516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2560,3584,0.008101333346631791
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2048,6144,0.010535111029942831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,1024,3072,0.008378666308191087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,16384,8192,0.41354754236009383
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,1536,12288,0.03858222232924567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1536,3584,0.018011555075645447
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,4096,4096,0.015639111399650574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3584,1024,0.0032986667421129015
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,256,256,0.003568888952334722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,7168,128,0.006735999965005451
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3584,32,0.0028479999552170434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,1024,10240,0.026492445005310908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,8192,512,0.05423733260896471
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,65536,3584,0.09304710891511704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1536,12288,0.028377778000301782
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,3584,6144,0.03621777892112732
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,768,768,0.005674666828579373
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2048,128,0.006956444846259222
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,7168,2560,0.011824888487656912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3072,32,0.002255111104912228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,256,3584,0.027423111928833857
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,32,7168,0.009704888694816166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,128,7168,0.015265777707099915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1024,2048,0.006095110956165526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,6144,16384,0.024879111184014216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,512,3072,0.005578666511509154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2560,6144,0.01662488944000668
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,32,10240,0.005326222214433882
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,7168,3584,0.008729777402347988
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,128,4096,0.007657777931955125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,6144,65536,0.1299146678712633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,4096,6144,0.11187556054857041
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,6144,1024,0.003918222255176968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,4096,512,0.005072000126043956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3072,6144,0.012068444656001197
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,5120,32,0.01275111072593265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,4096,10240,0.024863999750879075
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,10240,8192,0.4981839921739366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1024,16384,0.03436355458365546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,10240,2048,0.011981333295504252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2048,1536,0.020411555965741474
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,10240,65536,0.24536355336507162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,4096,1024,0.00591555568906996
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,12288,64,0.006623111251327727
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2560,2560,0.007368889119890001
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,5120,128,0.014446222119861178
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3072,6144,0.01590044465329912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,128,1536,0.014119111829333834
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,65536,6144,0.30314577950371635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,512,512,0.0028284444577164124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2048,6144,0.01702577703528934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,768,10240,0.015793777174419828
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,8192,3072,0.026837334036827087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,128,8192,0.018664000762833487
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,256,7168,0.009986666341622671
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,65536,3072,0.032291554742389254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,65536,256,0.027154665854242113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,10240,12288,1.055060492621528
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,12288,1024,0.04105066590838962
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,512,2560,0.004886222382386525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,768,12288,0.032812443044450544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,2560,2048,0.01689511040846507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,5120,256,0.018253333038753934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,65536,1024,0.017248888810475666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3072,32,0.003152888889114062
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,2560,2560,0.009342222577995723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,3584,3584,0.007471111085679796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,256,512,0.006294222341643439
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3584,16384,0.02479377720091078
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,128,1536,0.012392889294359418
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,128,3584,0.004902222090297275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,5120,16384,0.0233288887474272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,512,65536,0.052892443206575185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,65536,3072,0.1166764497756958
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1536,16384,0.020576889316240948
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,768,10240,0.0871279968155755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,65536,5120,6.181880103217231
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,10240,3584,0.017180444465743173
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,6144,3584,0.015003555350833468
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,64,5120,0.009537777966923183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,64,8192,0.01056177748574151
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,8192,1024,0.009376000199053023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,10240,16384,0.04373066624005636
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,8192,10240,0.038110223081376814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2048,10240,0.012435555458068848
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,6144,256,0.002510222295920054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,65536,2048,0.21953511238098145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,1536,3072,0.018238221605618794
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,5120,3072,0.02236533330546485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,12288,4096,0.03131822082731459
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,768,512,0.005927111125654644
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,65536,768,0.48759735955132383
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1536,7168,0.00908088849650489
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,1024,2560,0.019254222512245178
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,65536,5120,0.04457777738571167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,768,4096,0.00566133318675889
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,512,512,0.005564444594913059
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,3584,1536,0.01940444442960951
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,65536,768,0.2417893409729004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,256,3584,0.005649777750174205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3072,3072,0.00868800034125646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,5120,10240,0.17172000143263075
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,10240,1024,0.011685333318180509
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,16384,3584,0.016369778248998854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,6144,512,0.0029066666546795103
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,128,64,0.006203555398517185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,256,1024,0.014172444740931192
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,768,10240,0.008384000096056197
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,4096,3584,0.00814844419558843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,7168,7168,0.019448889626397025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2560,1024,0.005995555470387141
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,1536,5120,0.008416888614495596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3072,768,0.005747555444637935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,512,2048,0.006612444503439798
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,1536,1536,0.01649777756796943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,1536,3584,0.0075502221783002215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,768,2048,0.006856888946559694
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,64,10240,0.005927111125654644
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3072,65536,0.06187911166085137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,7168,4096,0.020999999509917364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,8192,4096,0.013200889031092325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,32,128,0.0053377776510185665
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,4096,7168,0.013487111363146039
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,5120,128,0.0024666666156715816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2560,1024,0.0031893334041039148
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2048,2048,0.006595555692911148
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,7168,5120,0.013482666677898832
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,512,65536,0.02257333364751604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,10240,512,0.006278222219811545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,256,5120,0.0069351109365622205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,128,3072,0.010751999914646149
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,32,128,0.002206222257680363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,8192,16384,0.0878480010562473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,768,512,0.012737777498033313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3584,6144,0.1574666632546319
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,2560,1024,0.01609599921438429
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3072,768,0.00628977765639623
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,5120,2048,0.006680888848172293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,768,65536,0.11629954973856609
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1024,3072,0.011413333316644033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2048,7168,0.011928000383906893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,16384,5120,0.033752888441085815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,768,2048,0.014807111687130399
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,4096,65536,0.08061599731445312
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,128,2048,0.006688888702127669
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,1024,256,0.0053297777970631914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,6144,8192,0.02037244372897678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,5120,7168,0.02827022141880459
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,768,1024,0.014182221558358936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,64,8192,0.005909333212508096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,7168,16384,0.03944177760018243
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,10240,1536,0.009712888962692684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,128,8192,0.00565155554148886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,10240,3584,0.06083910995059543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,1024,512,0.012812444733248817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,128,4096,0.010113777385817634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,10240,4096,0.05485777722464668
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,32,64,0.0018337778747081757
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,4096,256,0.016185777054892648
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,768,1536,0.003912888881233004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,64,4096,0.004628444297446145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,32,2048,0.006335999816656113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2048,8192,0.07572888665729098
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,65536,6144,0.24444622463650176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,8192,768,0.01756533318095737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,768,512,0.01273333364062839
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3072,3584,0.01197866681549284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,10240,6144,0.018466666340827942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3072,64,0.012887111140622033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1024,2048,0.004558222161398994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,12288,16384,0.06123289134767321
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,1024,65536,0.06258489025963677
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,16384,128,0.006636444479227066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,128,65536,0.04721422327889336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,12288,10240,0.026886221435334947
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2048,12288,0.031173331869973078
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3072,128,0.002472888885272874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,1024,16384,0.020624000165197585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,16384,768,0.033942222595214844
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,5120,12288,0.29497512181599933
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1024,1536,0.006312888943486744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,16384,4096,0.023188443647490606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,8192,10240,0.021702221698231165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1536,7168,0.009827555881606208
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3072,1024,0.007679999702506595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,128,5120,0.031113776895735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3584,768,0.0033324443631701996
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,65536,256,0.02685600022474925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,256,3584,0.00563733321097162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,8192,64,0.005912000106440649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3584,6144,0.010773333410422007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,3072,10240,0.28097955385843915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,12288,5120,0.021470222208235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,16384,3584,0.17097866535186768
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,16384,512,0.02233955595228407
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,2048,16384,0.025966222087542217
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,128,2560,0.009351111120647853
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,5120,7168,0.14385955863528782
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,128,6144,0.016311999824311998
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,768,128,0.005596444424655702
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2048,3072,0.05788533555136787
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,8192,5120,0.020263110597928364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,7168,4096,0.022649778260125056
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,7168,4096,0.03259377678235372
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,64,1024,0.005967111223273807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,4096,3072,0.00777688870827357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,10240,3584,0.013479111095269522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,64,8192,0.007648000286685095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,768,5120,0.00832444429397583
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2048,5120,0.028897778855429754
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,8192,6144,0.017680888374646504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,12288,128,0.00904444439543618
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,1536,3072,0.009317333499590555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,6144,8192,0.3431137667761908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,3584,2560,0.01716977854569753
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,2560,6144,0.02128711177243127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1024,1024,0.0038977778620190094
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,65536,768,0.021520889467663232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,64,256,0.002238222294383579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,10240,768,0.006909333169460297
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,6144,1536,0.02167644434505039
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,16384,65536,0.3131457699669732
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2048,7168,0.019345778557989333
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,8192,5120,0.028036445379257202
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,2048,256,0.005420444326268301
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3584,2560,0.007302222152551015
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,2048,16384,0.020593777298927307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,4096,256,0.002509333400262727
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,768,7168,0.01039644413524204
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,12288,2048,0.011759111450778114
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,128,7168,0.009684444301658207
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1536,1536,0.013433777623706393
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3584,768,0.006264888991912206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2048,65536,0.058908442656199135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,1536,3072,0.007280000381999546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,16384,768,0.007999999655617608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,32,8192,0.010458666417333813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,768,6144,0.03861688905292087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,10240,512,0.006259555617968242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,768,1536,0.004380444271696938
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,6144,64,0.007647111184067196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1024,10240,0.015442666080262927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,1024,128,0.00526311124364535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,65536,64,0.01070577816830741
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,12288,8192,0.032418668270111084
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,256,65536,0.017703110973040264
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,7168,10240,0.026965333355797663
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2048,10240,0.012337778177526263
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,2048,2560,0.02128177715672387
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,256,256,0.01274577776590983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,5120,5120,0.25021688143412274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,512,12288,0.013419555293189155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,8192,128,0.020950223008791607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,10240,2560,0.016000888413853116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,65536,512,0.012855110896958245
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,2560,512,0.014151111245155334
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,768,5120,0.006353777729802662
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,512,5120,0.020049777295854356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,128,7168,0.00960888879166709
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,65536,6144,0.09194933043585883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,8192,768,0.018972444865438674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,10240,64,0.0031884445084465873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,64,3584,0.006221333311663733
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,512,32,0.00555288874440723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,768,6144,0.008183110919263627
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,16384,12288,0.06657155354817708
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,5120,256,0.016560888952679105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,65536,4096,0.08505422539181179
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,8192,4096,0.015150222513410779
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,64,2048,0.009347555538018545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,768,2048,0.006730666591061487
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,2048,3584,0.03808088766203986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2560,8192,0.02003466586271922
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,12288,512,0.006480000085300869
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2048,3584,0.012622222304344177
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,4096,256,0.01520088811715444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,768,3584,0.005250666704442766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,1536,2560,0.06409866942299737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,256,768,0.0028479999552170434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1024,32,0.005618666609128316
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1536,1024,0.007252444823582967
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,64,64,0.0018159999615616268
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,1024,128,0.014783110883500842
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,5120,2560,0.0069306666652361555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,3072,2048,0.016189333465364244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,1024,3072,0.007305777735180325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,16384,32,0.005997333261701796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,512,10240,0.012463111016485425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,3584,1024,0.03273511264059279
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1536,256,0.002412444394495752
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,12288,32,0.00703999979628457
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,32,2048,0.00666222224632899
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1024,768,0.0031893334041039148
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,65536,4096,0.10064178042941624
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3072,16384,0.03368799885114034
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,128,3072,0.006051555689838197
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,512,768,0.005942222144868638
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,6144,65536,0.078402664926317
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,32,6144,0.03037422233157688
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,65536,4096,0.060580445660485156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2048,4096,0.008091555701361762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,5120,6144,0.014583999911944071
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,7168,65536,0.16342133945888943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3584,3072,0.007611555357774098
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,12288,7168,0.0739226672384474
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3584,256,0.014831110835075378
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,16384,256,0.0047146665553251905
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,4096,512,0.014450665977266101
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,5120,8192,0.04641244477695889
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1536,5120,0.00792266676823298
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,5120,64,0.003918222255176968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,768,16384,0.010046222143703038
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,3072,6144,0.009026666482289633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,128,10240,0.006603555546866522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,1024,10240,0.020311110549502902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,256,3072,0.00741244438621733
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,6144,3584,0.029504888587527808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,1024,4096,0.0076657773719893555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,10240,1536,0.022064000368118286
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,2048,10240,0.024477332830429077
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,7168,6144,0.1957768864101834
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,5120,4096,0.014498665928840637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,32,16384,0.020397333635224234
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,768,256,0.02271288964483473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,7168,6144,0.03144888745413886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,4096,512,0.005622222191757626
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3072,3584,0.007726222276687622
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,3584,10240,0.026661333110597398
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,16384,65536,0.17943999502393934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1536,3072,0.01015022231472863
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,65536,7168,0.2096231116188897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,1024,768,0.015525332755512662
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,1536,2560,0.015857777661747403
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,8192,128,0.005296000176005893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,65536,2048,0.07129955291748047
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,512,1024,0.005948444621430502
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,7168,12288,0.03139288889037238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3072,10240,0.013257777525318993
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,768,10240,0.012509333590666452
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,768,1024,0.013450666434235044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,3584,1024,0.04762755499945747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2560,512,0.0034391110142072043
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,512,7168,0.005559999909665849
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3072,768,0.005793777604897817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,8192,1024,0.00628977765639623
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2560,2560,0.007668444679843054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,12288,768,0.05085066623157925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,768,32,0.005647111270162795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,768,2048,0.020426667398876615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,1024,3072,0.019288000133302476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,512,7168,0.009753777749008602
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,2560,512,0.015848888291252982
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,5120,32,0.005315555466545953
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,1024,65536,0.077583114306132
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,5120,16384,0.03948622279696994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,2560,2048,0.0069395556218094295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,3584,10240,0.014856888188256158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,8192,1536,0.006602666858169768
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,8192,6144,0.041407111618253924
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,512,16384,0.015837333268589444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,128,5120,0.01036444471942054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,1024,3072,0.007282666862010956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1024,512,0.006592000110281839
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1536,3072,0.007280000381999546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1024,512,0.0031840000301599503
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,1024,3584,0.00906844437122345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,65536,3584,0.054414222637812294
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,1024,10240,0.02121777832508087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,16384,10240,0.16548089186350504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,7168,1536,0.011379555695586734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,256,8192,0.017516444126764934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1536,10240,0.014750222365061441
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,10240,12288,1.8787182701958551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,768,3072,0.009673777553770278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,3584,3072,0.027757333384619817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2048,65536,0.04566133353445265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,32,16384,0.03559199968973795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,16384,512,0.0075795559419526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1536,2560,0.005195555587609609
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,128,10240,0.025368889172871906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,768,5120,0.020048000746303134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1536,64,0.0021457777669032416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,2048,16384,0.020271110865804885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,7168,128,0.006955555743641323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2560,2048,0.0069715554515520734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,768,2048,0.006601777755551868
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,65536,7168,0.1607324414783054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,16384,7168,0.31956712404886883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,4096,2048,0.024015110399987962
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,128,5120,0.006586666736337874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3584,1024,0.006247111078765657
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,8192,768,0.011448000040319232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1536,256,0.0025031111306614345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,12288,8192,0.09644711017608643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,256,5120,0.008612444831265343
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,8192,512,0.00590133335855272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3584,512,0.003536888708670934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,6144,10240,0.019282667173279654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,32,1536,0.003530666646030214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1536,3072,0.005503111001518037
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,65536,7168,0.31144534216986763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,6144,128,0.020938666330443487
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,768,256,0.0038933331767717996
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,3072,3072,0.007298666569921706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,10240,10240,0.025248888466093276
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2048,256,0.00701422244310379
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,6144,768,0.005924444645643234
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,16384,128,0.08239377869500054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3584,4096,0.009034666750166152
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,5120,16384,0.03014311194419861
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,32,3584,0.004550222307443619
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,5120,6144,0.012506666282812754
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,4096,128,0.0052355556852287715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1024,16384,0.014898666077189975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,2560,12288,0.12599733140733507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,256,512,0.00249422217408816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,4096,64,0.00795377790927887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,7168,64,0.0024906665914588505
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,32,64,0.005615111026499007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,12288,8192,0.05894400013817681
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3072,2560,0.005686222265164058
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,65536,128,0.004914666629499859
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,1536,12288,0.026364444030655756
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3072,6144,0.013199110825856527
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,3584,768,0.00591466658645206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,6144,512,0.01657777859105004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,3584,10240,0.03712088863054911
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,7168,64,0.037081778049468994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,512,16384,0.029891557163662378
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,64,256,0.00250133333934678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,6144,32,0.0052933332820733385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,7168,5120,0.050386667251586914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2048,2048,0.006774222271309958
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2048,6144,0.017906667457686532
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,3584,4096,0.007987555530336168
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,6144,65536,4.840196397569445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2560,16384,0.034929778840806745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,768,5120,0.022318222456508215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,16384,5120,0.032700445916917585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,32,512,0.0025111111915773815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,64,1024,0.005910222315125995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,32,64,0.005647999958859549
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,128,3072,0.004635555462704765
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,1024,2560,0.006959999601046245
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,6144,16384,0.05181866553094652
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,2560,8192,0.11654844548967148
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,16384,512,0.0069875551594628235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,12288,768,0.0070053330726093715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,32,8192,0.005640000104904175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,768,6144,0.009728000395827824
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,8192,768,0.018938667244381376
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,16384,3584,0.01594311164485084
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,128,64,0.0019164443429973389
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1536,4096,0.007999111380841997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,6144,2048,0.008714666797055138
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,7168,3072,0.007780444290902879
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,12288,64,0.0027466666781240036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,12288,65536,0.22537689738803438
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,7168,6144,0.03989244500796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,12288,16384,0.08796088563071357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,65536,12288,1.5579741795857747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,768,1024,0.005717333406209946
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,7168,512,0.01678400072786543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,4096,1536,0.0057964444988303715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,16384,8192,0.10561422506968181
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,6144,3072,0.011247110863526663
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,8192,16384,0.030423998832702637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,2560,7168,0.029838220940695867
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,32,4096,0.011447110937701331
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,1536,1536,0.01457244488928053
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3584,65536,0.06978044244978163
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,512,1536,0.006178666734033161
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,6144,3072,0.008007999923494127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,256,1536,0.003916444463862313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2560,3584,0.07253244188096789
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,64,2560,0.006705777926577463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,32,7168,0.010410666465759277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,7168,1536,0.006590222318967183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1536,7168,0.0344195564587911
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,3072,2560,0.01591999994383918
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,3584,5120,0.03221688999070062
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,3584,2560,0.019225777851210702
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,65536,3072,0.6561511357625326
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,1024,3072,0.02268444498380025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3072,64,0.009688888986905416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,8192,4096,0.03271733389960395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,6144,128,0.014456000592973499
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2560,16384,0.02604088849491543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,65536,128,0.021739555729760066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,5120,5120,0.012095111111799875
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,4096,512,0.005364444520738389
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,128,65536,0.0455022222465939
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,4096,16384,0.02502044373088413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,2560,16384,0.03291644321547614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,2048,65536,0.0636426673995124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,65536,128,0.012788444757461548
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,256,8192,0.00889155517021815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,256,1536,0.006232000059551663
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3584,128,0.005518222020732031
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,512,65536,0.07659910784827338
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2048,1024,0.023559111687872145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2560,1536,0.005602666487296422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,512,3072,0.007265778051482306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,65536,1536,0.03651644455062018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,32,12288,0.004927999857399198
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1536,10240,0.014130666851997375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,64,12288,0.004922666483455234
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3072,12288,0.026292444931136236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,5120,1536,0.04607199960284763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,7168,4096,0.015571556157535978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,1024,4096,0.021319111188252766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,65536,10240,0.10424977540969849
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,2048,10240,0.024413333998786077
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,4096,32,0.0020942222326993942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,512,10240,0.008356444537639618
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,16384,768,0.022152889106008742
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,5120,16384,0.0844586690266927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,128,3072,0.021873777111371357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,3584,1024,0.007968000239796108
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,12288,256,0.006583111153708563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,3072,2048,0.018560889694425795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2048,2560,0.005584888988071018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,6144,768,0.005597333527273602
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,4096,2560,0.00654755574133661
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2560,3072,0.007361777954631382
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,8192,2048,0.00832177781396442
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,12288,512,0.0069724445541699724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,128,256,0.0052284445199701525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,16384,12288,0.19863111442989775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,6144,3072,0.020621332857343886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,512,7168,0.014184888866212634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,4096,128,0.002509333400262727
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,256,12288,0.0185208883550432
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,512,128,0.008403555386596257
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,256,256,0.008366222182909647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,10240,1024,0.005895111295912001
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,512,4096,0.006584888945023219
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,8192,65536,0.6318480173746744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,7168,3072,0.010388444695207808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1536,32,0.005668444352017508
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,12288,3072,0.015266666809717814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,5120,3584,0.02165511084927453
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,256,768,0.00701155596309238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,1024,3072,0.007304888632562425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,32,1024,0.014292443792025248
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,65536,8192,0.07295644283294678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,3584,512,0.012503999802801343
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1024,256,0.002519999941190084
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,768,3072,0.007321777443091075
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,512,10240,0.012072000238630505
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,4096,1536,0.005010666532648934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,8192,768,0.005981333139869902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,12288,256,0.00599911105301645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,6144,10240,0.033766223324669734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,7168,16384,0.026364444030655756
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2048,128,0.006584888945023219
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3072,7168,0.018823110395007663
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,16384,512,0.008708444734414419
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1536,65536,0.2675786548190647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,7168,2560,0.0277493331167433
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,512,3072,0.0052800000541739995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,256,1024,0.005584888988071018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,256,16384,0.020571556356218126
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,512,6144,0.010103999740547603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,768,4096,0.007939555578761632
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,7168,6144,0.03642933236228095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1024,32,0.002404444540540377
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,1024,3072,0.01810933318403032
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,512,16384,0.04526933365397983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,64,3072,0.015088889333936902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,32,8192,0.015284443895022074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,8192,7168,0.020661332541041903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,12288,5120,0.026722666290071275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,768,256,0.005239999956554837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,12288,1024,0.007260444263617198
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,128,65536,0.05400088760587904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,1536,1024,0.008368888662921058
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3584,1536,0.02845066785812378
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,5120,6144,0.014330666926172046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,8192,6144,0.012896888785892062
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,256,16384,0.02180355621708764
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,65536,4096,0.06650133265389337
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1536,64,0.004938666605287128
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,3584,128,0.01519555515713162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,4096,12288,0.04717777834998237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3584,8192,0.01479644411140018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,64,768,0.0041146667467223275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,6144,128,0.017892445127169292
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,5120,8192,0.031750222047170006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,7168,64,0.005752888818581899
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2048,10240,0.05440622236993578
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,6144,128,0.005623999983072281
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3072,65536,0.06438666582107544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,128,256,0.006001777946949005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3072,65536,0.22193155023786756
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,2048,7168,0.021504888931910198
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3584,6144,0.01310222182008955
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,768,1024,0.0057502223385704895
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1536,3072,0.00756177802880605
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,8192,6144,0.021386666430367365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3584,65536,3.9605305989583335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,512,7168,0.0354968872335222
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3584,3072,0.007105777661005656
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,1536,2560,0.007311111523045435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,1536,4096,0.016939555605252583
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,128,7168,0.009755555954244401
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,512,2048,0.01514844430817498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,65536,10240,0.3180924521552192
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,10240,10240,0.30030756526523167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,64,1024,0.00630044440428416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,768,10240,0.11359288957383896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3584,768,0.00462666650613149
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,16384,32,0.004954666727119022
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,256,768,0.003175111073586676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,768,7168,0.008980444735950893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2048,2048,0.006950221955776215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,65536,768,0.031176000833511353
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,256,64,0.006319111006127463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,5120,8192,0.02869510981771681
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2560,8192,0.02337155573897892
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,4096,65536,0.1831822262869941
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,65536,128,0.009739555418491364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,512,32,0.002061333300338851
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,65536,32,0.018838221828142803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2048,32,0.009339555270142024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,2560,65536,0.3609395556979709
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,4096,1536,0.006292444550328785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,4096,3584,0.1342702176835802
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,12288,2048,0.012447111308574677
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,512,128,0.005599110904667113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,512,7168,0.00562755556570159
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,16384,32,0.012310222619109683
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,128,65536,0.010370666782061258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,256,3584,0.015219555960761176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,7168,3072,0.051613334152433604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,3584,8192,0.02957333458794488
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2048,16384,0.01982311076588101
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,7168,5120,0.010808889236715106
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,5120,3584,0.007328888608349695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,768,1536,0.014166222678290473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,65536,128,0.019517333971129525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1536,2560,0.00556355549229516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,5120,768,0.007344000041484833
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,3584,512,0.009688888986905416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,128,512,0.006616888774765863
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,7168,6144,0.015875554747051664
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,4096,10240,0.03279377685652839
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2048,256,0.0035128887328836652
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1024,2048,0.007287999822033777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3072,1536,0.006272888845867581
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3584,7168,0.030221333106358845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,12288,10240,0.36940977308485246
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,65536,16384,0.16550221708085802
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,768,2560,0.007314666277832455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3072,64,0.006623111251327727
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,64,768,0.0029297777348094513
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,4096,8192,0.1437493297788832
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,2048,10240,0.014904888139830695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,256,2560,0.00481244424978892
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2048,1024,0.006308444258239534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,8192,65536,0.41058932410346133
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,3584,10240,0.01516355574131012
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,5120,64,0.008049777812427944
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1024,7168,0.026170666019121807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3072,12288,0.01627288924323188
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2048,65536,0.10886310868793064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,128,4096,0.008402666283978356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2048,1536,0.006299555715587404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,64,256,0.002201777779393726
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,8192,12288,0.025736888249715168
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,7168,7168,0.16086667113833958
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,6144,5120,0.011715555356608497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,4096,10240,0.016240888171725802
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,7168,65536,0.0749066670735677
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,3584,8192,0.13464266724056667
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1024,128,0.002181333386235767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1024,32,0.005719999886221356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,64,65536,0.006695111178689533
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2048,1024,0.0035866668654812705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,12288,2048,0.028360890017615423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,4096,32,0.006045333213276333
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,6144,3584,0.009762666291660732
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,7168,64,0.006486222147941589
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,512,8192,0.011164444188276926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,1536,6144,0.011776000261306763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,128,8192,0.026872888207435608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,64,16384,0.030975110001034204
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,5120,12288,0.018801778554916382
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,64,64,0.00563822231358952
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,1536,768,0.005573333137565189
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,6144,2048,0.009382222261693742
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,2560,256,0.014984000060293408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,5120,3072,0.01715377800994449
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2048,1536,0.006715555571847492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2048,256,0.006467555546098285
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,12288,2560,0.01109955542617374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,512,3072,0.007296889192528195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,6144,10240,0.030235555436876085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,10240,16384,0.08883733219570583
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,1536,3072,0.02029600077205234
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,32,2560,0.006983111302057902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,6144,6144,0.011417778001891242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2048,1536,0.00905688852071762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,2048,65536,0.08368533187442356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,10240,1024,0.006629333313968446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,12288,4096,0.02166044380929735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,1536,3072,0.007343110938866933
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,4096,2560,0.00666222224632899
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,2048,5120,0.0288426677385966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,768,10240,0.023689778314696416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,2048,128,0.0052560000783867305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,128,8192,0.01941333380010393
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,16384,8192,0.031025777260462444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,16384,7168,0.027280888623661462
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,768,65536,0.07358311282263862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,256,2048,0.004655999855862724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,768,4096,0.017421333326233756
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,4096,8192,0.016608888904253643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,128,16384,0.008839999636014303
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,10240,2048,0.00980799976322386
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2048,5120,0.05446222093370226
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,6144,4096,0.03656355539957682
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2560,16384,0.017312000195185345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1024,768,0.003592888928121991
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,12288,65536,0.1917795605129666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,4096,4096,0.01205599970287747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,256,3584,0.014861333701345654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,16384,4096,0.036836445331573486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,16384,256,0.018409777018758986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,3584,768,0.005681777579916849
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,512,256,0.006953777538405524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,64,3584,0.028596444262398615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,512,32,0.002238222294383579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1536,4096,0.008387555678685507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,2048,2560,0.02250755495495266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,256,10240,0.017639999588330586
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,512,2048,0.013835555149449242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,65536,12288,0.10699999994701809
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,512,768,0.0028684443483750024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,2560,128,0.014696889453464083
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2560,10240,0.012006222373909421
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,768,5120,0.02244088881545597
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,5120,12288,0.023631110787391663
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,32,12288,0.0052488889131281115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,2048,4096,0.02772088845570882
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,128,3072,0.007625777688291337
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,10240,8192,0.022230222821235657
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3072,8192,0.022270222504933674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,256,65536,0.11940622329711914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,768,1536,0.006022222340106964
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,10240,12288,0.06614310873879327
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1536,7168,0.01038755559258991
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,512,3584,0.00599644457300504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,5120,768,0.005994666781690385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,768,2560,0.006672888994216919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,3072,256,0.02315555512905121
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,1536,10240,0.0229413327243593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3584,768,0.0035626664757728577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,3584,2048,0.0075039996041191956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,5120,1536,0.0052728888889153796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,1536,6144,0.04600444436073303
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,512,128,0.0021911110315057966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,512,64,0.001993777851263682
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,512,10240,0.023026666707462732
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,8192,12288,0.046839111381106906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,1536,1024,0.013557333085272046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,7168,6144,0.014463999205165438
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,6144,4096,0.011057777537239922
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,8192,6144,0.041329777903027005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,256,1536,0.01239822225438224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,6144,3072,0.022295110755496558
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,16384,3584,0.03808888792991638
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1536,128,0.0022035555707083806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3072,16384,0.2416435612572564
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2560,128,0.0031848889258172777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,1024,2560,0.008000888758235509
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,6144,65536,0.07436266872617933
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,10240,256,0.01995822290579478
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,4096,1536,0.0064399999876817065
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,256,12288,0.032481776343451604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,3072,1536,0.01683288812637329
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,768,1536,0.007308444215191736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,5120,6144,0.014209777116775513
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,256,3072,0.007289778027269576
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,2560,65536,0.1339413324991862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2048,3584,0.006678222368160884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,1536,10240,0.03195555673705207
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,6144,65536,0.12941600216759577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,7168,1536,0.10579644309149848
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,64,8192,0.005585777676767773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2048,16384,0.02197955548763275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,16384,2048,0.0120319997270902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1024,3584,0.006062222023804982
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,768,10240,0.0646159980032179
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,32,16384,0.029230223761664495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,3584,6144,0.02309155629740821
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,512,1024,0.012812444733248817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,64,10240,0.01181244436237547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,768,1024,0.0068764446510208984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,1024,3072,0.06842311223347981
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,10240,16384,0.03781955440839132
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,4096,2048,0.03767733441458808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1024,1536,0.006627555522653792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1536,256,0.005872000008821487
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,7168,4096,0.09976266490088569
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1536,768,0.016760889026853774
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,12288,768,0.007085333267847697
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,512,6144,0.006711111300521427
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,1024,12288,0.028619554307725694
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,10240,5120,0.15604266855451795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,2560,256,0.015797333584891427
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,12288,8192,0.026527111728986103
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,65536,64,0.01204533295498954
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,16384,128,0.010379555324713389
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,12288,10240,0.036947555012173124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2560,512,0.002537777854336633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,5120,2048,0.008753777378135258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,12288,65536,0.33067109849717885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,10240,8192,0.0722222195731269
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,1024,8192,0.010399999717871347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,5120,12288,0.017455110947291057
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,5120,512,0.014471999473041959
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,6144,65536,0.1781466669506497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,256,64,0.0052720002002186244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,7168,12288,0.03789955708715651
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,512,2048,0.006326222171386083
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,3072,256,0.010552888943089379
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,128,3072,0.007656888829337225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,128,128,0.0021751111166344746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,3584,768,0.005683555371231503
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,5120,6144,0.012291555603345236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,8192,6144,0.047882666190465294
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,4096,256,0.0052693333062860705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3072,16384,0.023034666975339253
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,512,3072,0.007317333585686154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,65536,64,0.08645155694749619
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,2560,256,0.014878221684032016
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1024,65536,0.10534933540556167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,10240,3072,0.015309333801269531
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,1536,128,0.0052871112194326185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,6144,2048,0.023545778459972803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1536,64,0.0022008888837363985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,2560,256,0.015220445063379077
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,1536,768,0.005608888963858287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,256,3072,0.014462222655614218
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,10240,256,0.01759022143152025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1536,5120,0.07850755585564508
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,7168,256,0.02381244467364417
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,8192,1536,0.02130399975511763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,512,4096,0.00904888908068339
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,1536,256,0.0053173332578606075
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,128,2560,0.013468444347381592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,16384,512,0.009731555150614845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,65536,768,0.013872000078360239
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,16384,65536,0.8488746749030219
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,3072,768,0.017687110437287223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,65536,512,0.06557866599824694
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,12288,512,0.010210666391584609
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,5120,6144,0.011057777537239922
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,7168,16384,0.03883644607332017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2048,5120,0.02847822176085578
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3584,65536,0.10645866394042969
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,3584,16384,0.022263111339675054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,8192,1024,0.007836444510353936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1536,3072,0.04978488882382711
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3072,8192,0.013991999957296582
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,32,2560,0.0069191112286514705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,65536,1536,1.750276459587945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,12288,2560,0.012920888761679331
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,3072,3584,0.06224444177415636
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,768,6144,0.01573777695496877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,65536,32,0.004204444587230682
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,1536,12288,0.03835733400450812
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2560,4096,0.013337777720557319
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,16384,512,0.0046302220887608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,32,256,0.0063457778758472866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,1536,512,0.007348444726732042
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,7168,2560,0.009324444664849175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,128,4096,0.007693332930405934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2560,1024,0.009770666559537252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,12288,3072,0.03573866685231527
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,768,768,0.00319377767542998
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,64,32,0.0020986667109860317
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,1024,3072,0.007332444190979004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2560,2560,0.011071110765139261
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,12288,2560,0.02646044393380483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,6144,5120,0.01275288893116845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,1536,128,0.019323555959595572
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,768,128,0.0052968888647026485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,2048,768,0.017250667015711468
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1536,3584,0.006280888699822956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,7168,4096,0.033003555403815374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,3584,2560,0.023018666439586218
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,5120,8192,0.04601422283384535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,8192,128,0.020121778051058452
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,5120,10240,0.017603556315104168
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,10240,1024,0.00742133292886946
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,2560,512,0.016388444436921012
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,768,8192,0.026770666241645813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,10240,128,0.054786665572060474
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,256,6144,0.01650311052799225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2560,8192,0.010842666857772403
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,4096,65536,0.10513244734870063
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,768,7168,0.04483111037148369
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1536,10240,0.15731199582417807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1536,2048,0.004650666481918759
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,5120,3584,0.013732444081041547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,65536,5120,0.07726933558781941
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1024,128,0.0021626665774318907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,5120,3072,0.009320889082219865
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2048,512,0.0028275555620590844
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,1024,12288,0.013144888811641268
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,32,3584,0.007663999994595845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,7168,3072,0.08821244372261895
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2048,12288,0.11961244212256537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,12288,3584,0.0214355554845598
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,7168,65536,0.3307110998365614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,12288,2048,0.011379555695586734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,12288,12288,0.048954665660858154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,128,2048,0.006949333681000604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,3584,256,0.016855110724767048
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,768,5120,0.017186666528383892
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,5120,10240,0.030730668041441176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,4096,64,0.00221422231859631
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,3072,16384,0.061977777216169566
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,16384,1536,0.019687111179033916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,8192,5120,0.03766311208407084
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,3584,512,0.005551110953092575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,16384,10240,0.06436800294452243
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,128,3584,0.007963555554548899
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,1536,12288,0.0346684455871582
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2048,512,0.0038906666967603895
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,2048,3072,0.007290666302045186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,12288,12288,0.04102844330999587
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,4096,128,0.005961777849329843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,256,7168,0.017813333206706576
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,4096,10240,0.01809866726398468
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2048,2560,0.004934222333961063
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,16384,1536,0.010847999817795224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,10240,256,0.016875555117925007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,2048,512,0.005565333283609814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,512,65536,0.074599113729265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,768,256,0.00833777752187517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3584,4096,0.01037333326207267
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,128,1536,0.006276444428496891
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,6144,64,0.005879999862776862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,65536,512,0.028600000672870215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2048,1024,0.006270222365856171
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,4096,1024,0.0076657773719893555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,128,16384,0.0314666661951277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,5120,2560,0.007213333414660559
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1536,2048,0.00452533322903845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,128,4096,0.008303110798199972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,4096,16384,0.08007022407319811
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,16384,1024,0.011794666449228922
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,65536,5120,0.04675999946064419
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,32,8192,0.010393777655230628
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2560,16384,0.3179697725507948
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,8192,65536,0.08593244685067071
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,2048,128,0.005617777920431561
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2048,512,0.004046222401989831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,5120,4096,0.010053333308961656
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2048,16384,0.022517333428064983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2048,768,0.0032231110251612137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,2560,12288,0.016705777910020616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,64,64,0.00203822222020891
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,16384,5120,0.018213333355055917
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,128,1536,0.006325333482689328
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,8192,5120,0.03762844536039565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1024,6144,0.009720000127951304
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2048,256,0.005706666658322017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,6144,5120,0.011542222566074796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,2048,12288,0.015282667345470853
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,3072,512,0.014127110441525778
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,2048,3072,0.08467111322614883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,256,2048,0.006711111300521427
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,2560,768,0.016593777471118502
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,7168,65536,0.08485688765843709
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,2560,768,0.01796800063716041
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,4096,768,0.038047999143600464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,7168,768,0.009004444711738164
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,4096,2560,0.006801777829726537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,2048,2048,0.006617777877383762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,7168,128,0.005404444618357553
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,5120,12288,0.036413331826527916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,64,6144,0.0052933332820733385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,10240,5120,0.02345688806639777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2048,16384,0.04059733284844293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,1536,512,0.014138667119873894
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,2560,1024,0.014892444014549255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,768,10240,0.012481778032249875
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,8192,3072,0.016812443733215332
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,7168,3072,0.15996532970004612
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,32,5120,0.009329777624871995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,6144,4096,0.012541333006487953
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,7168,10240,0.3408364454905192
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,7168,5120,0.034640888373057045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,8192,6144,0.30383022626241046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,2560,3584,0.023657777243190344
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,768,65536,0.057224889596303306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,1536,7168,0.020788444413079154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,12288,1536,0.04283644424544441
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,768,16384,0.05683910846710205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3072,10240,0.01348444405529234
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3584,3584,0.008637333081828224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,4096,3584,0.010954666468832227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,1536,7168,0.03566933340496487
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,2048,1024,0.014089778065681458
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,128,7168,0.007772444850868649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,7168,128,0.016203555795881484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,1024,8192,0.02885777751604716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,7168,128,0.0021546667234765156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,10240,5120,0.01828177769978841
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,128,4096,0.017158221867349412
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,512,8192,0.021690666675567627
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2560,6144,0.02630666726165348
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,16384,3584,0.01605866683853997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,3072,1024,0.007312888900438945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,12288,12288,0.04881244566705492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,768,5120,0.009015999734401703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,8192,2560,0.022988445229000513
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,1536,6144,0.052301334010230176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,10240,10240,0.05975111325581869
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,12288,8192,0.0502017769548628
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,768,3072,0.011734222372372946
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,4096,2048,0.0069751110341813825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,16384,1024,0.030994666947258845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,768,10240,0.025604445073339675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,3584,2048,0.02143200072977278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,128,8192,0.010817777779367236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,512,16384,0.008167111211352879
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,64,32,0.0018764444523387486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,12288,2048,0.011948444777064852
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,6144,32,0.00592533333433999
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,2560,7168,0.01109155515829722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,64,16384,0.00563822231358952
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,2048,128,0.005341333233647876
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,256,8192,0.012099555797047086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3584,4096,0.01367022262679206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,256,768,0.0057848890622456866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,65536,512,0.022065778573354084
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,8192,65536,0.15251999431186253
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,16384,2560,0.014744000302420722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,2560,6144,0.009696889254781935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,5120,1536,0.00629333323902554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,10240,1536,0.007721777591440413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,768,3072,0.016192888220151264
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,4096,2560,0.007250666618347168
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,6144,768,0.018274666534529794
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3584,512,0.005892444401979446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,768,8192,0.01086400035354826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,1536,16384,0.041478223270840116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,1536,16384,0.019380443625979953
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,10240,3072,0.032263110081354775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,5120,16384,0.02331200076474084
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,7168,1024,0.005894222193294101
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,8192,65536,0.15152355035146078
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,2048,2048,0.026679111851586237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,6144,1536,0.006286222073766921
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,16384,65536,0.325492435031467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1024,12288,0.06412355767356025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,5120,10240,0.020576889316240948
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,5120,6144,0.014115555418862237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,64,1024,0.0059279998143513995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,2560,768,0.005584888988071018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,1024,16384,0.016135111451148987
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1536,8192,0.07253333595063952
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,16384,512,0.022143999735514324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,2560,3072,0.03574488891495599
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,7168,7168,0.2369004355536567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,64,8192,0.00555377784702513
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,4096,3584,0.007277333074145847
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,128,1024,0.007613333563009898
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,16384,6144,0.043613334496816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,128,1024,0.012073777616024017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,3584,12288,0.029712888929578993
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,768,128,0.014077333940400017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1536,10240,0.04491555690765381
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,7168,2560,0.07002311282687716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,3584,2048,0.006620444357395172
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,6144,1024,0.020263999700546265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,768,16384,0.033237334754731916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,64,768,0.0029440000653266907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,1536,8192,0.011741333537631564
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,10240,3072,0.012498666842778524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,512,6144,0.009651555783218807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,256,1536,0.006722666737106111
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,512,1024,0.013032000097963544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,2048,3072,0.021320889393488567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,512,16384,0.015863999724388123
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,512,3584,0.009711999860074785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1536,65536,0.057725330193837486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,2560,128,0.005615111026499007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,512,256,0.002504000026318762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,10240,5120,0.020052444603708055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3072,1024,0.004274666723277834
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,3072,256,0.014817777607176038
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,8192,1024,0.0063884444534778595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,768,128,0.002222222172551685
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,16384,768,0.007387555307812161
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,16384,32,0.007065777977307637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,512,10240,0.0069884442620807225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1536,3584,0.007650666766696506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,768,1024,0.0059573331640826324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,2048,4096,0.10969689157274033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,64,2560,0.005239999956554837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,256,32,0.002191999927163124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2560,1024,0.006181333214044571
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,128,1536,0.003544888976547453
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,6144,2048,0.009681777821646797
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,256,5120,0.016576889488432143
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,4096,10240,0.016607999801635742
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,65536,6144,3.1035003662109375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,3584,16384,0.03792000148031447
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,768,65536,0.0643084446589152
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,1024,10240,0.041272001134024724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,12288,10240,0.057935112052493624
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,6144,16384,0.03622666663593716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,10240,512,0.005981333139869902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,1024,1536,0.01620533400111728
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3584,3584,0.009910222556855943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,3584,1536,0.006148444281684027
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,3584,16384,0.0436719987127516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2048,64,0.00563822231358952
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,8192,2560,0.008536888493431939
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,4096,512,0.013472889032628802
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,8192,1024,0.005582222094138463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,64,1536,0.003624888757864634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,2560,3584,0.007721777591440413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,1536,256,0.015098666151364645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,128,6144,0.005940444353553984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,512,12288,0.03847644395298428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,256,512,0.0028124445428450904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,6144,2048,0.007347555624114142
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,16384,7168,0.7057030995686849
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,1536,768,0.02550933261712392
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,64,65536,0.014814222852389017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2560,3072,0.012435555458068848
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,16384,256,0.003149333306484752
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,12288,768,0.05379466546906365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,128,12288,0.031028442912631567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,12288,1024,0.031222220924165513
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,10240,3072,0.013566222455766467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1536,32,0.0022275555464956495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,5120,7168,0.01570755574438307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,4096,512,0.015531554818153381
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,16384,64,0.003956444561481476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,1024,4096,0.019347555107540555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,5120,2048,0.07829688655005561
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,65536,10240,0.4338017834557428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,1536,1536,0.006609777609507243
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,7168,5120,0.029186666011810303
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3072,32,0.006274666637182236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,128,3072,0.056194669670528836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,64,1536,0.004230222354332606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,5120,256,0.005585777676767773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,128,8192,0.020021332634819877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2560,2048,0.004926222066084544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,32,8192,0.00444444434510337
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,6144,2560,0.1085653305053711
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,6144,3072,0.013974222871992322
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1536,64,0.002242666565709644
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,4096,2048,0.07964533567428589
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,1024,6144,0.010422222316265106
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,2560,3072,0.016280000408490498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,256,8192,0.018806222412321303
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1024,10240,0.009006222089131674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,65536,6144,0.0937439997990926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,10240,6144,0.036377777655919395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,12288,7168,0.02032711108525594
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,4096,12288,0.019554666346973844
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,3072,3072,0.01735822194152408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1024,4096,0.006272000157170826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,65536,768,0.01716977854569753
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,1024,256,0.014899555179807874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,2560,2560,0.021573333276642695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,10240,512,0.006591111007663939
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,8192,8192,0.2069751156700982
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1024,8192,0.013628444737858243
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,8192,2048,0.00960000024901496
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,10240,1024,0.05270311236381531
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3072,256,0.00254844439526399
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3072,64,0.008566222257084316
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,7168,3584,0.01074399964676963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,1536,768,0.01758044461409251
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,16384,3072,0.0491920014222463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,8192,512,0.006970666348934174
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,10240,768,0.029601779248979356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,16384,6144,0.040516446034113564
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,64,10240,0.028419554233551025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2560,16384,0.021553777986102637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,8192,128,0.005685333162546158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,7168,65536,0.1290488905376858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,512,32,0.003936000168323517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,10240,128,0.006027555714050929
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,32,65536,0.016584888100624084
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1536,8192,0.01072622256146537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,128,6144,0.005230222311284807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,6144,512,0.0035626664757728577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,32,1536,0.003579555700222651
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,6144,12288,0.018203554881943595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,32,65536,0.27620800336201984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,3584,256,0.0165955556763543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2048,5120,0.009476444787449306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,3584,512,0.015121777852376303
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,65536,768,0.12298222382863362
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,8192,2560,0.007809778054555257
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,128,32,0.007985778152942657
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,8192,64,0.040842665566338435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,16384,768,0.018945778409639995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,10240,1024,0.007652444144090016
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,1024,6144,0.025552888711293537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,4096,1024,0.00600355573826366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3584,5120,0.010303110712104374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,3584,256,0.016559999850061204
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,10240,128,0.015508444772826301
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,16384,1024,0.0076142218377855085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,128,64,0.0019146667586432563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,10240,16384,0.3903368843926324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,2048,256,0.0052533331844541765
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1024,2048,0.0052800000541739995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,2560,12288,0.03687999977005853
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,6144,16384,0.025399999486075506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,12288,3584,0.20087911023033989
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,512,5120,0.009034666750166152
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,256,2560,0.004953777624501122
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,256,6144,0.014774221513006421
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,256,128,0.005615111026499007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,16384,2048,0.01218577805492613
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,768,1536,0.014328888720936246
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,65536,1536,0.07389511002434625
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,16384,128,0.009200000100665623
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,128,1024,0.012158222496509552
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,2560,1536,0.05579288800557455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,16384,256,0.0028684443483750024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,1024,8192,0.02032088902261522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,128,5120,0.008812444077597724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,65536,512,0.07657066980997722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,10240,2560,0.0137155552705129
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,128,3072,0.013962666193644205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,3072,2048,0.04298933347066244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,256,256,0.0028533333291610083
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,1536,8192,0.0181342214345932
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,7168,5120,0.023157333334287006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2048,4096,0.009893333746327294
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,6144,3072,0.012412444584899478
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,2560,2560,0.010255999863147736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,5120,10240,0.030151999659008447
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,8192,768,0.005912888795137405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,1536,1536,0.04500088757938809
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,65536,2048,0.023271110322740342
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,8192,1024,0.004893333133724001
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,4096,256,0.016647110382715862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1024,1024,0.00741244438621733
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,512,5120,0.018232888645595975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,10240,2048,0.1461697816848755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,256,1024,0.0057546666098965546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,2048,256,0.014829332629839579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,512,3072,0.007300444775157505
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,768,3072,0.022681777675946552
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,128,256,0.011648889217111798
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,5120,16384,0.04695466823048062
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2560,10240,0.2680186695522732
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,768,256,0.002483555633160803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,10240,10240,0.06610577636294894
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3584,256,0.005853333406978183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,10240,1536,0.027927110592524212
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,65536,8192,2.468627505832248
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,512,3072,0.011620444556077322
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2048,16384,0.021687999367713928
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2560,4096,0.008007999923494127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,256,5120,0.009696000152164036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,10240,12288,0.03334755698839823
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,16384,512,0.02397511071628994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,4096,5120,0.1041226651933458
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,12288,7168,0.037909332248899676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,4096,1024,0.0034773333205117118
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,10240,128,0.018031110366185505
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,10240,16384,0.23838400840759277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,16384,32,0.006263111200597551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,5120,32,0.003499555504984326
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,768,256,0.014848000473446317
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,1536,2560,0.017986666825082567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,16384,3584,0.019288889235920377
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,6144,32,0.006032889087994893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,1024,1536,0.016864889197879367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,16384,7168,0.034620443979899086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,65536,3072,0.03190755512979295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,3072,10240,0.019202666150199044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,10240,768,0.004977777600288391
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1536,8192,0.013752889302041797
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,8192,768,0.004041777716742622
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,4096,65536,0.09661333428488837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2048,64,0.0059262220230367445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,128,2560,0.005254222287072076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,128,65536,0.0908728904194302
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,7168,512,0.050423112180497914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,8192,128,0.01520088811715444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,3072,10240,0.01586844523747762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1024,10240,0.012099555797047086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,12288,16384,0.1277208858066135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,10240,5120,0.015879111157523263
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,7168,3072,0.007627555893527136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,65536,8192,0.1435342232386271
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,12288,6144,0.02050488856103685
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,512,32,0.006225777582989798
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,8192,1536,0.00665244460105896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,3072,3072,0.007160888777838812
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,256,128,0.011038222246699862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,128,1536,0.012961777547995249
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,10240,1536,0.00812711069981257
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,65536,512,0.06051377455393473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,2560,4096,0.11880711714426677
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,1024,2048,0.014109333356221518
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,16384,65536,0.40054043134053546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2048,16384,0.031742221779293485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1024,3072,0.00794488853878445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,256,3584,0.014811555544535318
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,2048,8192,0.012080888781282636
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,128,6144,0.016536000702116225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,768,5120,0.01886488828394148
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,7168,2048,0.020503110355801053
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,512,5120,0.01719288859102461
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,4096,2560,0.0069146665434042616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,4096,32,0.005917333480384614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3072,768,0.0029173334025674393
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1024,3072,0.011733333269755045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,3584,3072,0.007636444436179266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,128,6144,0.009684444301658207
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,16384,65536,0.15492355823516846
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,3072,7168,0.026877333720525105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,16384,3072,0.019415999452273052
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1536,6144,0.01648177703221639
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,4096,4096,0.023930667175187006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,32,3584,0.007602666815121968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,3072,6144,0.022056000100241765
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,768,3584,0.03899022274547153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,32,1024,0.006980444822046492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,10240,65536,0.19024266137017143
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1024,1536,0.005856888989607493
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,16384,10240,0.02889866630236308
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2048,5120,0.008986666798591614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,8192,4096,0.03481511274973551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,512,7168,0.01000711073478063
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,512,2048,0.00656088896923595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,12288,64,0.0032391111469931076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,1024,5120,0.10420178042517768
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,5120,2048,0.006630222416586346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,768,5120,0.033519112401538424
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,7168,10240,0.027083555857340496
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2560,128,0.005603555589914322
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,1024,2048,0.014446222119861178
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,7168,12288,0.0399617784553104
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,128,10240,0.01995022263791826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,768,32,0.005633777628342311
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,2560,2560,0.022436444958051045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,6144,65536,2.323931588066949
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,10240,4096,0.03529422150717841
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,7168,768,0.0059004442559348206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,256,3584,0.007669332954618666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1536,65536,0.975828488667806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,128,512,0.019254222512245178
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,10240,4096,0.03721244467629327
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2048,8192,0.01958666741847992
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1536,1024,0.004695999953481886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,8192,16384,0.11129866706000434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,12288,128,0.00627377794848548
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,4096,768,0.005603555589914322
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3072,128,0.012267555627557965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,65536,4096,0.1515102254019843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,3584,2560,0.01644533375898997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,6144,768,0.005759999983840519
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,768,10240,0.02259822189807892
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,10240,1536,0.011305777562989129
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,65536,5120,0.13428000609079996
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,6144,32,0.0069422221018208405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1536,4096,0.007714666426181793
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,256,3584,0.015178667174445258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,65536,1536,0.1810906728108724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,2048,6144,0.00904799997806549
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2048,4096,0.008040889269775814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,256,65536,0.105185780260298
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,6144,1536,0.011747555600272285
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,6144,3584,0.008176888856622908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,512,8192,0.0058604445722368025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,1024,1024,0.014476444986131458
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,10240,3072,0.01219555570019616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1024,16384,0.017916444275114272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,128,2560,0.006969778074158563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,3072,1536,0.015101333459218344
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,10240,3072,0.028152889675564233
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,8192,768,0.018578666779730055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,16384,128,0.0064382221963670515
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,256,65536,0.08071288797590467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1024,2048,0.0069057775868309875
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,3072,768,0.018241778016090393
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,512,7168,0.024870221813519795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,1536,4096,0.01110755569405026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,1024,1024,0.005653333332803514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,16384,65536,0.17227466901143393
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2560,16384,0.0226248883538776
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,256,7168,0.06486933098898993
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,6144,10240,0.05387555559476217
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,512,2048,0.007991111112965478
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,2048,8192,0.010747555229398938
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,32,2048,0.004546666724814309
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,128,12288,0.022981334063741896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,5120,128,0.014852444330851236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,16384,512,0.024639111426141527
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3072,1024,0.0032213332338465583
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,1536,128,0.014491554763582019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,7168,1536,0.006534222099516127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,8192,1024,0.0052391112678580815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,65536,1536,0.03504800134234958
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,768,6144,0.009094222552246517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1024,1024,0.003964444415436851
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,3584,7168,0.012483555409643384
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,512,128,0.0021582223061058256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,16384,1024,0.13320799668629965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,64,65536,0.5095306502448188
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,4096,1536,0.005965333431959152
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,7168,4096,0.009367110828558603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,1024,16384,0.030738668309317693
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,16384,3072,0.29349422454833984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3584,5120,0.010835555692513784
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3584,5120,0.07071022192637126
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,6144,768,0.0047066667013698155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,6144,2048,0.007883555359310573
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,7168,128,0.013809777796268463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,10240,7168,0.02034311162100898
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,8192,12288,0.02146222194035848
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,5120,4096,0.07214222351710002
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,16384,1024,0.06704088714387682
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,768,10240,0.031743112537595965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,12288,3072,0.012668444050682915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,256,512,0.011754666765530905
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1536,3584,0.00775022225247489
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,256,768,0.002899555489420891
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,12288,256,0.02278222143650055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,64,65536,0.1746008925967746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,512,6144,0.009019555317031013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,256,1536,0.003960888832807541
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,65536,2048,0.0588773356543647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,1024,12288,0.023519999451107446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2560,1024,0.006008000009589725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,64,12288,0.013122667041089801
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,512,2560,0.009454222189055549
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,8192,1536,0.011872889267073737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,7168,3584,0.011744889120260874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,256,3584,0.015556444724400839
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,256,16384,0.1339822212855021
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2560,12288,0.017627555463049147
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,7168,768,0.018943111101786297
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,6144,1536,0.010103111465771994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,2560,512,0.005561777700980504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,65536,12288,0.26853420999315053
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,7168,32,0.007322666545708974
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,12288,10240,0.027965333726671007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,32,32,0.001959111127588484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,512,65536,0.09860622220569187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1536,1024,0.0063573333124319715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,768,128,0.005761777775155173
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,6144,5120,0.015944000747468736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,5120,3072,0.016442666451136272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,16384,3072,0.036630223194758095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,128,7168,0.019270223047998216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,256,4096,0.007996444072988298
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,10240,2560,0.010488000180986194
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3584,7168,0.01408266690042284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,768,1536,0.006237333433495627
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,2560,8192,0.023974221613672044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,5120,8192,0.03363822235001458
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,12288,2560,0.012871999707486896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,768,3584,0.04426311122046577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,65536,128,0.008347555167145198
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,12288,3072,0.013159111142158508
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,12288,10240,0.041533334387673274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,128,3584,0.018486221631368
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2560,64,0.0021235555824306277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2560,128,0.005568000177542369
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,1536,6144,0.009023110899660323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1536,3584,0.007623111208279927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1024,8192,0.09745511081483628
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,7168,768,0.0042204442951414315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,16384,10240,0.5564453336927626
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,10240,768,0.04763022065162659
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2560,1536,0.0058311112225055695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2048,32,0.006605333338181178
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,512,3584,0.009017777939637503
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1536,6144,0.011704000333944956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,16384,32,0.015351111690203348
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,1536,7168,0.01961511042382982
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,32,1024,0.002951111023624738
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,12288,4096,0.03722666700681051
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1536,7168,0.010229333407349056
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2560,8192,0.01529600057337019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2048,1536,0.005941333456171884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,8192,2560,0.023129777775870428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,512,1024,0.012737777498033313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,10240,65536,0.15411555767059326
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,12288,4096,0.03044355577892727
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,5120,128,0.016537777251667447
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,7168,7168,0.3403022289276123
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,4096,12288,0.06710933314429389
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,512,12288,0.02231377859910329
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,128,10240,0.030897776285807293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3072,10240,0.017532444662517972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,2048,256,0.014782221780882942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,8192,1536,0.006606222026877933
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,128,3584,0.00794666674402025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2560,1024,0.005920889063013925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,10240,32,0.002159111201763153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,256,3072,0.007339555356237624
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,7168,1024,0.009333333207501305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,8192,5120,0.17075378364986846
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,7168,768,0.0036435553597079385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,6144,3584,0.023665777511066858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,256,6144,0.009129777550697327
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3584,10240,0.01748533381356133
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,8192,7168,0.019297776950730216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,256,7168,0.010427555276287926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,2560,1536,0.01714933415253957
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,4096,4096,0.02306933369901445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,7168,12288,0.03480711248185899
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,256,1024,0.007240888973077138
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3072,4096,0.008564444051848518
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,7168,65536,0.24305510520935059
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,4096,3072,0.017228444417317707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,16384,8192,0.16741511556837294
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,256,7168,0.018858666221300762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,10240,4096,0.017799999978807237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3072,4096,0.01257333325015174
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1024,64,0.005544888890451855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,10240,3584,0.01348977784315745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,10240,768,0.008360000120268928
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2048,32,0.0022008888837363985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,6144,3584,0.01763288842307197
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,6144,128,0.005559999909665849
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,7168,7168,0.04894666539298164
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,2048,3072,0.015481778317027621
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2048,3584,0.007623111208279927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,4096,12288,0.020252444677882724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,128,10240,0.006274666637182236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,3584,256,0.005551110953092575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,12288,4096,0.04987466666433546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,7168,6144,0.014780445231331719
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,1024,7168,0.012631999949614206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,64,4096,0.00796088907453749
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,7168,5120,0.02639111048645443
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,6144,3072,0.02714577813943227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,6144,2048,0.04840622345606486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,128,10240,0.005605333381228977
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2560,4096,0.00908622228437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,768,6144,0.009078222016493479
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,256,128,0.005924444645643234
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,8192,128,0.005965333431959152
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,5120,65536,1.3893199496799047
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2048,5120,0.10802844497892593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,768,512,0.012615111139085559
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,256,12288,0.013395555317401886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,10240,256,0.00600000015563435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,128,3584,0.004590222405062782
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,7168,65536,0.9571182462904188
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,6144,256,0.01242844429281023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,6144,12288,0.029319110843870375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,128,3584,0.006046222315894232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,64,7168,0.009755555954244401
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,6144,512,0.020940444535679285
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2048,3072,0.011228444675604502
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,512,4096,0.008391111261314815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,256,256,0.015231110983424716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,512,5120,0.017494221528371174
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,64,10240,0.004909333255555895
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,6144,7168,0.01346933344999949
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,7168,1536,0.021700445148679946
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1024,3584,0.016689777374267578
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,12288,6144,0.01828888886504703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,768,16384,0.01902755598227183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,128,65536,0.011776000261306763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,4096,5120,0.09765511088901097
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,12288,4096,0.016404444972674053
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,7168,65536,0.4483644697401259
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,4096,512,0.011459555890825061
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,6144,6144,0.016972444123691983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,768,256,0.0052915554907586836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,5120,6144,0.01961599952644772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,8192,6144,0.04297777679231432
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,2048,7168,0.020888888173633154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,2560,1536,0.006259555617968242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2048,1536,0.00526577772365676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,8192,4096,0.01332355539004008
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3584,512,0.0029404444826973807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2560,3584,0.008369777765538957
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,256,5120,0.00489155575633049
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,5120,10240,0.054550220568974815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,6144,64,0.00903733323017756
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,8192,5120,0.017210667332013447
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2048,64,0.0069617778062820435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,10240,128,0.007296889192528195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,256,6144,0.005903111149867375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,4096,64,0.0021644443687465456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,6144,256,0.00675644435816341
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,3584,1024,0.005605333381228977
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,7168,1024,0.006266666783226862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,6144,3584,0.01035911093155543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,256,1536,0.008166222108734978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2560,2048,0.005199999858935674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3072,1536,0.00406666679514779
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,2048,1536,0.006305777778228124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,32,5120,0.008758222063382467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,4096,128,0.0052764444715446895
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,5120,6144,0.033161779244740806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,256,1536,0.01275022245115704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,65536,512,0.029029332929187354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,128,4096,0.007994666695594788
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1536,16384,0.019302222463819716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3584,5120,0.012067555553383298
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,65536,3584,0.09056088659498428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,3584,3072,0.007368000017272101
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,2560,512,0.013826666606797112
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,5120,64,0.0022035555707083806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,8192,64,0.007622222105662028
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3584,2560,0.007258666886223688
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,5120,3584,0.022219555245505437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,4096,1536,0.007080000307824876
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2048,7168,0.013518222504191928
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,5120,2560,0.006960888703664144
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,1536,3584,0.015531554818153381
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,7168,64,0.014801777071423002
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,1536,2048,0.011503111157152386
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,3584,256,0.0052391112678580815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,7168,7168,0.015986666083335876
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,8192,3584,0.011504000259770287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,4096,16384,0.034781333472993635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,6144,256,0.01755822201569875
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,768,2048,0.006908444480763541
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1024,3584,0.0069448889957533936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,6144,2048,0.018923555811246235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,512,7168,0.009661333428488838
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,6144,6144,0.025164445241292317
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,1024,8192,0.011209777659840055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,1024,1536,0.006068444500366847
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,8192,2560,0.02845422095722622
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1024,64,0.006934222247865465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,4096,1536,0.006817777951558431
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,65536,10240,0.6232382456461588
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1024,1024,0.005251555393139522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,2560,256,0.005240889059172736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,3584,1536,0.06027999851438734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,7168,65536,0.13388533062405056
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,4096,12288,0.02162844439347585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,8192,6144,0.013254221942689685
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,768,6144,0.11519022782643636
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,1536,1536,0.018219555417696636
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,256,1536,0.006603555546866522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,4096,65536,0.09221600161658393
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,3584,4096,0.018401778406567044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,5120,32,0.0022897778285874263
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,128,6144,0.018966222802797954
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,768,2048,0.00703111125363244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,1536,2560,0.022679999470710754
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,1024,5120,0.022775999373859827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,2560,5120,0.029917332861158583
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3072,5120,0.02309155629740821
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,512,3584,0.021126222279336717
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,64,16384,0.010011555420027839
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,10240,7168,0.43315643734402126
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,3584,8192,0.02581600017017788
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,4096,64,0.002616000051299731
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,6144,2048,0.006618666566080517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,512,5120,0.091684447394477
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,1536,4096,0.01723911033736335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,7168,6144,0.015112888481881885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,12288,64,0.010049777726332346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,10240,7168,0.026439110438028973
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,5120,10240,0.033857779370413885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3584,64,0.002493333278430833
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,12288,65536,0.16063288847605386
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,65536,2560,0.04828177889188131
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,8192,512,0.017828444639841717
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,65536,8192,0.3067226674821642
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,6144,5120,0.010417777631017896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,128,65536,0.10481244325637817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,3584,12288,0.020620443754725985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,768,5120,0.009318222602208456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,10240,64,0.0025128889828920364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,4096,256,0.002415111081467734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,512,1024,0.005559111220969095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,12288,3072,0.01308444473478529
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,32,3584,0.004876444323195351
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,65536,1536,0.029901332325405542
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,2048,6144,0.01868533260292477
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,12288,768,0.00868977771864997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,10240,768,0.006064888917737537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,65536,768,0.05213155680232578
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3584,1024,0.0031893334041039148
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,128,4096,0.007650666766696506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,65536,2560,0.1599457793765598
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,2048,5120,0.008332444561852349
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,256,4096,0.023007111416922674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,32,3072,0.004670222186379963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,8192,2560,0.028472890456517536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,7168,5120,0.036375999450683594
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,128,768,0.01346666696998808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,3584,5120,0.024459555745124817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,3584,256,0.005554666535721884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,12288,5120,0.01829955478509267
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,512,2048,0.004928888960017098
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2560,256,0.00254133323000537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1024,3072,0.005586666779385672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,10240,5120,0.015840000576443143
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,5120,256,0.028577778074476454
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,12288,4096,0.01720266706413693
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1536,2560,0.0075599998235702515
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,8192,128,0.005943999936183293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3584,65536,0.06550577614042494
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3072,1024,0.0035537779331207275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3584,16384,0.029106666644414265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,2560,128,0.015164444843928019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,768,128,0.005643555687533484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,16384,3072,0.02166755497455597
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,768,65536,0.29913422796461314
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,128,16384,0.015493333339691162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,2560,65536,0.15029422442118326
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,4096,1536,0.0196435550848643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,4096,16384,0.024311110377311707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,128,4096,0.014481777946154276
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,2560,512,0.005932444499598608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,16384,1536,0.022760889596409265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,1536,256,0.01595288846227858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,6144,3072,0.009367999931176504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,7168,8192,0.056743999322255455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,12288,32,0.007298666569921706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1024,8192,0.012089778151777057
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,65536,10240,7.450252109103733
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,2560,7168,0.009655999640623728
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,10240,256,0.006599111275540457
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,12288,8192,0.0229768885506524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,7168,768,0.01961599952644772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,7168,2048,0.03183111217286851
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1024,8192,0.011944889194435544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,32,1536,0.006598222172922558
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,16384,256,0.010370666782061258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,6144,1024,0.006256888724035687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,65536,3584,0.03617066807217068
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,256,1024,0.007287999822033777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,256,512,0.005330666485759947
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,10240,5120,0.021074665917290583
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3584,4096,0.011152000062995486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,12288,8192,0.039239999320771955
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,8192,512,0.017873777283562552
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,768,8192,0.049351112710105054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,7168,16384,0.07687644163767497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,768,2560,0.01926488843229082
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,768,3072,0.007012444237867991
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,3584,65536,0.08130488793055217
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,2560,16384,0.020298666424221463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2560,16384,0.02075199948416816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,12288,16384,0.06822933091057672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,5120,64,0.002534222271707323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3072,8192,0.014448889427714877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,256,64,0.0069653333889113525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,10240,3072,0.013277333643701343
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,512,6144,0.017271111408869427
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2048,2048,0.0074684446056683855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,16384,7168,0.045441776514053345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,7168,4096,0.2695884439680311
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2048,10240,0.013281777501106262
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,3072,3584,0.00979733301533593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,7168,3584,0.011360888679822287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,256,256,0.0025920000755124623
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,32,64,0.00665244460105896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,12288,768,0.0069573331210348345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,32,1024,0.003014222201373842
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,4096,8192,0.06364088588290744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,512,1024,0.005519111123349931
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,3584,10240,0.047255999512142606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,32,4096,0.007887110941939885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,3072,4096,0.012468444804350534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,16384,2048,0.01131377783086565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,5120,5120,0.1721564398871528
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,2560,65536,0.06528711318969727
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,32,3584,0.00507733349998792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,64,2560,0.0074240002367231585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,8192,512,0.0029128889242808023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2048,5120,0.01205599970287747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,65536,1536,0.07637955745061238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,128,12288,0.01314133322901196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,2048,128,0.014767110347747803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,12288,2560,0.013112000293201871
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,2048,2560,0.008958222137557136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,6144,768,0.01716711123784383
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,4096,64,0.006603555546866522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,768,7168,0.008953777452309927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,8192,128,0.004020444634887907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3584,8192,0.04488977789878845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,1536,16384,0.06263733572430082
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,16384,12288,0.1654320028093126
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,6144,1536,0.04971111151907179
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,512,128,0.013085333009560903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,65536,1536,0.01922044489118788
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,512,2560,0.006678222368160884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,768,3584,0.007309333317809635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,8192,128,0.007074666519959767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,5120,1024,0.006311999840868845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,3072,1536,0.015141333142916361
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,3584,12288,0.017961778574519686
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,12288,64,0.007630222373538547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1024,5120,0.008402666283978356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,6144,32,0.03382666574584113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,12288,3072,0.014021333720948962
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,5120,768,0.029277334610621136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,6144,65536,1.5939786699083116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,7168,4096,0.044771555397245616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,256,1536,0.027078222897317674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,2560,2048,0.0069431112044387394
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,12288,6144,0.0675848921140035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2560,65536,0.12712177965376112
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,1536,256,0.004879999905824661
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,5120,1024,0.009008000294367472
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,512,768,0.002879111096262932
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,4096,8192,0.08476355340745713
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2560,1536,0.0038542221817705366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,256,768,0.00556355549229516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,1024,128,0.01426666643884447
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1024,8192,0.01921422282854716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3072,2048,0.006568888823191325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,16384,1024,0.014546665880415173
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1536,512,0.005591111050711737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2560,1024,0.003167111012670729
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,128,2560,0.006618666566080517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,3584,512,0.014095111025704278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,4096,3584,0.07146488957934909
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,65536,512,0.013872000078360239
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,1024,128,0.017679999272028606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,768,1024,0.011287111375066968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1536,4096,0.011734222372372946
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,5120,12288,0.046015110280778676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1024,64,0.011704000333944956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,12288,16384,0.040904889504114784
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,512,1536,0.0038568890757030914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,12288,32,0.005325333525737126
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,3072,8192,0.011882666912343768
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,32,65536,0.04554133282767402
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,128,1536,0.006287999865081575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,4096,768,0.019243554936514962
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,4096,1536,0.011017777853541903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,4096,256,0.0024986666523747975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,768,65536,0.08452444606357151
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,12288,3072,0.012708444562223224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,8192,1536,0.0076657773719893555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,65536,2048,0.12330044640435113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,8192,6144,0.013165333204799227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,7168,65536,0.13295911418067083
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,2560,4096,0.04170755545298258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,128,16384,0.03601600064171685
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,512,128,0.0052382221652401825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,6144,8192,0.048728889889187284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,128,1024,0.012087110843923358
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,7168,128,0.021427555216683283
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,7168,1024,0.005993777679072485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1024,32,0.006583111153708563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,768,16384,0.14943467246161565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,12288,2048,0.028796444336573284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,65536,6144,0.06468177504009671
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,128,2048,0.013082666529549493
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1536,512,0.0032231110251612137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1024,512,0.0032755554550223877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2560,1024,0.006501333167155583
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,12288,5120,0.016159110599093966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,6144,10240,0.025765332910749648
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,1024,65536,0.06243466668658786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,10240,8192,0.02491644521554311
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,6144,10240,0.022481777601771887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1536,32,0.0021928888228204516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,3072,3584,0.02314222190115187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,768,1024,0.0032311110860771606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,65536,2560,0.05731733640034994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,8192,4096,0.15702666176689997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,64,128,0.006272888845867581
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,1024,6144,0.042782220575544566
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,3072,128,0.0052933332820733385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,7168,10240,0.01846311158604092
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,512,2048,0.006639110959238476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,4096,1024,0.006016888966162999
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,65536,32,0.013962666193644205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,65536,256,0.009719111025333405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,8192,1024,0.043452445003721446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,5120,5120,0.012461333639091916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,4096,256,0.005315555466545953
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3584,5120,0.029108444849650066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,5120,8192,0.018603555030292936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,512,65536,0.05413866705364651
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,7168,2048,0.008126222425036961
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,32,8192,0.005691555639108022
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1024,32,0.00573600000805325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3584,3584,0.010073777702119615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,5120,512,0.005601777798599667
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,7168,2560,0.015400000744395785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3584,512,0.008444444172912175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,10240,2560,0.05140977766778734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,6144,2560,0.008775110873911116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,5120,7168,0.021350221501456365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,10240,4096,0.016599110431141324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,6144,10240,0.016397333807415433
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,4096,8192,0.015207999282413058
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,7168,256,0.005676444619894028
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,65536,1536,0.06755910979376899
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,3584,1024,0.018907555275493197
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,16384,2560,0.02014666630162133
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,2048,8192,0.04098577631844415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,12288,2048,0.028530667225519817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,6144,65536,0.14555554919772679
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,6144,4096,0.019745777050654095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,65536,2560,0.04331644376118978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,768,12288,0.011044444309340583
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,128,1024,0.01146044416560067
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,4096,512,0.015232000086042615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,128,2048,0.004662222332424588
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,10240,768,0.018608889646000333
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1024,7168,0.008088000118732452
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1024,16384,0.013052444491121503
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,768,128,0.0052791109515561005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1536,5120,0.008405333591832055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,7168,12288,0.022442667020691767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,12288,128,0.03765955567359924
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,10240,256,0.006309333360857434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,4096,2560,0.02370133333735996
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,2048,2560,0.07316266828113131
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,128,2560,0.007052444749408298
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,768,3584,0.005965333431959152
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,512,1024,0.013417777915795645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1024,128,0.0028275555620590844
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,32,2560,0.005072000126043956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1536,10240,0.014482667048772177
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,768,4096,0.031311111317740545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,512,5120,0.011095999843544431
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,4096,6144,0.011601777540312873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,65536,256,0.12021066082848443
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,12288,1536,0.008544888761308458
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3072,8192,0.011940444509188334
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,1536,1024,0.0059262220230367445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,8192,6144,0.01538755496342977
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,4096,3584,0.032418668270111084
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3072,256,0.022057778305477563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,7168,2048,0.008967111508051554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,4096,3072,0.008268444074524773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,64,3072,0.007249777515729268
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,64,512,0.00319377767542998
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2048,4096,0.007685333490371704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,8192,2560,0.028132445282406274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,6144,7168,0.012861332959598966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,65536,7168,0.5441679954528809
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,5120,4096,0.03111999895837572
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,16384,2048,0.020827554994159274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,16384,16384,0.25245687696668834
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,5120,512,0.0035786665976047516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,128,2560,0.00797155582242542
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,128,2048,0.006612444503439798
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3072,12288,0.3853413263956706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,12288,10240,0.04169866773817274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,65536,12288,0.12972622447543675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3584,32,0.006056889063782162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,32,12288,0.013579555683665805
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,512,512,0.006433777925040986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,768,7168,0.06195288896560669
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,3072,3072,0.02168088820245531
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3584,128,0.002537777854336633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,8192,2560,0.012050666742854647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,512,10240,0.011370666325092316
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,256,2560,0.014112888111008538
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,3072,1536,0.00851022203763326
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,256,3584,0.00907555553648207
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,2560,256,0.005951111101441913
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,16384,2560,0.023697778582572937
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,768,512,0.006315555423498154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,2560,1536,0.006623111251327727
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,7168,12288,0.08383466800053914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3584,3072,0.008956444760163626
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,3584,4096,0.022775999373859827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,4096,768,0.014985778265529208
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,128,512,0.005612444546487596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,8192,256,0.011674666570292579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,512,256,0.014592889282438489
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,128,16384,0.030151999659008447
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,3584,3584,0.02382844520939721
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,8192,64,0.00665155549844106
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,128,1024,0.005992888990375731
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,512,12288,0.06433688931994967
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,65536,256,0.12746311558617487
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,128,768,0.013734222286277346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1536,1536,0.004320888883537716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,12288,65536,3.4170479244656033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,16384,4096,0.0513955553372701
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,7168,8192,0.018858666221300762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,512,7168,0.12405067019992405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,65536,12288,3.6655938890245228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,12288,10240,0.02713511056370205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,10240,8192,0.07157599925994873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,12288,6144,0.35998577541775173
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,1024,16384,0.016202666693263583
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1024,6144,0.007716444631417592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,6144,12288,0.040693332751592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,1536,6144,0.015517334143320719
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1024,128,0.005304000029961268
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,512,256,0.014529777897728814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,10240,1024,0.007723555796676212
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,5120,2560,0.030079109801186457
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,16384,2048,0.013578666581047906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,768,64,0.005986666513813867
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,1536,65536,0.1707519955105252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2048,3072,0.0073626670572492815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,7168,16384,0.04663911130693224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,16384,1536,0.010031999813185798
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,512,3072,0.005619555711746216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,7168,16384,0.023426666855812073
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,768,16384,0.018989332848125033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3584,10240,0.015807999504937064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,128,256,0.011815110842386881
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,5120,1024,0.005974222388532426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,10240,1536,0.008013333711359236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,512,65536,0.06231377522150675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3072,10240,0.016156444946924847
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1536,65536,0.48935731252034503
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3584,512,0.02542488939232296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3072,65536,0.08089777496125963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,2560,3072,0.02237422267595927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3584,4096,0.011327111058764987
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,65536,512,0.04332000017166138
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,128,10240,0.006051555689838197
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2560,6144,0.011838222543398539
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2560,64,0.0021973333011070886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3072,2048,0.011392888923486074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,4096,12288,0.09080088800854152
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,12288,10240,0.07263733281029595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,65536,12288,0.10362222459581162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,1024,3072,0.019460444649060566
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,12288,1536,0.010048888623714447
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1536,12288,0.01458577811717987
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,2560,512,0.014192000031471252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,8192,768,0.003370666669474708
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2560,64,0.017553778158293832
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,3584,65536,0.06619466675652398
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,12288,6144,0.024915556112925213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,6144,3584,0.00870311094654931
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,2560,2048,0.017623111605644226
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,1024,5120,0.02218222286966112
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,8192,4096,0.03401511245303684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,64,64,0.005692444327804778
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,5120,2560,0.008183110919263627
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,768,5120,0.015976000163290236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,7168,6144,0.014864888456132678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3584,4096,0.009706666900051964
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,512,65536,0.05624444617165459
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,12288,1536,0.01168444421556261
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,7168,10240,0.0162124451663759
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,768,1024,0.004326222257481681
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,2560,8192,0.010774222513039907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3584,4096,0.009942221972677443
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,2560,3072,0.008360889222886827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,3072,512,0.006766222417354584
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,7168,12288,0.023431999815834895
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,64,10240,0.004583999928500918
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,6144,128,0.003933333274390963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,4096,32,0.0032017777363459268
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2048,4096,0.00832622249921163
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,8192,4096,0.011521778172916837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,512,1536,0.013570666313171387
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,256,10240,0.03371466530693902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,16384,4096,0.018585777944988675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,10240,65536,0.19765778382619223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2560,8192,0.015386665860811869
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,768,65536,0.6980133586459689
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,128,3072,0.007288888924651676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,12288,128,0.0025519999778932994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,3584,65536,0.07576533158620198
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2048,512,0.0031475555151700974
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,3584,8192,0.012286221815480126
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,6144,2560,0.022521777285469904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,1024,3072,0.007315555380450354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,3584,16384,0.02142400046189626
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,2048,4096,0.007983999947706858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,3584,2560,0.0210817770825492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,32,128,0.005663110978073544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,7168,5120,0.017027555240525138
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,7168,5120,0.024119999673631456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,1536,16384,0.029804444975323145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,32,8192,0.0391226675775316
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,512,128,0.0021626665774318907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,768,3584,0.005992888990375731
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,256,4096,0.008654221892356873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,64,3584,0.004908444566859139
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,6144,2560,0.020237333244747586
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,1024,2048,0.014510222607188754
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,10240,64,0.03197866678237915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,768,1024,0.005959110955397288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,128,256,0.012070222033394707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,512,256,0.006325333482689328
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3072,32,0.0026026666164398193
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,16384,3072,0.016912889149453905
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2560,8192,0.010493333141009012
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,65536,16384,0.5597857899136013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,3584,768,0.00563733321097162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,1536,5120,0.017995556195576985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,64,5120,0.004442666553788715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2048,16384,0.019352888067563374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1536,1536,0.017999110950364005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,16384,5120,0.5774097972446018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,4096,512,0.0025964443468385274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,5120,256,0.003286222202910317
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,4096,7168,0.020260444945759244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,768,1024,0.012811555630630918
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,768,256,0.0029440000653266907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,3072,128,0.0148062225845125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3584,6144,0.0480968885951572
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,8192,10240,0.01791200041770935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,10240,12288,0.032496889432271324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,4096,512,0.005607999861240387
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,1536,1024,0.016673778494199116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3584,4096,0.009188444250159794
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,2560,2048,0.013839999834696451
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,12288,16384,0.04270133376121521
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,16384,1536,0.016387555334303115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,512,6144,0.018588445252842374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,32,8192,0.00527022240890397
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,512,64,0.0019413334214025075
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3584,16384,0.0243857784403695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,65536,1024,0.06795555353164673
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,10240,2048,0.00871288859181934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,256,1536,0.008022222254011366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,512,256,0.004998222407367495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,256,65536,0.055008888244628906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,128,1024,0.006664888726340399
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,768,12288,0.03563022282388475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,1024,5120,0.017839110559887357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,8192,4096,0.08614844746059841
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1024,64,0.00498933345079422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,3584,12288,0.019346666004922655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,3072,5120,0.14995377593570286
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,10240,32,0.003563555578390757
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,256,3072,0.011179555621412067
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,256,5120,0.013233777549531726
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,128,5120,0.007622222105662028
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3584,4096,0.009123555488056606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,64,32,0.005598222215970357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,65536,3584,0.07955111397637261
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,768,4096,0.018271999226676095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,10240,256,0.0074026667409473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3072,16384,0.022269333402315777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,128,2048,0.014309333430396186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2560,2560,0.007978666987684038
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,10240,5120,0.02589688864019182
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,65536,16384,0.3499155574374729
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,6144,1024,0.006663110935025745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,65536,6144,0.05823733409245809
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,6144,8192,0.014857777290874057
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2048,3072,0.01074577785200543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,8192,512,0.006279110908508301
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,10240,256,0.01756533318095737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,4096,64,0.007678222325113084
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,512,256,0.005281777845488654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,6144,4096,0.008823110825485652
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,8192,3584,0.011386666860845355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,256,128,0.011800888511869641
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,768,64,0.002424888933698336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,8192,768,0.006269333263238271
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,10240,1024,0.006751110984219446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,5120,8192,0.021327111456129286
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,128,12288,0.014477334088749357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,3072,2048,0.009879999690585667
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,512,768,0.0063031112982167145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1024,256,0.002712888850106133
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,1536,10240,0.02414488957987891
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,512,32,0.005943111247486538
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,6144,1536,0.005609777652555042
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1024,2048,0.004598222259018156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,1024,12288,0.023393778337372675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,8192,1024,0.004862222406599256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,7168,256,0.00591555568906996
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,128,3072,0.007365333537260692
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,3072,10240,0.028157333532969158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2560,768,0.0063582224150498705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,65536,3072,0.05705866548750135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1024,128,0.005973333285914526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3584,16384,0.029334220621320937
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1024,768,0.005976888868543837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2048,12288,0.017218665944205392
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,10240,768,0.019221333993805777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,12288,512,0.006897777732875612
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3072,2560,0.010602666272057427
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2560,3072,0.008388444781303406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,16384,256,0.006616000086069107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,7168,2560,0.015045333239767285
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1536,1536,0.00871288859181934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,10240,1536,0.010809777511490716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,3584,3584,0.02756711178355747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,16384,1536,0.009848000274764167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,512,1536,0.013883555928866068
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3072,6144,0.011018666956159802
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,10240,12288,0.7463342348734537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,4096,768,0.003942222230964237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,32,128,0.002622222320901023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,768,12288,0.012276444170210095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3584,65536,0.08262133598327637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2560,7168,0.18604444132910836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,3072,8192,0.12366044521331787
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,768,3072,0.00851466672288047
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,768,12288,0.013906665974193148
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,4096,16384,0.038388444317711726
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,4096,32,0.007823111282454597
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1024,65536,0.0570160018073188
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,5120,5120,0.013101333545313941
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,64,32,0.0019013333237833446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,12288,8192,0.14412089188893637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,1024,3584,0.007982222570313348
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3584,12288,0.020647999313142564
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,32,1536,0.013532444834709167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,8192,5120,0.03606222073237101
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,512,1536,0.008017777568764156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,6144,6144,0.26770223511589897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,10240,16384,0.03928000066015456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3072,32,0.002226666650838322
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,1024,6144,0.023741333021057978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,512,1536,0.003846222327815162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,5120,2048,0.01792888840039571
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,1024,1024,0.014175999495718213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,3072,12288,0.07486933469772339
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,65536,6144,0.15636444091796875
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3072,2560,0.005941333456171884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,3072,3584,0.021470222208235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,6144,7168,0.03037422233157688
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,10240,3072,0.01716088917520311
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,3072,7168,0.01238844460911221
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,16384,128,0.006568000134494569
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,16384,32,0.0021520000365045336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,16384,128,0.007965332931942409
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,512,3072,0.005250666704442766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,16384,2560,0.01239111108912362
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1536,3584,0.012665777570671506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,128,64,0.0018613334331247543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,512,4096,0.015936889582210116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,128,7168,0.009995555712117089
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,768,10240,0.011706666813956367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1024,3072,0.005560889012283749
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,1536,8192,0.014526221487257214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,512,10240,0.011718221836619906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,65536,2560,0.036279112100601196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,65536,16384,0.22907466358608672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,256,6144,0.011063111325105032
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,7168,2048,0.02435644467671712
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3584,64,0.005567111074924469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,4096,8192,0.015827554795477126
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,5120,1024,0.02293333411216736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,256,5120,0.01955911186006334
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,8192,1536,0.007860444486141205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,5120,768,0.007277333074145847
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,12288,2560,0.04191111193762886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2048,3584,0.008440888590282863
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,1024,256,0.015011555618709989
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2048,6144,0.03497955534193251
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,1024,16384,0.03378399875428941
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,65536,3584,0.042168001333872475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,12288,3584,0.016907556189431083
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,512,8192,0.010059555371602377
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,64,1024,0.0029191111938820947
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,7168,8192,0.03314310974544949
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,256,2048,0.013136000268989138
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,8192,64,0.0028799999919202593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,64,65536,0.017952000101407368
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,768,1536,0.015538665983412
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,8192,6144,0.028431110911899145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,10240,64,0.00869955536392
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,12288,3584,0.017323555217848886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,6144,256,0.005595555735958948
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,12288,65536,0.14145778285132513
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,5120,512,0.005928888916969299
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,5120,64,0.0022293333378103045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,10240,1024,0.049999998675452344
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,768,128,0.0063386667105886675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,128,32,0.005294222384691238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,16384,512,0.006684444430801604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3072,32,0.010837333069907295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,128,512,0.011773332953453064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,16384,65536,5.738118913438584
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,10240,128,0.013478221992651621
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2048,10240,0.03495111068089803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,32,10240,0.010822222464614444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,10240,3584,0.07960266537136503
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,3584,3584,0.053559998671213783
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,128,4096,0.0716257757610745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,65536,2560,0.08657688564724392
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,16384,2048,0.016368000043763053
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,4096,65536,1.1389724943372939
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,256,256,0.007939555578761632
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,7168,128,0.005967111223273807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,5120,2048,0.006739555547634761
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,6144,6144,0.027152000202072993
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,128,6144,0.009005332986513773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,32,3072,0.020382222202089097
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2560,1536,0.00868977771864997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,768,3072,0.017180444465743173
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,1024,2560,0.016199999385409884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,8192,3072,0.011085333095656501
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1536,12288,0.016231111354298063
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,2560,1536,0.006288888967699475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,7168,128,0.005295111073387994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,768,6144,0.009038222332795462
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,64,32,0.00775911079512702
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,2560,12288,0.016902221573723685
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,5120,3584,0.0151591118839052
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1024,2560,0.007307555940416124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,5120,7168,0.02586488922437032
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,7168,128,0.0155102229780621
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,512,8192,0.008040000167157915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2560,12288,0.018070222602950204
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,6144,32,0.003949333396222857
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,10240,128,0.002564444517095884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,128,8192,0.010767111347781287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,7168,128,0.014920888675583733
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,16384,16384,0.048100445005628795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,1536,8192,0.0232968893316057
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,65536,8192,0.08919111225340103
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,128,6144,0.009780444204807281
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,4096,128,0.01443822185198466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2560,7168,0.012148444851239523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,2048,1024,0.006559999866618051
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,8192,256,0.002862222285734283
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,128,10240,0.0129457778400845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,7168,3072,0.010757333702511258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2048,16384,0.0826888879140218
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,65536,4096,0.051827556557125516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,6144,12288,0.05134222242567274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,3584,768,0.008242666721343994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2048,256,0.005630222045713001
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,3072,2048,0.01623911162217458
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,7168,8192,0.02236266599761115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1536,5120,0.012727999852763282
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,2560,768,0.005933333188295364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3584,1536,0.006583111153708563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,8192,12288,0.042295998997158475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,8192,7168,0.031138665146297876
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,65536,256,0.018948444061809115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,7168,2560,0.009121778110663096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,16384,3072,0.03686577743954129
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3584,2560,0.008279110822412703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,3584,16384,0.020999110407299463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,3584,7168,0.11989688873291016
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,5120,128,0.0052693333062860705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,2560,4096,0.01756800048881107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,32,16384,0.07693155606587727
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3072,6144,0.01204533295498954
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,10240,1024,0.08228977521260579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,2560,8192,0.011384888655609555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,6144,4096,0.08452444606357151
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3584,16384,0.021720000439220007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,7168,256,0.017285333739386666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,12288,12288,0.056223114331563316
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,512,768,0.015110222829712762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,3584,7168,0.041949331760406494
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3584,8192,0.02933599882655674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,5120,512,0.007658667034573025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,128,2560,0.014156444205178155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,256,8192,0.024801777468787298
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,512,10240,0.013079110946920184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,3584,256,0.015422221687104968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,8192,3584,0.03183555603027344
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,5120,512,0.0059057776298787855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,768,10240,0.025435555312368605
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2048,128,0.0059164443777667145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,6144,512,0.0069520001610120135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,65536,256,0.019742222295867074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,4096,768,0.003248888792263137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,1024,4096,0.016190222567982145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,5120,1024,0.007283555964628856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,3584,512,0.006221333311663733
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,12288,3584,0.0192266669538286
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,256,2560,0.0068853331936730286
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,6144,768,0.0032924444725116095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1024,1536,0.006592888798978593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,2048,16384,0.03631466627120972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,32,3584,0.0052720002002186244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,3584,8192,0.012903999951150684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,128,1536,0.005963555640644497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,256,128,0.004972444640265571
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3584,8192,0.060976001951429576
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,1024,65536,0.06195200151867337
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,8192,16384,0.04605422086185879
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,6144,3072,0.008230222596062554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,10240,6144,0.03408799899948968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,8192,7168,0.028902222712834675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,2048,1536,0.014807111687130399
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,5120,768,0.017872888180944655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,128,3072,0.007078222102589077
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,4096,10240,0.017419555120997958
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,10240,3584,0.04134844409094916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,4096,2560,0.043495999442206494
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,7168,2560,0.00794488853878445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3584,6144,0.014109333356221518
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,6144,2560,0.007661333514584436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,256,512,0.012455110748608908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,4096,2560,0.010306666294733683
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,6144,7168,0.014864888456132678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,65536,10240,0.23307556576199004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,4096,2048,0.007607111500369177
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,256,768,0.005923555543025334
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,65536,7168,0.10213777754041885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,1024,128,0.014480888843536377
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,16384,10240,0.03576088945070902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1024,3584,0.008302222523424361
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,768,256,0.00590844452381134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,3584,3584,0.025418667329682246
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,7168,7168,0.15912532806396484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,6144,12288,0.02968533171547784
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,2048,3072,0.007328888608349695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,16384,3584,0.01443199978934394
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,10240,8192,0.02237777743074629
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,256,1024,0.01239377756913503
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,5120,3584,0.06425955560472277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,768,10240,0.02645333276854621
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,5120,6144,0.011369778050316704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,8192,128,0.007690666450394525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,65536,7168,0.2417671150631375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,10240,512,0.02093422247303857
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,6144,768,0.005604444278611078
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,32,512,0.0024862223201327855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1024,8192,0.007877333296669854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,12288,512,0.0063386667105886675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,8192,2560,0.09615555736753677
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,7168,12288,0.04002044598261515
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,1536,16384,0.0188737776544359
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1536,256,0.00590844452381134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,65536,2048,0.08072355720731947
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,8192,4096,0.22495823436313203
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,768,512,0.005642666584915585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,128,768,0.005628444254398346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,3072,256,0.015666666958067153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,4096,4096,0.008638222184446123
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,16384,2560,0.012108444339699216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,16384,65536,0.26587022675408256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,12288,1024,0.09831466939714219
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,10240,3072,0.013914666242069669
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,7168,1024,0.02404444416364034
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,7168,1024,0.006624889042642381
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,768,2560,0.005251555393139522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,5120,2048,0.006592888798978593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,7168,65536,0.1912400060229831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,12288,512,0.005092444519201915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,10240,7168,0.038586666186650596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,65536,16384,0.37789156701829696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,6144,16384,1.5447662141587999
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,4096,768,0.016528889536857605
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,12288,5120,0.04518844352828132
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,10240,6144,0.027256889475716486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,7168,768,0.005912000106440649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,16384,768,0.12378933694627549
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,12288,768,0.02003466586271922
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,5120,3584,0.02656800051530202
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,5120,2560,0.008532444636027018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,3072,12288,0.10270577669143677
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,3072,256,0.015465777781274585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2560,1024,0.006245333287451003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,512,3072,0.01479822231663598
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,8192,32,0.005615111026499007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,32,16384,0.13391289446089003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,7168,16384,0.05021511183844673
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,1536,6144,0.009082666701740688
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,12288,2560,0.015096000499195524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,4096,16384,0.021007110675175984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,256,7168,0.010216889282067617
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,12288,64,0.00666133314371109
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,32,5120,0.004924444274769889
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,4096,16384,0.4609662161933051
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,6144,7168,0.07619200150171916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,6144,8192,0.06564533048205905
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,7168,128,0.006622222148709827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,12288,32,0.010801778071456485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,65536,64,0.010043555663691627
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,8192,4096,0.035959998766581215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,65536,128,0.030084444416893855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,2048,65536,0.06317955255508423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,64,12288,0.023057777020666335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,16384,1024,0.007271111011505127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,128,768,0.004208000169859992
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,6144,7168,0.02815377712249756
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3584,4096,0.035649776458740234
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,128,16384,0.026766222384240892
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,768,256,0.002459555657373534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3584,8192,0.01642933322323693
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,2560,3072,0.00871733327706655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,32,128,0.005607111172543631
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,64,3584,0.007595555649863348
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,768,2560,0.009706666900051964
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,4096,8192,0.014783999986118741
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,1536,65536,0.10179199775060017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,2560,128,0.014768888552983602
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,8192,3584,0.009881777895821465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,64,512,0.009340444372759925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,128,3584,0.02415200074513753
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,65536,32,0.08827733331256443
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,16384,12288,0.06167822413974338
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,10240,1536,0.023834667272037927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,10240,10240,0.030527111556794908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,7168,16384,0.5283199946085612
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,128,65536,0.04653333293067085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,8192,2560,0.037476443582110934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,128,6144,0.01643022232585483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,8192,6144,0.0396364430586497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,5120,512,0.0069573331210348345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,64,2048,0.007027555671003129
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3072,65536,0.09351022375954522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,12288,512,0.01757244434621599
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,8192,6144,0.12929511070251465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,10240,6144,0.023429332507981196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,5120,12288,0.0792177783118354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,32,5120,0.02050488856103685
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,7168,5120,0.03276266654332479
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3072,128,0.00656355544924736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,6144,6144,0.0395084453953637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,256,16384,0.005701333284378052
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,3584,6144,0.008978666530715095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,1024,4096,0.021293333835071985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,5120,3584,0.012529777983824411
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,6144,6144,0.01185955521133211
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,2048,6144,0.009016888837019602
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,8192,5120,0.02715377840730879
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,8192,32,0.007644444704055786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,5120,1024,0.03052799900372823
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,5120,10240,0.015246222416559855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,5120,64,0.004304888761705822
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,256,5120,0.006707555717892117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,2560,10240,0.01497244503762987
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,7168,8192,0.022848889231681824
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,1024,3072,0.01564800077014499
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,6144,256,0.0029004443850782183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1536,128,0.006598222172922558
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,5120,4096,0.008379555410808986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,768,3584,0.010053333308961656
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,3584,3584,0.019928000039524503
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,12288,5120,0.2873920069800483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,8192,1536,0.12002221743265788
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,10240,3072,0.025822222232818604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3584,6144,0.015809777710172866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,5120,1024,0.004291555533806483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,7168,12288,0.05713511175579495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,5120,128,0.005522666705979242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,7168,1024,0.005960000058015187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,12288,2048,0.011184000306659274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,10240,16384,0.22082577811347115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,5120,65536,0.07304977708392672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,512,7168,0.011177777416176267
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,16384,2560,0.04095288780000474
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,512,32,0.006664000037643645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,4096,16384,0.03398933344417148
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,1024,10240,0.011759999725553723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,65536,8192,0.35192177030775285
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,6144,256,0.005593777944644292
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,16384,64,0.0028826666788922418
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3584,10240,0.016128889388508268
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,8192,10240,0.027659555276234944
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,768,2048,0.009371555513805812
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,65536,8192,0.14653333028157553
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,5120,1024,0.007626666790909237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3072,16384,0.022673777408070032
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,12288,2048,0.016211556063758004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,128,7168,0.013360888593726687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,768,8192,0.009694221946928237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,32,16384,0.044237332211600415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,768,12288,0.01426488823360867
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,16384,3584,0.04075466593106588
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2048,12288,0.11018488804499309
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,4096,2048,0.019574221637513902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,65536,2048,0.030332446098327637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,32,3584,0.015799999237060547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,65536,3584,0.09073244200812446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,2048,128,0.014827556080288358
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2048,2560,0.0069360000391801195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,256,256,0.0021475555582178966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,512,4096,0.016427555018001132
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,10240,32,0.002171555534005165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3584,16384,0.02199555602338579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,6144,768,0.01867733399073283
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,4096,7168,0.02030311193731096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,3584,2048,0.016882666283183627
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,4096,2560,0.00850755555762185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,768,5120,0.005548444473081165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3584,5120,0.009666666388511658
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,6144,768,0.028892444239722356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,2048,128,0.014783999986118741
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,3072,6144,0.021949332621362474
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,2560,1024,0.005912888795137405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,7168,768,0.01754399968518151
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,10240,7168,0.026750221848487854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,65536,32,0.05439200003941854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,2048,1536,0.051038222180472485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,256,5120,0.03387199838956197
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,512,7168,0.007328888608349695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,256,10240,0.00907555553648207
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,8192,1024,0.004448889030350579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,256,16384,0.03129244512981839
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,65536,1536,0.025476444098684523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,768,65536,0.20574577649434408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,6144,2560,0.007298666569921706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,12288,6144,0.01809866726398468
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,10240,10240,0.02832088867823283
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,5120,6144,0.013433777623706393
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,512,8192,0.010632000035709806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,5120,128,0.0052791109515561005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,768,1536,0.006599999964237213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1536,3584,0.059642665916019015
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,256,256,0.006663110935025745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,768,1024,0.005889777921968036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,3584,1536,0.028912888632880315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,6144,768,0.0059279998143513995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,16384,128,0.021507556239763897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2560,7168,0.012327111429638334
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,16384,6144,0.02033422225051456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,128,16384,0.016228444046444364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,10240,2560,0.011439111497667102
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,3072,256,0.015180443723996481
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1536,12288,0.01644977761639489
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,32,3584,0.009264888862768808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,128,5120,0.008741333252853816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,2560,65536,0.2294631004333496
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,12288,12288,0.06490044461356269
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,10240,128,0.0063182223174307085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,2048,128,0.005249777601824866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,64,7168,0.0052720002002186244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,7168,128,0.005616888817813661
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,6144,5120,0.024892444411913555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,4096,10240,0.028530667225519817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1536,2048,0.015849777393870883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,2048,768,0.01825066738658481
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,256,256,0.0052977779673205475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1536,256,0.006664888726340399
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,8192,16384,0.030584888325797185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,3584,4096,0.02420888841152191
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,3072,3072,0.02103999919361538
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,65536,3072,0.03168977631462945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,12288,4096,0.0210533340771993
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3584,1536,0.05314844515588549
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,512,512,0.005577777822812398
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2048,3584,0.06519555383258395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,512,512,0.005897777775923411
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,7168,12288,0.3887511094411214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,7168,512,0.034225778447257146
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,10240,8192,0.025850666893853083
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,65536,2048,0.026438222991095647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,5120,128,0.007609777980380588
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,10240,3072,0.18880621592203775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2048,768,0.004531555705600315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,6144,2048,0.013053333593739403
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,64,3072,0.004575110971927643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,128,128,0.002185777864522404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,7168,512,0.005730666634109285
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,7168,512,0.00702311098575592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,2048,1024,0.014916444818178812
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,64,2560,0.005271111097600725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,6144,3584,0.011058666639857821
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,10240,6144,0.029754665162828233
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,4096,6144,0.03128977616628011
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3584,16384,0.02253955602645874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,10240,16384,0.053299556175867714
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,12288,8192,0.039552887280782066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,6144,128,0.006581333362393909
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,256,128,0.012121777567598553
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,6144,1024,0.0076657773719893555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3584,3584,0.008353778057628209
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,256,2560,0.007288888924651676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,512,12288,0.013502221968438892
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,16384,5120,0.02572088936964671
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,7168,3072,0.010648888846238455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,256,8192,0.010956444674068026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3072,16384,0.3655324512057834
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1024,256,0.005950222412745158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,12288,3584,0.035720000664393105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,5120,7168,0.026838221483760413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,7168,65536,0.12648267216152614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,7168,12288,0.031332443157831825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,5120,10240,0.017621333400408428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3072,2048,0.0332897769080268
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,8192,2048,0.01752088963985443
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,4096,6144,0.03180977702140808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,10240,5120,0.04018311036957635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,768,16384,0.027090667022599116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,12288,32,0.005939555664857228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,4096,2048,0.006648889018429651
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,3072,768,0.029363556040657893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3584,10240,0.0151591118839052
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,512,3072,0.007639110916190677
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,5120,5120,0.010363555616802638
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,5120,8192,0.04402933186954922
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,16384,1024,0.007975999679830339
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,128,4096,0.008024888734022776
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,8192,16384,0.10998755693435669
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,1024,2560,0.007326222128338284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3072,768,0.00906133320596483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,6144,64,0.006687110910813014
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1024,16384,0.013561777770519257
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,6144,2048,0.0069671107663048645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,768,2048,0.00869599978129069
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2048,2048,0.006959111326270633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,16384,128,0.006659555352396435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3584,768,0.0028719999310043124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1024,65536,0.06564178069432576
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,65536,5120,0.10515022277832031
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3584,12288,0.018586667047606576
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,512,12288,0.013353777428468069
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,10240,128,0.004597333156400257
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,4096,6144,0.2277759975857205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,768,512,0.0069982219073507525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,2560,3584,0.02072355482313368
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,7168,1024,0.007651555869314406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2560,1536,0.0063235556913746735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,5120,5120,0.01868088874551985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,3072,128,0.02630666726165348
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,1024,12288,0.029370665550231934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,512,65536,0.19291910860273573
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1536,128,0.00227111101978355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,768,1024,0.005947555518812603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3072,3072,0.006795555353164673
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,256,3584,0.008105778031879002
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,64,7168,0.004632888982693355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,256,1536,0.007337777978844113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,10240,128,0.006224888894293044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,128,2560,0.007335111498832703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,12288,1024,0.01074577785200543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2048,5120,0.008979555633332994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,1024,10240,0.011781333221329583
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2560,512,0.003580444388919406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,128,10240,0.011424889167149862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2048,1536,0.01919555498494042
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1536,3584,0.017438222964604694
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,32,1024,0.003272888975010978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,6144,12288,0.027631999717818365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,256,2560,0.013853333062595792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,1536,128,0.014508444401952954
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,10240,2048,0.07288977834913465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,256,1024,0.01330844478474723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,256,1536,0.008662222160233391
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,8192,12288,0.12196977933247884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,2560,3072,0.007425777614116669
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,5120,3584,0.01054133309258355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,16384,4096,0.015495111544926962
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,2560,7168,0.023337778117921617
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,10240,128,0.008728000024954477
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,5120,6144,0.012889777620633444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2560,768,0.025145777397685583
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3584,7168,0.012961777547995249
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1536,2560,0.0070159998204973005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,256,3584,0.007659555309348636
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3584,64,0.0032515554792351196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,8192,256,0.005588444570700328
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,128,3072,0.005599110904667113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,64,16384,0.07719110780292086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,65536,10240,5.9411366780598955
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,512,64,0.009826666778988307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,1536,2560,0.028934223784340754
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,3584,1024,0.01754666699303521
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,3072,16384,0.44237600432501895
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1536,512,0.010377777947319878
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,3072,256,0.005277333160241445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,65536,6144,0.1200302177005344
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,4096,128,0.014557333456145393
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,2560,10240,0.031445334355036415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,2560,10240,0.01276622215906779
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,32,2048,0.004271111140648524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,3584,10240,0.026477333572175767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,8192,768,0.006670222100284364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1024,8192,0.01068800025516086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,256,3584,0.007408888803588019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,6144,10240,0.3973706563313802
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,4096,1536,0.0069360000391801195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,4096,256,0.0024746666765875286
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,32,12288,0.022407111194398668
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1536,768,0.009547555612193214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,1024,4096,0.01074577785200543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,8192,65536,0.09060355689790513
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,12288,6144,0.034485333495669894
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,65536,128,0.013208000196350945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,32,7168,0.005261333452330695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,4096,6144,0.012253333297040729
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,5120,4096,0.0240657776594162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,128,256,0.011394666300879585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1024,1536,0.023503111468421087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,3072,2048,0.008619555168681675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,5120,1536,0.00665866666369968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,768,32,0.0021155555215146807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,4096,65536,0.1565422217051188
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,4096,10240,0.0151591118839052
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,5120,1536,0.006610666712125142
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,32,768,0.005608888963858287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,5120,1536,0.01908888916174571
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,10240,4096,0.02882844540807936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,5120,2560,0.006973333656787872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1024,2560,0.004883555488453971
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2048,256,0.00258577780591117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,7168,1536,0.006240889016124938
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,65536,1536,0.03726844324005975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,2048,1024,0.006254222244024277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,1536,5120,0.008352888955010308
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,1536,1536,0.006276444428496891
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,4096,16384,0.08278489112854004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1536,3072,0.0196426659822464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1536,8192,0.037845333417256675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,4096,1024,0.017875555488798354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,12288,512,0.0039822223285834
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1536,512,0.003170666595300039
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,4096,512,0.01551822159025404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,5120,64,0.0021439999755885866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,2048,12288,0.017487999465730455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,2048,3584,0.007636444436179266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,16384,3584,0.015690666106012132
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,3584,65536,0.06894044743643866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2560,1536,0.006367110957702001
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,1536,16384,0.02012800011369917
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1024,64,0.002128888956374592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,12288,2048,0.02537600033813053
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,7168,4096,0.025128000312381323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,3584,5120,0.008653333617581261
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,128,4096,0.01609244445959727
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,512,256,0.003576888806290097
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,5120,3584,0.026828444666332666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,768,256,0.01510933372709486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,5120,1024,0.018552889426549275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1536,4096,0.006904888898134232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2560,2560,0.006550222221348021
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,128,64,0.006627555522653792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,1536,16384,0.022250667214393616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,64,8192,0.00563466673096021
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,768,65536,0.0638328856892056
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,65536,4096,0.0389395554860433
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,128,1536,0.006288888967699475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,32,1536,0.0036888888312710654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,10240,3584,0.016939555605252583
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3072,32,0.006301333506902059
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,12288,6144,0.028441776831944782
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,6144,128,0.005623999983072281
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,512,1536,0.038176000118255615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,12288,7168,0.02852355440457662
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,10240,1536,0.0756044454044766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,1536,16384,0.016973333226309884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,5120,64,0.002287111141615444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,8192,2048,0.009306666751702627
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,6144,65536,3.2055511474609375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,4096,5120,0.033950222863091364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,65536,4096,0.05128622386190626
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3072,2560,0.009746666583749983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,64,2560,0.00702133360836241
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,768,10240,0.011788444386588203
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,4096,2048,0.005274666680230035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,8192,10240,0.5253475507100424
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,8192,8192,0.044826666514078774
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,12288,512,0.004267555558019214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,5120,4096,0.008287111090289222
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,256,8192,0.026855111122131348
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,16384,1024,0.00738933351304796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,16384,1024,0.009287111461162567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,3072,7168,0.03808533483081394
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,16384,2560,0.034854223330815635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3072,32,0.0021582223061058256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,5120,16384,0.0837635530365838
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,512,8192,0.01885511146651374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,512,1024,0.003301333429084884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,5120,8192,0.015912888778580558
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,5120,3584,0.007831999825106727
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,3584,12288,0.055330667230818004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2048,512,0.007280000381999546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,5120,2048,0.012423999607563019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,5120,10240,0.016930666234758165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,7168,3584,0.02407111061943902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,2048,1024,0.013811555173661975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,128,1024,0.007272889216740926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,128,768,0.005570666657553778
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3072,512,0.002991111121243901
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,6144,2560,0.01144444445768992
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,6144,8192,0.025097777446111042
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,1024,768,0.016591999265882704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,8192,2560,0.029576000240114
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,10240,6144,0.01812977757718828
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,12288,7168,0.032765332195493914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,3584,1024,0.016180444094869826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,1536,1024,0.005649777750174205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,16384,768,0.024724443753560383
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3584,256,0.005555555638339784
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,256,65536,0.05957511398527357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,6144,128,0.0148044443792767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,10240,128,0.0059262220230367445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2560,3072,0.007461333440409766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,12288,5120,0.018223999275101554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,1024,2560,0.026791110634803772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,2048,768,0.017640888690948486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,768,128,0.0162444445821974
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,6144,12288,0.03187822302182516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,5120,8192,0.13978666729397243
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,3584,16384,0.06732710864808825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,2048,128,0.014517333772447375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,16384,2048,0.21906044748094347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,256,3584,0.01220266686545478
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,8192,12288,0.02405599918630388
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,128,1536,0.006695111178689533
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,10240,16384,0.053802665736940175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,65536,128,0.01755288905567593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,16384,8192,0.040538665321138166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2048,7168,0.011872889267073737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,3584,512,0.015125332607163323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,7168,16384,0.09812177552117242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,4096,3584,0.012615111139085559
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,512,2048,0.0069671107663048645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,16384,256,0.017521777086787753
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,65536,256,0.009986666341622671
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,16384,7168,0.046335998508665294
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,16384,2048,0.012164444559150271
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,128,4096,0.008050666915045844
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2048,512,0.0032879999942249725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,1024,10240,0.012096000214417776
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,16384,16384,0.04541066620084974
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,6144,768,0.004499555461936527
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,6144,7168,0.09008622169494629
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,768,2560,0.016194666425387066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,4096,10240,0.016588444511095684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,7168,5120,0.01352355546421475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1024,10240,0.0140346669488483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,5120,8192,0.16221600108676487
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1536,32,0.005596444424655702
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,10240,10240,0.027089777919981215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,512,768,0.015243555108706156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,3072,16384,0.0451546675629086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2048,32,0.0021831111775504216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,512,768,0.016912000046836007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,4096,65536,0.6211342281765407
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,5120,128,0.002234666711754269
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,768,7168,0.0478097763326433
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,6144,5120,0.013704000247849358
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,512,4096,0.01106844428512785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,512,1024,0.003232888877391815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,12288,4096,0.01733955575360192
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3072,2560,0.00702133360836241
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,12288,3584,0.015048000547620984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,32,64,0.008042666647169325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2560,256,0.005940444353553984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,16384,4096,0.07243288887871636
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,65536,10240,0.9971164067586263
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,12288,2560,0.02235377828280131
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,5120,256,0.00904444439543618
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,4096,8192,0.043319112724728055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2560,6144,0.01001422190003925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,3072,2560,0.01626755628320906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,8192,128,0.0052977779673205475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,5120,3072,0.019055111540688407
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,768,64,0.0021057778762446507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,4096,7168,0.014887111054526435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,256,16384,0.016206221448050607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,8192,128,0.015504888362354703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,8192,3072,0.026134222745895386
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,6144,65536,0.08022133509318034
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,128,8192,0.02370400064521366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,32,32,0.0020088888704776764
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,5120,768,0.01124533348613315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,2048,1024,0.01533600025706821
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,8192,5120,0.08247022496329413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,2048,512,0.005592000153329637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,5120,65536,0.12290310859680176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,65536,8192,0.07111110952165392
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,64,12288,0.005577777822812398
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,6144,1024,0.006631111105283101
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,512,8192,0.010397333237859938
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,512,65536,0.062294224898020424
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2048,64,0.0021822222818930945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,5120,65536,0.0734400020705329
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,64,5120,0.00869777798652649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,3584,16384,0.19858845074971518
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,7168,2560,0.13514399528503418
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,768,2560,0.00488800017370118
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,128,8192,0.011408000356621213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,65536,6144,0.12239644262525771
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,5120,128,0.0021342221233579847
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,3072,1024,0.005594666633341048
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,64,4096,0.014175999495718213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,1536,6144,0.018559111489189994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,7168,6144,0.0177528891298506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,8192,10240,0.03644888930850559
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3584,128,0.002149333349532551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,64,5120,0.040545778142081365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,10240,8192,0.05782310830222236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,65536,7168,0.0790862242380778
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,8192,65536,0.15718666712443033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,128,512,0.0025617778301239014
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3584,2048,0.005419555637571547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,2560,65536,0.065065774652693
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,16384,2048,0.029264890485339697
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,32,2560,0.004903111192915174
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,65536,12288,0.19020356072319877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,65536,256,0.023457777169015672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,8192,2048,0.0069146665434042616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,4096,256,0.002420444455411699
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,5120,3072,0.06104177898830838
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,65536,7168,0.0751840008629693
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,6144,512,0.005997333261701796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,6144,65536,0.1161688831117418
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3584,10240,0.01514844430817498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,2560,10240,0.01575199928548601
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,5120,10240,0.02527644402450985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,2048,7168,0.020643555455737643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,6144,2560,0.008355555435021719
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,64,32,0.0020844443804687923
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,128,7168,0.01755199995305803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,5120,512,0.01495733360449473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,768,3584,0.016536000702116225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,768,1024,0.005934222290913264
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,12288,3584,0.016866667403115165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,4096,3584,0.021021333005693223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,8192,256,0.005613333235184352
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,256,12288,0.0073724447025193115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,8192,10240,0.08185066779454549
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,5120,4096,0.05304355422655741
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,4096,12288,0.20637155903710258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,768,768,0.006308444258239534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,65536,7168,0.062071111467149526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,768,1536,0.016396444704797532
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2560,3072,0.007221333682537079
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,12288,6144,0.04515022370550367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,768,8192,0.019483556350072224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,10240,256,0.016900445024172466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,10240,4096,0.3797902266184489
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,16384,10240,0.06715110937754314
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,4096,6144,0.02815999918513828
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,16384,65536,0.24359732204013398
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,10240,16384,0.05259199937184652
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,6144,65536,0.11757511562771267
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,512,2560,0.008768888811270395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,1024,256,0.005592888842026393
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3584,4096,0.10721866952048408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3584,4096,0.009220444493823582
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,4096,256,0.007722666694058313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,256,12288,0.024093333217832778
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,5120,64,0.007328000333574083
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,512,1536,0.008357333640257517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,256,256,0.01351377781894472
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,32,64,0.005963555640644497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,5120,2048,0.006931555353932911
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,256,2048,0.006682666639486949
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,12288,768,0.005342222336265776
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,768,1024,0.01477688882086012
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,16384,768,0.00794311116139094
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,256,256,0.012849777936935425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,10240,512,0.0032053333189752367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,3072,512,0.005643555687533484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,6144,1536,0.03443377878930833
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,65536,12288,0.31843021180894643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,128,3072,0.007996444072988298
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,16384,8192,0.056913779841529004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,5120,16384,0.07610844241248237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,5120,32,0.0052782222628593445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1536,32,0.002228444442152977
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2048,768,0.003960888832807541
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,2560,3584,0.018583999739752877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,768,65536,0.07758044534259372
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,128,4096,0.009769777456919352
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,256,8192,0.010415111151006486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,6144,2560,0.007881777981917063
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,256,8192,0.011454222102959951
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,10240,3072,0.011928000383906893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,512,5120,0.008064000142945183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,64,2048,0.0063324446479479475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1536,3584,0.011397333608733283
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,16384,12288,0.0617777771419949
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3584,3584,0.009158222211731805
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,8192,16384,0.8968995412190756
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3072,2560,0.06407999992370605
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,16384,6144,0.04103733433617486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,3072,8192,0.07362933291329278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,6144,32,0.0022071111533376905
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,256,3072,0.009208888643317753
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,3584,2048,0.01588711142539978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,256,1536,0.00665244460105896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,256,128,0.005280888742870755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,32,1536,0.0035111109415690103
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,8192,2048,0.008536888493431939
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,768,8192,0.01242844429281023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,10240,6144,0.02402577797571818
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,10240,7168,0.048431111706627734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,256,64,0.007999999655617608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,3584,3584,0.04330577784114414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,16384,32,0.047170665529039174
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3584,65536,0.06555377774768405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1024,10240,0.010168888502650792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,16384,512,0.010123555858929953
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,12288,1024,0.015088889333936902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,16384,768,0.006956444846259222
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,10240,2560,0.02781244450145298
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,128,16384,0.025088000628683303
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,64,10240,0.012042666474978128
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1024,256,0.0038168889780839286
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,768,3072,0.0073520003093613525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,10240,3584,0.026787555880016748
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,4096,16384,0.02535466684235467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,10240,5120,0.039042668210135564
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1536,768,0.005902222047249476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,1024,128,0.01457244488928053
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1024,256,0.0032497776879204642
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,10240,7168,0.019327999817000497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3584,1536,0.02834933333926731
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,10240,10240,0.03577155537075467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,256,1024,0.0031866667171319327
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,6144,1536,0.005924444645643234
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,8192,256,0.01684711045689053
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,6144,1024,0.006008000009589725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,32,16384,0.033181332879596286
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,256,3584,0.014839111102951897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,12288,3072,0.06455200248294406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,3584,768,0.005601777798599667
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,7168,65536,0.07963111003239949
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3584,768,0.008986666798591614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,7168,8192,0.13762311140696207
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,768,768,0.0059004442559348206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,4096,2048,0.006717333363162146
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,4096,128,0.014799111419253878
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,32,32,0.006646222124497096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,256,1024,0.013797333670987023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,10240,3584,0.013172444370057849
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1024,4096,0.008300444318188561
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,3584,1024,0.0059279998143513995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2048,16384,0.03581600056754218
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,7168,32,0.01479822231663598
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2560,32,0.006611555400821898
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,16384,512,0.022367111510700647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,2560,7168,0.022661333282788593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1536,12288,0.02674755619631873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,5120,7168,0.012799999780125089
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,2048,10240,0.01240799989965227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,8192,256,0.005575110928879843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,10240,6144,0.030984888474146526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,7168,3072,0.008022222254011366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,5120,6144,0.02639377779430813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2560,32,0.005952000204059813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,512,1024,0.0031822222388452957
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,1024,512,0.0052373334765434265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,768,5120,0.008975110948085785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,4096,64,0.0025111111915773815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,12288,1536,0.09080621931287979
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,8192,1024,0.01126844435930252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,12288,4096,0.020482665962643094
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,64,3072,0.005608888963858287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,5120,4096,0.016321778297424316
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1024,1024,0.0062977779242727495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,1536,6144,0.020543111695183646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,65536,7168,0.7261502477857801
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2048,2048,0.06640711095598009
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,5120,256,0.004879999905824661
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,65536,256,0.02439733346303304
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,5120,2560,0.017687110437287223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,32,7168,0.0053013331360287136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,256,128,0.011056889262464313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,10240,8192,0.04292710953288608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,128,2560,0.006303999986913469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,4096,7168,0.03393955694304572
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,256,7168,0.005640888793600931
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,65536,10240,0.24331998825073242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,768,1536,0.004930666751331753
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,64,3584,0.004935111022657818
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,512,4096,0.0068862222962909276
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3584,768,0.011568889021873474
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2048,5120,0.01772622267405192
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,768,65536,0.08036533329221937
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,64,64,0.002181333386235767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3584,128,0.014505777094099255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,64,7168,0.010606221854686737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,5120,1536,0.006206222292449739
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,65536,5120,0.12579554981655544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,64,2560,0.007068444457319047
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,7168,3072,0.009299555586444007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,12288,4096,0.02216888964176178
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,6144,10240,0.02632355524433984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,768,12288,0.011752000285519494
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,4096,3584,0.00852177788813909
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,128,10240,0.005274666680230035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,2560,10240,0.01421333352724711
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,768,2560,0.006275555739800136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,16384,2048,0.031814220878813006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,8192,768,0.006256888724035687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,256,2560,0.004959110998445087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,512,6144,0.01852888862291972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,8192,65536,0.15759644243452284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,3584,1024,0.005585777676767773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,1024,1536,0.014813333749771118
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,256,5120,0.009343110852771336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,16384,64,0.023640000157886084
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,4096,3072,0.024031110935741003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,768,3584,0.015840000576443143
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,1024,768,0.015512888630231222
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,12288,1024,0.054089777999454074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,7168,16384,0.02568000058333079
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,10240,16384,0.13473333252800837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,1536,3072,0.008290666672918532
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,65536,2048,0.803567091623942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,2048,7168,0.17645955085754395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,65536,12288,0.18620267179277208
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1024,1024,0.004602666530344221
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,8192,16384,0.08233066399892171
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,2560,5120,0.019903999235894945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,128,2560,0.031938665442996554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,256,768,0.005612444546487596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,7168,1024,0.01789155602455139
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,4096,3584,0.01090222183201048
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,128,8192,0.018906666172875296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,8192,10240,0.018755555152893066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,1024,1536,0.00628266649113761
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3072,64,0.0022053333620230355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,16384,5120,0.03846755623817444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,5120,3072,0.02207022243075901
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,16384,64,0.0069395556218094295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,10240,8192,0.02828977836502923
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,256,3584,0.01070844464831882
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,2560,8192,0.03254755669169956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,5120,8192,0.021557332740889654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,128,7168,0.005554666535721884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,12288,768,0.007266666326257918
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,128,12288,0.01683466633160909
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,6144,6144,0.14247200224134657
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,16384,2048,0.014652444256676568
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,65536,8192,0.08847733338673909
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2048,32,0.006285333385070165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,64,4096,0.009448889229032729
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3584,32,0.002177777803606457
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,1536,2048,0.019675556156370375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,10240,3072,0.14686755339304605
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,4096,32,0.0058702222175068325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,768,2560,0.016200888488027785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,768,8192,0.010757333702511258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2048,512,0.007056000332037608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2048,7168,0.03968266646067301
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,768,1536,0.013953778478834363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,8192,512,0.003937777959638172
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,16384,128,0.003264888914095031
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,128,4096,0.005295111073387994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,256,8192,0.011388444238238864
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,768,8192,0.022946667340066697
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,768,256,0.015169777803950839
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,128,3584,0.009024000002278222
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,1536,128,0.014584889014561972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,16384,65536,0.24275554551018608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,8192,1024,0.02586488922437032
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,12288,256,0.009810667071077559
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,5120,128,0.017869333426157635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,32,128,0.007999999655617608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,4096,12288,0.0211173329088423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,64,128,0.006989333364698622
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,4096,2560,0.016985777351591323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,8192,3072,0.009018667042255402
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,12288,5120,0.039525333378050066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,5120,256,0.016891555653678048
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,16384,256,0.0069022224181228215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,7168,2048,0.010627555350462595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,4096,768,0.011251555548773872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,10240,16384,0.03982133335537381
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,7168,8192,0.0158897770775689
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,10240,7168,0.03077600068516201
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,256,512,0.016040888097551134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3584,3072,0.007730666961934831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,6144,512,0.00286666676402092
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,6144,32,0.0021608887861172357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,8192,64,0.005927111125654644
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1024,65536,0.06014221906661987
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,16384,12288,2.4567572275797525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,4096,128,0.014863111906581454
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,8192,4096,0.04647555616166857
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,32,2560,0.012789333032237159
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2560,1024,0.007700444095664554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,7168,3584,0.012320000264379712
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,6144,2560,0.007268444531493717
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,7168,16384,0.03878133495648702
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,12288,6144,0.029191109869215224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,8192,768,0.003563555578390757
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,64,3072,0.007611555357774098
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2048,128,0.0052622221410274506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2048,65536,0.046421332491768726
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,10240,2560,0.01795288920402527
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,16384,6144,0.4610444704691569
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,10240,2560,0.027473777532577515
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1024,512,0.005557333429654439
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,256,64,0.00755555596616533
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,7168,2560,0.025566221939192876
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,4096,512,0.002604444407754474
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,2560,12288,0.03576888971858554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2560,7168,0.009825777676370408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,10240,3584,0.03171555532349481
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,512,6144,0.005626666463083691
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,32,1536,0.006594666590293248
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,2560,2560,0.007306666837798224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,6144,512,0.005554666535721884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,64,7168,0.0052879999081293745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,4096,128,0.005582222094138463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,6144,512,0.00286311118139161
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,6144,512,0.006244444598754247
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,16384,2560,0.013247110777431063
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,2560,3584,0.016196444630622864
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,16384,128,0.047557334105173744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,65536,256,0.2395146687825521
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,5120,1024,0.018552889426549275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,64,1024,0.01647733317481147
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1024,12288,0.14850489298502603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,4096,256,0.01682488951418135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2048,6144,0.012087999946541257
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,16384,128,0.033985777033699885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,16384,6144,0.031228442986806233
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,10240,12288,0.04125688804520501
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1024,768,0.0069297779765393995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,3584,8192,0.026757333013746474
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,768,65536,0.06012088722652859
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,10240,128,0.011428444749779172
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,4096,5120,0.08065955506430732
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,7168,10240,0.01992177797688378
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,12288,1024,0.015763555963834126
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,65536,2048,0.6971004274156359
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,65536,32,0.0441599984963735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,3584,10240,0.06688533226648967
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,3072,2560,0.025742222865422566
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,7168,5120,0.015211555692884656
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,7168,3072,0.01036888857682546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,16384,3072,0.017027555240525138
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3584,1536,0.00665866666369968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,5120,64,0.0021751111166344746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,6144,65536,0.13218843936920166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,16384,8192,0.04328355524275038
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,2560,4096,0.008678221868144141
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,1024,2048,0.0068915556702348925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2048,1024,0.007359111474619971
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,768,512,0.006311111152172089
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,128,3584,0.03821511069933573
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,512,1536,0.010447111394670276
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1536,5120,0.008644444247086843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3072,128,0.0022444443570242985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3584,64,0.0019297777778572505
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,64,2560,0.009033777647548253
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,10240,3072,0.05420533484882779
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,128,16384,0.022082666556040447
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,256,16384,0.016901334126790363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,16384,7168,0.025426665941874187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,768,8192,0.025161777933438618
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,3584,128,0.014496000276671516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,128,1024,0.0032337777730491427
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,65536,256,0.010835555692513784
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,32,65536,0.5103564262390137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,8192,12288,0.025126222107145522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,3584,128,0.01518133282661438
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,128,8192,0.010815111299355825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,6144,512,0.00563822231358952
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,768,4096,0.005964444329341252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3072,1536,0.0038897775941424896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,7168,2560,0.009704888694816166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,64,8192,0.022727110319667395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,7168,512,0.007048000064161088
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,256,256,0.009992000129487779
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,5120,10240,0.06924533181720309
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,128,4096,0.01346933344999949
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,4096,10240,0.07605866591135661
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,7168,32,0.005605333381228977
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,8192,1024,0.0762595534324646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,256,8192,0.019242667489581637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,3072,1536,0.018915555543369718
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,128,8192,0.005576888720194499
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,3072,2560,0.0069573331210348345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,4096,2560,0.016872889465755887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,768,4096,0.04424622323777941
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,128,16384,0.005909333212508096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,16384,5120,0.044734223021401294
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,8192,4096,0.02703999976317088
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3584,1536,0.01143200033240848
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,256,3072,0.01520088811715444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,256,768,0.0029493332323100832
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,8192,4096,0.033492445945739746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,10240,10240,0.0343155562877655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,12288,2560,0.014921777778201632
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3584,8192,0.01314133322901196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,512,8192,0.023923556009928387
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3072,256,0.005943111247486538
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,768,7168,0.07814844449361165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,16384,256,0.050329777929517955
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,3072,3584,0.007328888608349695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2048,8192,0.013004444539546967
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,128,2560,0.013473778135246701
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3584,16384,0.022008889251285132
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,10240,64,0.005943111247486538
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,64,8192,0.005281777845488654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,8192,32,0.006302222195598815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3072,1024,0.00738844441043006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1024,1536,0.005617777920431561
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,12288,2048,0.011407111254003314
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2560,768,0.005886222339338726
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,6144,128,0.01516177753607432
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,512,65536,0.024570667081409033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,4096,1024,0.004201777693298128
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,6144,16384,0.04806133442454868
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1536,4096,0.009326222042242685
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,16384,128,0.015189333094490899
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,256,65536,0.14572532971700033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,5120,256,0.027273777458402846
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,4096,768,0.006284444282452266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,10240,128,0.031113776895735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,5120,32,0.005405333307054307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,768,65536,0.10492177804311116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3072,65536,0.07248977820078532
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,2048,3584,0.016149333781666227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2048,256,0.0023111111174027124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,5120,3072,0.07192622290717231
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1024,4096,0.005919111271699269
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,5120,12288,0.0188773340649075
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,2048,10240,0.02532800038655599
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,16384,128,0.007991111112965478
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,12288,128,0.015815999772813585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,128,10240,0.012099555797047086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3584,4096,0.023539554741647508
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,6144,3072,0.008650666309727563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,1536,2560,0.007287110719415877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,512,6144,0.010283555421564314
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,512,10240,0.027446221974160936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,10240,2560,0.17586044470469156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,512,32,0.0021644443687465456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,4096,65536,4.33024639553494
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,12288,3584,0.024353777368863423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,64,4096,0.007461333440409766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,65536,512,0.3189751042260064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3072,3584,0.007374222079912822
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,256,7168,0.010875555376211802
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2560,3072,0.006606222026877933
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,16384,768,0.008020444048775567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,16384,1536,0.01075644459989336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1536,3072,0.005574222240183089
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1536,6144,0.009336888790130615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,6144,768,0.003945777813593547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2048,10240,0.01610666679011451
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2560,256,0.005609777652555042
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,1024,10240,0.06107022364934286
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,5120,7168,0.014137778017255994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,2048,2560,0.016514667206340365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,4096,128,0.0052977779673205475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,8192,512,0.017587555779351127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1536,10240,0.016705777910020616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,1024,512,0.005283555636803309
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,128,3584,0.014870221416155497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,64,12288,0.008364444805516137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,12288,3072,0.013936888840463428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3072,128,0.002201777779393726
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,8192,3072,0.030913776821560327
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,1024,5120,0.06043733490837944
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3584,1024,0.004994666824738185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,12288,3584,0.043896890348858304
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,128,128,0.005356444252861871
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,8192,32,0.00776088900036282
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,4096,4096,0.007707555260923173
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,1536,8192,0.0321359998650021
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,512,3072,0.014874666929244995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,32,3584,0.012127111355463663
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3584,1024,0.00847111145655314
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,2048,3072,0.01588800052801768
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,65536,256,0.009739555418491364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,5120,6144,0.02754666739039951
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2048,6144,0.009743111001120673
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,8192,5120,0.016927111479971144
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,32,2560,0.017961778574519686
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3584,2560,0.010572444233629437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,16384,7168,0.04875555634498596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,8192,1536,0.006759111252095964
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,10240,768,0.004951111144489712
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,64,10240,0.01218577805492613
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1024,3584,0.006269333263238271
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1536,7168,0.05539200041029188
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,4096,768,0.005623999983072281
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,512,5120,0.009384888741705153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,256,65536,0.06367999977535672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,2560,128,0.00527555536892679
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,128,768,0.00564444437623024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,5120,768,0.0035777779089079965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,4096,32,0.005941333456171884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,256,1536,0.006041777630647023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,128,512,0.002979555477698644
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,2560,7168,0.02678933408525255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,512,10240,0.050943109724256724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2560,65536,0.07006755802366468
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2048,65536,0.05343822307056851
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,12288,3072,0.012344000240166983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2560,2560,0.007821333077218797
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,3584,512,0.014412444498803882
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,8192,3584,0.013702222042613559
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,512,256,0.014592889282438489
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,128,2048,0.015416888727082146
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,6144,128,0.0055440002017551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,10240,65536,0.15096799532572427
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,8192,1536,0.025116443634033203
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3584,3072,0.011981333295504252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,32,6144,0.010022222167915767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,7168,2560,0.008387555678685507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,4096,2560,0.01811822255452474
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,4096,128,0.006047111004590988
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3584,10240,0.04046044415897793
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,256,1024,0.0032702222880389956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,12288,3584,0.018179555733998615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2048,2560,0.005567111074924469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,5120,2048,0.007080888582600488
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,12288,3072,0.016403555870056152
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,16384,2560,0.012549333274364471
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,8192,64,0.006259555617968242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,12288,12288,0.03447644578086005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,64,10240,0.011910222470760345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,1024,2048,0.015504888362354703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,7168,1024,0.006280888699822956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,65536,1024,0.47099023395114475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,12288,32,0.0028168888141711554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,10240,3584,0.013436444103717804
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,1024,16384,0.10409155819151138
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,3072,1024,0.017217778497272067
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,768,4096,0.006083555519580841
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,16384,1536,0.013027555412716336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2560,12288,0.026913776993751526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,8192,1536,0.00665244460105896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,64,5120,0.02030755579471588
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,16384,1536,0.010093332992659675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,32,32,0.008009778128729926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,128,12288,0.013491555220550962
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,4096,7168,0.011071999867757162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3584,5120,0.01276088919904497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,16384,512,0.021240888370407954
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2048,256,0.002558222247494592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,1536,10240,0.04426133301522997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,128,768,0.013129777378506131
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,7168,5120,0.01348977784315745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2560,3072,0.007975999679830339
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,2048,1536,0.006509333435032103
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,6144,1024,0.0046826667255825466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,16384,5120,0.02125599980354309
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,64,65536,0.0598906675974528
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,768,128,0.005976888868543837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,1536,1536,0.00629688882165485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,256,12288,0.01348977784315745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,65536,4096,1.36571470896403
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1024,256,0.011437333292431302
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,128,6144,0.02944533361328973
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,3072,6144,0.02723377777470483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1536,1536,0.006284444282452266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,10240,16384,0.041859553919898145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,128,7168,0.009758222434255812
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,65536,768,0.01463733282354143
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,65536,5120,0.13356178336673313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,3584,12288,0.10585066345002915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,768,1536,0.04064888755480448
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1024,4096,0.007728888756699032
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,6144,1024,0.006102222121424145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,6144,128,0.0052826665341854095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,6144,2048,0.0068604445291890045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,1536,2560,0.006976889239417181
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,256,7168,0.009679999616410997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,32,5120,0.015260444747077094
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,12288,128,0.007336888876226213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,512,10240,0.023781332704755995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,3072,2048,0.020040889581044514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,3072,12288,0.017100445098347135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,5120,5120,0.013045333325862885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1536,3584,0.008355555435021719
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,768,3584,0.010836444795131683
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,32,65536,0.04555999901559618
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,5120,6144,0.02276355524857839
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,64,7168,0.005595555735958948
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,128,256,0.0022444443570242985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3072,3072,0.07310933536953397
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,768,256,0.005623111294375525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,12288,32,0.003924444317817688
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,256,3072,0.00544888898730278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,12288,768,0.02753688891728719
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,16384,4096,0.023357333408461675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,4096,6144,0.010383110907342698
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3584,32,0.005612444546487596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,256,16384,0.020041777027977836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3072,12288,0.021951110826598272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,4096,32,0.0019031111150979996
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,12288,512,0.024851555625597637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,16384,32,0.006024888820118374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,5120,2560,0.0075457774930530125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,2048,5120,0.00870577742656072
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,12288,10240,0.05743555227915446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3584,512,0.028318223026063707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,12288,768,0.01956355571746826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,16384,5120,0.0672266681989034
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,4096,32,0.01178311142656538
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,16384,512,0.022688888841205176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,64,3584,0.004927111168702443
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,4096,5120,0.17628355820973715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,65536,3072,1.1025937398274739
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,8192,8192,0.015942222542232938
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,65536,128,0.3078960047827827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,16384,12288,0.035487112071779035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1536,10240,0.01312444441848331
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,128,768,0.00564533347884814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,256,1536,0.013129777378506131
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,512,128,0.005673777725961473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,7168,2560,0.007070221834712558
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,5120,32,0.016970665918456185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,1536,128,0.005255110975768831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,512,2048,0.006650666809744305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,1536,2560,0.007654222349325816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,6144,16384,0.07870755592981975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,7168,1536,0.018017777138286166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,16384,16384,0.13555199570126003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,4096,65536,0.23423287603590223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,8192,2560,0.016688000824716356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,5120,6144,0.036512000693215266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,1024,16384,0.03053599927160475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,4096,7168,0.01332711097266939
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,7168,7168,0.012983111043771109
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,256,8192,0.020322667227851022
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3584,16384,0.022094221578703985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3072,512,0.0028595555987623003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,7168,4096,0.03818044397566054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,10240,3584,0.03742488887574937
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,10240,1024,0.09894044531716241
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,4096,10240,0.05114755696720547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,128,6144,0.010275555153687796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3584,10240,0.3315155506134033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,512,768,0.00629688882165485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,2560,1024,0.01955200069480472
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,32,512,0.0035573335157500375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,512,768,0.005609777652555042
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,512,4096,0.016879111528396606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,7168,7168,0.015863110621770222
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1024,65536,0.02491644521554311
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,256,768,0.0029297777348094513
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,12288,3584,0.014540443817774454
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,512,6144,0.009359111388524374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,256,1536,0.006839111033413146
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3584,12288,0.045850667688581676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,65536,6144,0.6406639946831597
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,2560,6144,0.03824444280730353
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2048,3072,0.006329777754015393
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,5120,2048,0.006672888994216919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,64,3072,0.006976000136799282
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3584,12288,0.0176862229903539
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,4096,768,0.00452888881166776
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,65536,1024,0.0239955551094479
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,5120,1024,0.003936000168323517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2048,2048,0.007709333466158972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,8192,8192,0.022239999638663396
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,1024,7168,0.047263112333085805
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,65536,7168,0.16196978092193604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,4096,12288,0.03572266631656223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,7168,4096,0.014156444205178155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,10240,7168,0.11100622018178304
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,16384,5120,0.13885422547658285
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,12288,2048,0.012727999852763282
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,128,16384,0.006720888945791457
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1536,3072,0.005709333138333426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,16384,3584,0.2803697850969103
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,12288,6144,0.029626667499542236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2560,256,0.002879111096262932
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,512,4096,0.008325333396593729
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,1536,7168,0.009699555734793345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,3072,7168,0.02316444449954563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,7168,3072,0.03319822086228265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2048,1536,0.00666933341158761
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1024,512,0.00573333352804184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,3584,5120,0.011759111450778114
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,64,3584,0.007664889097213745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,512,1536,0.00665866666369968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,7168,2048,0.017165333032608032
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,128,8192,0.01834488908449809
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,3584,256,0.016557332542207506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,768,512,0.005377777748637729
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,2048,6144,0.018551111221313477
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,3072,3072,0.01799199978510539
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,7168,4096,0.013802666631009845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,7168,16384,0.07644177807701959
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,65536,4096,0.09554755687713623
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,4096,768,0.017066667477289837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,4096,1536,0.06463999880684747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1024,32,0.0021946666141351065
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3584,5120,0.011708444191349877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,32,3584,0.011766222616036734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,64,7168,0.0573235551516215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,5120,128,0.006631111105283101
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,3584,7168,0.02438755498992072
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,10240,512,0.019261333677503798
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,65536,3584,0.10829955339431763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2048,64,0.0022035555707083806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,256,12288,0.014194667339324951
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,1536,4096,0.008452444440788692
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,512,8192,0.010751999914646149
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,6144,65536,0.0770311090681288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,16384,128,0.0606222218937344
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,65536,512,0.012327999704413943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,16384,2048,0.013119999733236102
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,8192,128,0.0031964443624019623
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,6144,3072,0.014007111390431723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,3584,16384,0.07450844181908502
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,5120,7168,0.18807466824849448
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2048,12288,0.028436445527606543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,12288,3072,0.016732444365819294
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,12288,6144,0.029468443658616807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,128,4096,0.008052444292439355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,10240,64,0.006309333360857434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1024,256,0.006628444625271692
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,2560,65536,0.16782222853766549
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,12288,3072,0.03257511059443156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,8192,65536,0.08786488903893365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,512,16384,0.08194933334986369
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,128,10240,0.012454222473833295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,768,12288,0.013425778183672162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3072,16384,0.47422133551703555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,6144,4096,0.014498665928840637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,3072,6144,0.009286222358544668
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,64,3072,0.01852888862291972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3072,3584,0.008303999900817871
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2048,65536,0.07737600141101413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,1024,2048,0.007999999655617608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,6144,32,0.018244443668259513
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1024,4096,0.011389333340856763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,3584,3072,0.007284444239404466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,256,16384,0.026456000076399908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,7168,7168,0.020789333515697055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,7168,768,0.005940444353553984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,5120,65536,0.34175732400682235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3584,4096,0.011424889167149862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,2560,256,0.005429333282841577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2560,6144,0.011756444142924415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,512,128,0.013175999952687157
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,10240,1536,0.008016000191370646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,2048,12288,0.1545999977323744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,10240,32,0.010016000105275048
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3072,65536,0.06419289112091064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,128,1536,0.008295111358165741
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,5120,8192,0.02197066611713833
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,3584,65536,0.24015556441413033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,128,1536,0.00797333319981893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1536,65536,0.06090488698747423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,10240,7168,0.02316355539692773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,65536,3584,0.10242666800816853
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,2048,10240,0.01148088855875863
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,7168,5120,0.013288000391589271
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,10240,65536,1.0377848943074544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,2048,65536,0.14565332730611166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3072,2048,0.03052266769938999
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,7168,256,0.024313777685165405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,128,2048,0.0045422220395671
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,12288,12288,0.43764355447557235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,65536,256,0.011519999967681037
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,8192,10240,0.2693520122104221
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,65536,4096,0.1442426708009508
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2048,16384,0.28480177455478245
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3072,2048,0.007360888852013483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,16384,3584,0.016568000117937725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,16384,5120,0.021006221572558086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,8192,16384,0.0557546681827969
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,256,12288,0.023021333747439917
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1024,3072,0.006877333339717653
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2048,6144,0.012799999780125089
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,3072,7168,0.04703999890221489
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,10240,10240,0.2630942132737902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,6144,256,0.01684266659948561
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,512,1536,0.008051555189821456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,7168,10240,0.24887998898824057
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,7168,12288,0.03718222181002299
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,8192,768,0.006259555617968242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,1536,512,0.014348444011476306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,1024,3072,0.01819022165404426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,768,65536,0.02086044516828325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,10240,5120,0.02917333443959554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,12288,512,0.02720622221628825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,8192,7168,0.04009866714477539
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,5120,768,0.005949333310127258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,256,5120,0.015504888362354703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,6144,12288,0.018272888329293992
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1536,3584,0.0063031112982167145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2560,768,0.0032088889016045462
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,1024,3584,0.007344000041484833
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,1536,3072,0.01586577792962392
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,65536,2048,0.13083377149369982
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,6144,2048,0.10242666800816853
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,16384,512,0.022724444667498272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,256,5120,0.009034666750166152
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,7168,256,0.015790222419632804
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,128,1024,0.005941333456171884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,12288,1536,0.008059555457697975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,4096,5120,0.022330666581789654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3072,5120,0.00943911075592041
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,16384,6144,0.019925332731670804
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,32,7168,0.004514666481150521
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2560,512,0.006716444260544247
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2048,7168,0.06682755549748738
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,8192,3584,0.011728000309732227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,32,10240,0.0120319997270902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,7168,12288,0.061066667238871254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,1024,1536,0.006287999865081575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,10240,5120,0.01591022147072686
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,64,16384,0.0158595558669832
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,4096,4096,0.03726666503482395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,256,7168,0.01036177741156684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,5120,3072,0.007600888609886169
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2560,3584,0.007661333514584436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,7168,3072,0.01257155587275823
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,5120,4096,0.009076444639099969
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,2560,5120,0.017992888887723286
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,8192,1536,0.006870222174459034
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,65536,3584,0.09776710801654392
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,256,2560,0.006812444577614467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,7168,1024,0.01862844493654039
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3584,2560,0.00602311102880372
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,6144,12288,0.020252444677882724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3584,128,0.005590222362014983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,7168,256,0.003962666624122196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,10240,2560,0.012840888566441007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,256,10240,0.011777777638700275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,768,12288,0.013152889079517789
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,256,4096,0.0046728890803125174
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,7168,512,0.002924444360865487
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,256,3584,0.0148026661740409
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,8192,2560,0.02474755545457204
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1536,256,0.009701333112186855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2048,8192,0.012824888858530255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,2560,2048,0.006588444527652528
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,256,16384,0.005884444548024072
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,10240,7168,0.020506666766272653
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2560,3584,0.00798044436507755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,6144,512,0.0028906667398081887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1024,16384,0.011443555355072021
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1024,2560,0.004927999857399198
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,1024,128,0.005240889059172736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,3584,512,0.015204444527626038
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,5120,5120,0.012413333687517377
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,6144,1024,0.03403910994529724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,512,4096,0.014677332507239448
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1536,8192,0.013003555436929068
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,65536,3584,2.3046275244818792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,5120,16384,0.023450666003757056
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,7168,3072,0.023616888456874426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,512,256,0.015277332729763456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,65536,10240,0.24019643995496961
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,64,2560,0.005205333232879639
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,128,512,0.005632888939645555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,256,6144,0.009098666409651438
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,768,1536,0.003951999876234266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,3072,6144,0.009310222334331935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,512,10240,0.006990222467316522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,10240,3072,0.011298666397730509
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3072,2048,0.004917333523432414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,8192,512,0.005592000153329637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2560,12288,0.01769333415561252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,512,256,0.0046364445653226645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,3584,6144,0.00905511114332411
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,256,65536,0.2076488865746392
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,256,12288,0.022900443938043382
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,16384,768,0.06789866420957777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1024,3072,0.007998222278224098
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,12288,3072,0.024618667032983568
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,8192,65536,0.09366222222646077
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,65536,6144,0.05520533190833198
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1536,12288,0.01443911095460256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1536,512,0.007592889169851939
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,6144,6144,0.011415111521879831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,7168,16384,0.03800977932082282
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,65536,256,0.023154666026433308
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,1024,5120,0.022355554832352534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,3072,10240,0.014339556296666464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,256,7168,0.017871111631393433
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,16384,65536,0.31915201081169975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,128,10240,0.005615111026499007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,65536,2560,0.11059733231862386
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1024,5120,0.008732444710201686
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,4096,8192,0.014588443769348992
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,3072,1024,0.0058337777025169795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,512,2560,0.0069751110341813825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,7168,7168,0.028618666860792372
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,8192,7168,0.2771848837534587
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,7168,10240,0.02036088870631324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,512,1536,0.006342222293217977
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,512,64,0.0023111111174027124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1536,2560,0.004976888911591636
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,6144,6144,0.5950409041510688
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,2048,768,0.016590222716331482
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,10240,10240,0.06928710805045234
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,12288,2048,0.027581334114074707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,128,12288,0.022678222921159532
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,512,10240,0.01239111108912362
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,6144,2560,0.025644444757037695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,1024,768,0.005669333454635408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,512,64,0.005578666511509154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,128,2560,0.008656000097592672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,2048,1536,0.033165332343843244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,3072,16384,0.04260622130499946
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3584,1024,0.014182221558358936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,8192,256,0.003007111036115222
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,128,4096,0.014845333165592618
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1536,128,0.0021893332401911416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,3584,2048,0.0063075555695427795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,3072,12288,0.017287999391555786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,65536,6144,0.05587110916773478
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,6144,5120,0.016653334101041157
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,8192,256,0.017228444417317707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2048,16384,0.017671111557218764
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,6144,3072,0.026181333594852026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,768,1024,0.01350844403107961
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2560,2560,0.005336888962321811
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,64,3584,0.007556444240940942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2048,3584,0.008979555633332994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,768,12288,0.019651555352740817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,512,32,0.0022257777551809945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,8192,256,0.014320888453059726
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,10240,256,0.0028799999919202593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,16384,128,0.01144977741771274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,5120,6144,0.012553777959611682
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,5120,3584,0.02161866592036353
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,7168,512,0.012801777985360889
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,16384,256,0.04477511180771721
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3584,3584,0.007625777688291337
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,8192,4096,0.013572444518407186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,768,512,0.005274666680230035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,768,64,0.005947555518812603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,256,3584,0.004927111168702443
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,128,12288,0.014070222775141398
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,12288,8192,0.06706311305363973
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,16384,2048,0.04820800158712599
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,2560,4096,0.008415111237102086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,16384,8192,0.057068447271982826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3584,3072,0.007061333292060428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,64,16384,0.03358666764365302
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,4096,2560,0.025959110922283594
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,8192,2048,0.12635733021630183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,6144,5120,0.06052622530195448
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2048,10240,0.015190222197108798
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,7168,768,0.003842666745185852
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,1024,16384,0.016206221448050607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,7168,10240,0.0385173327393002
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,256,10240,0.032348444064458214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,32,2048,0.008378666308191087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,256,16384,0.02680177821053399
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,512,7168,0.020294222566816542
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,12288,32,0.006044444524579578
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,4096,10240,0.01604977746804555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,512,512,0.012457778056462606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,7168,256,0.005929777605666055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,6144,512,0.0058346668051348785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,16384,1536,0.027884443600972492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1024,8192,0.01107911103301578
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,10240,4096,0.02859555681546529
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,5120,1536,0.018178666631380718
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2048,65536,0.06920088662041558
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3584,256,0.0028515555378463534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,1024,4096,0.009677333136399588
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,128,8192,0.018544889158672757
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2560,12288,0.014710222681363424
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,512,16384,0.0335804455810123
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,4096,128,0.016163556112183463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,3072,128,0.014768888552983602
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,6144,5120,0.021707556313938562
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,2048,8192,0.023384888966878254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,8192,2048,0.009006222089131674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,768,7168,0.010062221851613786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,65536,4096,0.052813334597481616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,512,4096,0.038559112283918597
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,65536,2560,0.026736888620588515
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2560,6144,0.010806221928861407
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,16384,12288,0.12670311662885878
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,5120,7168,0.05297866794798109
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,2560,512,0.013732444081041547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,4096,5120,0.032646222247017756
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,4096,768,0.018188445104493033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1536,2560,0.04105510976579454
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,3584,12288,0.029743111795849268
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,5120,1536,0.006248888870080312
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,2048,8192,0.039826667971081205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,256,7168,0.018350222044520907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1024,4096,0.018208000395033095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1536,2048,0.035000890493392944
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,128,1024,0.011699555648697747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,65536,12288,0.12815822495354548
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3072,10240,0.017468444175190397
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,5120,65536,0.104312002658844
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,768,12288,0.030924446052975122
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2560,128,0.002540444541308615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,6144,3072,0.01035644445154402
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,10240,768,0.008030222521887885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,7168,2560,0.01977333260907067
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,3584,2048,0.021407110823525324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,2560,12288,0.16939022805955675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,7168,1024,0.012074666718641916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,32,1024,0.0032257777121331957
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,512,5120,0.008339555727110969
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,7168,2048,0.019896888070636325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,5120,7168,0.12025154961480035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2560,7168,0.013455111119482251
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,6144,8192,0.015265777707099915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,128,2560,0.007311111523045435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3072,32,0.005607999861240387
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,512,3072,0.015465777781274585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,128,16384,0.015825778245925903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,32,5120,0.004222222086456087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,64,7168,0.02037244372897678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,5120,768,0.022277333670192297
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,65536,7168,0.5977324379814996
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,65536,512,0.04405155446794298
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3584,10240,0.19104888704088
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,10240,768,0.00871288859181934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,1024,65536,0.06249599986606174
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,2560,3584,0.0075804442167282104
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,3584,256,0.016466667254765827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3072,12288,0.030926220946841772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,16384,5120,0.017495999733606975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,7168,512,0.017629333668284945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,5120,1536,0.013023110727469126
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,256,4096,0.008373333348168267
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2048,8192,0.012829333543777466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1536,512,0.01683111157682207
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,7168,512,0.0028826666788922418
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,6144,1024,0.005978666659858491
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,5120,1024,0.006295111030340195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,65536,32,0.0042364444169733255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,6144,65536,0.12665600246853298
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,32,7168,0.005623999983072281
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,32,512,0.005895111295912001
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,65536,2560,0.0841582218805949
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,16384,10240,0.028229332632488672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,512,4096,0.008372444245550368
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,128,2048,0.0063324446479479475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,6144,768,0.003528888854715559
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,8192,6144,0.01907199952337477
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3584,512,0.008263111114501953
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,6144,1536,0.006262222097979651
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,10240,7168,0.019671110643280875
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,6144,7168,0.01926488843229082
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2560,1536,0.006268444574541516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1536,8192,0.010947555303573608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,32,32,0.005569777968857024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,256,768,0.005932444499598608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,3072,3584,0.007974222302436829
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,3584,256,0.016562667157914903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3584,8192,0.01481244464715322
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1024,6144,0.012230222423871359
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,7168,2560,0.015239111251301236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,6144,4096,0.013211555778980255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,10240,3584,0.03379377722740173
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,12288,6144,0.018057778477668762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2560,256,0.003563555578390757
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,512,4096,0.016536888149049547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,65536,768,0.03430666526158651
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,6144,768,0.006230222268237009
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,8192,12288,0.020223110914230347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,16384,8192,0.04660888844066196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,10240,6144,0.017703110973040264
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,4096,5120,0.010168888502650792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,65536,10240,0.10474577877256606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,256,32,0.0018568889548381169
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,128,2560,0.004924444274769889
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,7168,1536,0.009039999710188972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,2048,3584,0.008009778128729926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,7168,65536,0.13503466712103948
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,5120,256,0.0031946665710873077
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,12288,6144,0.6789884567260742
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,6144,5120,0.011266666981909009
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,3072,1536,0.006256000035338932
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,65536,1536,0.02402399977048238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,5120,32,0.002540444541308615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,5120,128,0.009036444127559662
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,2048,8192,0.0233271105421914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,5120,12288,0.020742222666740417
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,768,4096,0.006004444426960415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,65536,8192,0.07218755616082086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2048,1024,0.0032604444358083936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1536,65536,0.08251021967993842
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,4096,6144,0.09395466910468207
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,12288,10240,0.028275556034511987
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,65536,7168,0.339737786187066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,12288,2560,0.020853334003024634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,256,768,0.005596444424655702
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,65536,12288,0.12818843788570827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,6144,128,0.014871110518773397
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,64,4096,0.01311466677321328
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,10240,2048,0.03779200050565932
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,1024,256,0.005222222043408289
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3072,128,0.020239111449983384
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,32,1536,0.006246222390068903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,65536,16384,0.16995110776689318
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,16384,128,0.006281777802440856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,128,512,0.005688000056478713
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,2048,1024,0.016857778032620747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,128,12288,0.013991999957296582
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3584,1536,0.0038817777401871155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,12288,1024,0.007638221813572778
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,6144,768,0.036307556761635676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1024,128,0.006937777830494775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,16384,64,0.008008889026112026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,512,12288,0.017331555485725403
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,768,6144,0.010568000376224518
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,768,6144,0.02653955585426754
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,6144,65536,0.07710222403208415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,1024,2560,0.00685422205262714
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,32,4096,0.012087110843923358
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,10240,3072,0.013951111170980664
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2560,4096,0.00833511104186376
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,6144,3072,0.010451555252075195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,3072,3072,0.007154666715198093
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,4096,512,0.016948444975747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,12288,1024,0.022642667094866436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,64,4096,0.00424799985355801
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,3584,10240,0.06881688700781928
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,12288,2048,0.02832088867823283
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3072,2048,0.004867555366622077
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,512,768,0.005971555494599872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,1536,3584,0.007978666987684038
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,4096,4096,0.00797155582242542
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,768,65536,0.028299556838141546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,65536,64,0.02335111134582096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,16384,256,0.013572444518407186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,2048,10240,0.018351111147138808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1024,2560,0.007069333559936947
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,3072,512,0.014126222994592456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,256,3072,0.01423466702302297
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,3072,10240,0.06322044134140015
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,256,2048,0.006269333263238271
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,10240,6144,0.3597262170579698
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3584,2560,0.006986666884687211
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,7168,10240,0.020084444019529555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,32,2048,0.00425511101881663
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,8192,3584,0.052383999029795326
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,4096,3584,0.025009777810838487
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1536,1536,0.0047493332790003884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,10240,12288,0.030159999926884968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,64,7168,0.00535999983549118
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,128,4096,0.02093155516518487
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,6144,16384,0.025752888785468206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,7168,512,0.0031644445326593188
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,16384,5120,0.040942221879959106
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,8192,4096,0.01312622262371911
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,768,6144,0.011034666664070554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1024,512,0.005963555640644497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,768,16384,0.17488088872697619
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,12288,4096,0.11993243959214951
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,5120,3072,0.12724355856577554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,8192,256,0.006931555353932911
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,4096,12288,0.024440889557202656
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3584,10240,0.023888889286253188
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,128,512,0.005582222094138463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,2048,6144,0.00964355551534229
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,10240,3072,0.019136000010702345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,5120,7168,0.03604355454444885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,7168,1536,0.006640000061856375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,6144,3584,0.023207111491097346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,8192,3072,0.023317333724763658
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1536,3072,0.0063271112740039825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,5120,10240,0.025199999411900837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,128,6144,0.006261333409282897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,512,512,0.012089778151777057
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,16384,12288,0.07451911105049981
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2048,3072,0.013024888932704926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,6144,3584,0.01792711185084449
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,6144,32,0.0021626665774318907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,768,2048,0.00462666650613149
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,5120,32,0.005939555664857228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,12288,768,0.012308444413873883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,768,2560,0.01836266616980235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,7168,8192,0.05842489004135132
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,768,1024,0.013454222016864352
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,16384,12288,0.043783999151653714
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,32,12288,0.005584888988071018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1024,3072,0.005626666463083691
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,128,512,0.011737777955002256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2560,12288,0.017697778013017442
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,6144,7168,0.04362044400639004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3072,32,0.006042666733264923
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,10240,12288,0.06928710805045234
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,8192,10240,0.02201333310869005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,6144,64,0.00582311095462905
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,8192,3584,0.03140799866782294
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,128,256,0.005693333430422678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,4096,1024,0.050322665108574756
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,10240,4096,0.029639111624823675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,7168,16384,0.04439199964205424
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,6144,12288,0.03662577933735318
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,1536,8192,0.03695911169052124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2560,65536,0.05460177858670553
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,6144,10240,0.03748710950215658
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,12288,2560,0.11425244808197021
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,16384,512,0.006698666761318843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,768,128,0.003560888684458203
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1536,8192,0.01850044396188524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,12288,3072,0.012810666528013019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3584,1536,0.006694222076071634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3584,64,0.0018800000349680583
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3584,1536,0.004850666556093427
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,256,2560,0.009376889301670922
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,65536,1024,0.016121778223249648
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,8192,768,0.006923555499977536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,256,1024,0.0032302221904198327
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,32,3072,0.007296889192528195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,64,10240,0.012737777498033313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,65536,512,0.01220533334546619
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3584,64,0.011711111499203576
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,1024,7168,0.07873066928651598
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,10240,1024,0.0546248886320326
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,65536,1024,0.09193955527411567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,5120,1536,0.022126222650210064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,64,3072,0.0075662218862109715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,12288,2048,0.017350221673647564
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,2560,1536,0.025047999289300706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2560,3584,0.008660444782839881
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,512,16384,0.029126223590638902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,768,512,0.004184000194072723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,7168,10240,0.04936888813972473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,768,8192,0.050712890095180936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,10240,2560,0.025038222471872967
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,32,7168,0.005246222019195557
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1536,768,0.006255110932721033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1536,6144,0.010117332968446944
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3072,5120,0.012039110892348819
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,3584,6144,0.013421333498424955
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,256,128,0.0021422221842739317
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,7168,3584,0.03197688857714335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,4096,10240,0.025191111697090998
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,2048,1024,0.016564443707466125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,10240,1536,0.02424977719783783
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,1536,5120,0.0103511114915212
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,768,4096,0.00556355549229516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,32,10240,0.005600888695981767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,65536,5120,0.0757840010854933
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1536,4096,0.0075146663520071245
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,7168,12288,0.023824888798925612
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,64,256,0.002190222135848469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,512,1024,0.003162666741344664
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,16384,512,0.058042665322621666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,256,12288,0.022946667340066697
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3584,2048,0.0068400001360310455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,1536,512,0.013243555194801755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,768,512,0.01202133297920227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,7168,32,0.0032053333189752367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,10240,3584,0.0226275556617313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,2560,10240,0.015237333046065437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,4096,4096,0.010138666464222802
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,7168,3584,0.012861332959598966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,4096,4096,0.0355333321624332
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,16384,1536,0.010951999988820819
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,10240,2560,0.2460569010840522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,3584,12288,0.03068355719248454
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,512,2048,0.0045422220395671
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,64,10240,0.012290666500727335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,768,65536,0.0902622209654914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,512,1536,0.025412445267041523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2560,512,0.005881777654091517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,512,16384,0.0790497793091668
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,65536,1536,0.04126755396525065
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1024,12288,0.01532977819442749
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,256,6144,0.008009778128729926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,5120,8192,0.028997331857681274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,65536,4096,0.06819111108779907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,5120,3072,0.03448266784350077
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,512,8192,0.011392888923486074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,65536,8192,0.14264710744222006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1024,10240,0.06500177913241915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,10240,256,0.007224000162548489
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,16384,512,0.022679110368092854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,6144,64,0.013532444834709167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,3584,10240,0.015543111496501498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3584,64,0.005611555443869696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,4096,4096,0.009369778136412302
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3584,6144,0.022640888889630634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,1536,256,0.005577777822812398
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,4096,65536,0.08159289095136854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3072,2560,0.04054844379425049
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,256,3584,0.014479110638300577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1024,3072,0.005286222116814719
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,1024,12288,0.013420444395807056
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,65536,768,0.015689777003394235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2560,6144,0.039481778939565025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,512,7168,0.06683644321229723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2048,8192,0.011734222372372946
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,6144,32,0.0052577778697013855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,6144,6144,0.013952000273598565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,6144,65536,0.11616178353627522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2560,32,0.002157333410448498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,8192,256,0.0035182221068276297
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,128,8192,0.01069777790043089
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,6144,256,0.009472000102202097
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,1024,2560,0.015060444672902426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1024,1024,0.004655999855862724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3584,4096,0.00941777808798684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,4096,2560,0.0069395556218094295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,6144,12288,0.021640888518757288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,6144,3584,0.010760889285140567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,4096,768,0.007260444263617198
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,65536,12288,1.5905990600585938
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,7168,64,0.01441511180665758
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,6144,2560,0.007058666812049017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,512,65536,0.03601600064171685
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,32,256,0.002469333302643564
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,8192,128,0.021743110484547083
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,5120,3584,0.012561777399645912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,7168,64,0.0025031111306614345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,3584,16384,0.04311644368701511
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,768,12288,0.03815289007292853
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,8192,5120,0.014634667171372307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,16384,65536,0.16677244504292807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2560,5120,0.019117333822780185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,4096,768,0.0031831111345026228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3584,2048,0.009763555394278632
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,10240,5120,0.01568177839120229
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,6144,128,0.003882666842805015
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,32,32,0.005056000004212062
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,4096,5120,0.025349333882331848
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,512,2048,0.014094221923086377
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,6144,256,0.002875555513633622
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,8192,8192,0.19860533873240152
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,768,1536,0.014152889450391134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,16384,16384,0.05315555466545952
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2560,2560,0.006934222247865465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,3072,10240,0.013423999978436364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,3584,5120,0.008295111358165741
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,8192,3072,0.15721333026885986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,768,7168,0.033370667033725314
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,10240,2048,0.011152000062995486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2048,1536,0.006172444257471297
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1024,3072,0.005608888963858287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,6144,3072,0.009697777529557547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,16384,256,0.008328888979223039
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,768,4096,0.008034666379292807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,3072,8192,0.011490667031870948
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,5120,768,0.017680888374646504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,65536,2048,0.02504088812404209
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,3072,8192,0.024699555502997503
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,65536,3584,0.06882755623923408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,6144,256,0.023366222778956097
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,5120,6144,0.059347556696997754
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,768,1536,0.004911111046870549
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,6144,3584,0.01239999963177575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3072,32,0.010374222364690568
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,512,10240,0.024061333802011278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,256,1536,0.006233777850866318
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,6144,1024,0.004219555606444676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1024,2048,0.006566222343179915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,8192,512,0.017183999220530193
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,65536,2048,0.025433777107132807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,5120,4096,0.011056000159846412
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,6144,12288,0.018407111366589863
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1024,3072,0.023216888308525085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,10240,128,0.006270222365856171
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,16384,4096,0.0155102229780621
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,5120,512,0.005575110928879843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,12288,12288,0.0369759996732076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,7168,65536,8.070014953613281
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,768,3584,0.005687110953860813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,7168,6144,0.02765422231621212
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,512,4096,0.014458666245142618
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,768,256,0.017449777987268236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,3584,2048,0.0221413340833452
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,6144,5120,0.011031111081441244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,64,64,0.00666222224632899
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,8192,3072,0.010810666614108615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,16384,12288,0.038831111457612776
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,128,32,0.005580444302823808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,12288,7168,0.021368000242445204
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,8192,7168,0.17830310927497017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,256,768,0.015480000111791821
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,32,3584,0.005072000126043956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,32,768,0.00574933323595259
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3584,3584,0.018423999349276226
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,16384,2048,0.011872000164455838
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,4096,1536,0.006695111178689533
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,4096,7168,0.011335111326641507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,3072,2048,0.020227554771635268
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,5120,512,0.005608888963858287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,768,1536,0.00851822230550978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,8192,768,0.06421599785486858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,256,128,0.011928888658682505
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,128,1024,0.005688000056478713
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,6144,10240,0.0464897784921858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,256,8192,0.005767999837795894
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,32,65536,0.009505777723259395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,3072,512,0.01202400028705597
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,1024,6144,0.008789333204428354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,256,3072,0.014192000031471252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,32,16384,0.006593777901596493
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,16384,12288,0.061218665705786816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,512,512,0.009712888962692684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,65536,2048,0.392302221722073
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,6144,65536,0.3847084575229221
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,10240,768,0.0332640012105306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2560,4096,0.013559111290507846
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,64,3584,0.01181422256761127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,512,128,0.013348444468445249
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,3072,10240,0.05180533395873176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,6144,65536,0.11623199780782063
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,256,10240,0.020663999848895602
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,1536,7168,0.038719111018710665
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,2048,6144,0.02092711130777995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3072,4096,0.008052444292439355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,512,768,0.0028239999794297745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,16384,5120,0.018564444449212816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,256,7168,0.024695111645592585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,768,2560,0.01551822159025404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,128,3072,0.007111999723646376
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,6144,4096,0.008329778081840938
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,1024,6144,0.01777777738041348
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3584,2048,0.023772444989946153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,10240,4096,0.013993778162532382
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,768,7168,0.010141332944234213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2560,12288,0.017605332864655387
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,16384,12288,0.06267822451061673
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,10240,4096,0.015929778416951496
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,2560,12288,0.05071377754211426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2560,6144,0.0099982221921285
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,6144,1024,0.005947555518812603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,7168,512,0.004713777866628435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,8192,128,0.002838222309947014
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,128,128,0.0021475555582178966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,3072,128,0.015496888094478183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,16384,7168,0.035219556755489774
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,10240,2048,0.19584443834092882
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,6144,7168,0.0296088887585534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3072,32,0.006624889042642381
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,16384,1536,0.060792889859941274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,256,3072,0.00757244477669398
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3584,65536,0.07560177644093831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3584,3584,0.00756444450881746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,768,8192,0.03101955519782172
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,8192,4096,0.013799111048380533
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,7168,128,0.0029111111329661477
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,6144,16384,0.02314577831162347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,32,6144,0.005954666684071223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,32,128,0.0021768889079491296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3072,5120,0.01036888857682546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2560,12288,0.16339288817511663
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,6144,10240,0.018801778554916382
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,65536,3584,0.035974221097098455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,256,5120,0.008330666356616551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,7168,8192,0.01793066660563151
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,1024,65536,0.10106844372219509
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,128,768,0.01464000013139513
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,2048,8192,0.022707555029127333
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1536,3072,0.008303999900817871
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,12288,65536,0.22661866082085502
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,7168,6144,0.08919288714726765
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3072,12288,0.0191600008143319
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,4096,16384,0.06363022327423096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,65536,3072,0.09822399748696221
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3072,32,0.0022400000856982335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,5120,4096,0.01257066677014033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,7168,7168,0.013882666826248169
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,3584,10240,0.02933244572745429
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,256,512,0.005616000129116907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,16384,7168,0.02674755619631873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,12288,7168,0.023826667004161414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,256,6144,0.009033777647548253
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,16384,6144,0.051072001457214355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2048,768,0.005669333454635408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,6144,3584,0.007694222033023834
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,6144,1024,0.010114666488435533
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,2048,3584,0.02642222245534261
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2048,8192,0.013421333498424955
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,256,2560,0.007332444190979004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,64,5120,0.004927111168702443
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,5120,2560,0.006918222126033571
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,128,2560,0.013471110827393003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,512,128,0.01351466692156262
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,768,2048,0.006561777657932705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2048,5120,0.015303999185562134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,12288,16384,0.5698382059733073
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,128,3072,0.007304888632562425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,256,2560,0.005140444470776452
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,65536,1536,0.020842666427294414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,2560,8192,0.21942578421698677
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,256,2048,0.013440888788965015
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,768,1024,0.0035253332720862497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,256,3072,0.007303111255168915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,5120,7168,0.08462399906582302
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3072,6144,0.01273155543539259
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,4096,1536,0.03394577900568644
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,3072,4096,0.00797155582242542
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,128,16384,0.028838220569822524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,12288,6144,0.020558221472634208
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,8192,10240,0.07155377997292413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,7168,8192,0.2754586537679036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,8192,6144,0.015223110715548197
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,2048,2048,0.014808000789748298
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,6144,1536,0.021407999926143225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,1536,1024,0.005591111050711737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,65536,6144,0.19017688433329263
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1024,3584,0.012984889249006907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1024,3584,0.007755555212497711
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,3072,3584,0.04829422301716275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,2048,7168,0.009709333380063375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,256,32,0.001855111163523462
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,64,65536,0.053040888574388295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,4096,8192,0.021074665917290583
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,256,65536,0.06279466549555461
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,5120,2048,0.011577777564525604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,8192,8192,0.016598222984208
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1536,10240,0.02438755498992072
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,128,1536,0.0035146665241983202
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,4096,1536,0.006263999889294307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,128,1024,0.0081386665503184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1536,1536,0.02931555443339878
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,6144,128,0.005573333137565189
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,65536,2560,0.1389440033170912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2048,512,0.006632888896597757
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,10240,5120,0.018608000543382432
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,6144,2560,0.011407111254003314
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,256,10240,0.011333333121405708
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,6144,6144,0.012809777425395118
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2560,768,0.007653333246707916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,7168,3072,0.07718755801518758
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,768,512,0.005664888769388199
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,65536,1024,0.6308159828186035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,7168,768,0.018788443671332467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3072,65536,1.2410462697347004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2560,10240,0.028106666273540918
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,2048,3072,0.007686222592989604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,768,3584,0.013243555194801755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,6144,128,0.015184000134468079
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,64,3584,0.01145155562294854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,256,3584,0.007635555333561367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,8192,10240,0.05592177973853218
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,512,512,0.005581333405441708
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,16384,512,0.011624000138706632
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,3072,12288,0.041309333509869046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,4096,3072,0.007709333466158972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,256,3072,0.014760888285107084
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,128,65536,0.06209511227077908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3072,12288,0.02051911089155409
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,5120,5120,0.03291199935807122
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,16384,512,0.009348444640636444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,1024,512,0.02534399926662445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,12288,10240,0.17884533935123018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,10240,512,0.01962222158908844
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3072,512,0.00674222202764617
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,32,4096,0.018944889307022095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,7168,12288,0.044738666878806226
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,7168,4096,0.03300889001952277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,8192,16384,0.4198719925350613
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,768,2560,0.013567999833159976
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,6144,2560,0.013758222262064615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3584,128,0.005610666755172942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1536,512,0.0029208889851967492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,10240,3584,0.016898666818936665
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,512,65536,0.054959999190436475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,512,16384,0.016202666693263583
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,12288,3072,0.015592000550693937
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,3072,1024,0.014831999937693277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,768,6144,0.01959377858373854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3584,5120,0.012335999972290464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,3072,5120,0.05237422386805216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3072,5120,0.0107360002067354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,6144,4096,0.011083555718262991
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,16384,1536,0.010037333601050908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,7168,2048,0.008231110870838165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,128,512,0.002540444541308615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,512,4096,0.015857777661747403
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,4096,2560,0.054493334558275014
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,16384,8192,0.05863733424080742
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,7168,7168,0.020061333974202473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,1024,768,0.016501333978441026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,16384,512,0.006594666590293248
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,768,12288,0.030792888667848375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,3072,768,0.0057884446448749965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,65536,3072,0.031360887818866305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,128,65536,0.009707555174827576
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3072,7168,0.013049778011110095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,256,16384,0.034691555632485285
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,8192,6144,0.03568266828854879
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1536,5120,0.0083155557513237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,2048,65536,0.06319555309083727
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,10240,6144,0.019996444384257
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,3584,10240,0.03294488787651062
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,64,1536,0.01332800007528729
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,12288,65536,0.2568826675415039
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,2048,3584,0.007616000043021307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,64,4096,0.022634666826989915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,256,3072,0.014543111125628153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3584,16384,0.11412889427608913
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,256,768,0.005889777921968036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,8192,10240,0.10551466544469197
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,5120,2560,0.023077333966890972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,512,256,0.0052560000783867305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,2560,8192,0.032059556908077665
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,64,3584,0.006664000037643645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,768,16384,0.011352889239788055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1024,32,0.0028764444092909494
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,16384,768,0.010462221999963125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2560,2560,0.021700445148679946
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,7168,256,0.01690933273898231
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,2048,128,0.020298666424221463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,6144,512,0.003567999849716822
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,7168,128,0.007712889048788283
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,1536,6144,0.032630221711264715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2048,3072,0.007328888608349695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,5120,16384,0.031854222218195595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,4096,128,0.005312888986534542
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,65536,512,0.05766044722663032
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,3584,12288,0.03059200114674038
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,512,10240,0.013330666555298699
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1024,256,0.002530666689078013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,16384,256,0.006604444649484422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,64,6144,0.009398221969604492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,768,4096,0.007045333584149678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,12288,768,0.00941422250535753
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,512,256,0.0022453332526816265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,32,3584,0.005247999810510212
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,5120,3584,0.007672888537247975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1536,65536,0.03827022181616889
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,2048,7168,0.022290666898091633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,512,2560,0.0069271110826068455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2560,32,0.0024577778660588795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,6144,256,0.01854222185081906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,128,768,0.005541333307822545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,32,512,0.006596444381607904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,6144,12288,0.01818933255142636
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,512,7168,0.013257777525318993
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,8192,768,0.007982222570313348
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,768,65536,0.062272886435190834
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,7168,512,0.005251555393139522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2048,1024,0.006228444476922353
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,512,5120,0.01719111038578881
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1536,64,0.0058168888919883305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,128,5120,0.008681777450773451
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,256,10240,0.012806221842765808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,5120,5120,0.024350222614076402
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,4096,5120,0.021998221675554912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,16384,2048,0.011114666859308878
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,128,16384,0.025723555021815833
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,16384,1536,0.0306346681382921
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,5120,2048,0.00849333322710461
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,4096,8192,0.23109867837693956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,1536,2560,0.0069617778062820435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,128,6144,0.009700444837411245
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,65536,8192,0.09373599953121609
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,64,5120,0.009375111096435124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,65536,128,0.01960000064637926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,256,4096,0.03271822134653727
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,1536,3584,0.008206222620275285
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1536,2048,0.035473777188195124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,12288,768,0.0069288888739215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,6144,5120,0.20218755139244926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,768,768,0.0029324444217814338
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,4096,1536,0.011235555840863122
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,768,4096,0.03773422373665704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,65536,8192,0.07360711362626818
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,32,512,0.002534222271707323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,1024,3584,0.0199955552816391
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,6144,512,0.017510222064124215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,3072,3584,0.007659555309348636
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,128,6144,0.009073778159088558
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,128,64,0.0021831111775504216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3072,10240,0.024307555622524683
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,1024,768,0.015258666541841296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,4096,2048,0.0069777775141927935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,3584,10240,0.037608888414171
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,10240,6144,0.0566506650712755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,8192,4096,0.033668445216284856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,7168,32,0.003531555748648114
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3072,8192,0.014156444205178155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1536,1024,0.006267555471923616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,16384,16384,0.053156445423762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,1024,4096,0.020655110478401184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,6144,16384,0.24222755432128906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,1536,3072,0.007617777420414819
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,512,4096,0.01629066632853614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2048,10240,0.026050666968027752
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,5120,5120,0.02405422263675266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,512,2560,0.010099555883142684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,4096,512,0.006524444454246097
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,7168,7168,0.0295466681321462
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,2560,128,0.0053546664615472155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,4096,512,0.015164444843928019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2048,2560,0.007511999871995714
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,16384,512,0.004743111216359668
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,2048,512,0.013817778064144982
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3072,128,0.005997333261701796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1024,12288,0.018446221947669983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2048,7168,0.14823110898335776
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,4096,3584,0.014354666074117025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,512,128,0.005631999837027655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,8192,2560,0.008700444466537898
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,7168,2560,0.028194665908813477
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,768,1024,0.003228444399105178
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,4096,10240,0.02401777770784166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,32,1024,0.0029404444826973807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,1024,6144,0.021680000755521987
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,65536,768,0.020679111282030743
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,6144,10240,0.015729778342776828
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,4096,4096,0.010687111152542962
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,32,3584,0.007869333028793335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,32,3072,0.0042204442951414315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1024,128,0.005616888817813661
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1024,256,0.005365333209435145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,256,256,0.006634666687912411
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1536,32,0.0069351109365622205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,7168,7168,0.03144888745413886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3584,512,0.01643200053109063
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,8192,3584,0.02183377742767334
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,64,512,0.002529777793420686
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,512,1536,0.003920000046491623
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1536,2560,0.007278222176763747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1024,2560,0.029502222935358684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,8192,1536,0.009325332939624786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3072,128,0.006617777877383762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,3072,2048,0.014322666658295525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,4096,3584,0.011544889046086205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3584,768,0.00396799999806616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,64,32,0.005972444597217772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,4096,12288,0.05443288882573446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,65536,1536,0.021900445222854614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,10240,2048,0.010963555839326648
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,128,16384,0.023644444015291
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,128,65536,0.1438302199045817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1536,6144,0.00979911122057173
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,512,65536,0.0661733349164327
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,128,1536,0.009366222553782994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,768,65536,0.06368355618582831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,768,65536,0.078639997376336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,65536,6144,1.7718373404608831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,128,4096,0.012427555190192329
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,8192,5120,0.014500444134076437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,768,6144,0.009340444372759925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,4096,12288,0.24125154813130698
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,6144,8192,0.023048889305856492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2048,3072,0.005961777849329843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,12288,128,0.006249777972698212
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,32,128,0.002141333288616604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,1024,3072,0.029135998752382066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,65536,16384,0.3018639882405599
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,5120,3584,0.017902221944597032
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,16384,512,0.006588444527652528
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,512,256,0.015488000379668342
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,128,7168,0.01032266683048672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,8192,10240,0.043172445562150746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3072,8192,0.014289778139856128
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,12288,8192,0.03764977720048692
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,768,10240,0.008300444318188561
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,10240,12288,0.176910228199429
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,10240,1024,0.02201866606871287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,256,10240,0.01203555530971951
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,16384,64,0.0021564445147911706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,1024,1024,0.013779555757840475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3072,8192,0.09999377859963311
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,128,65536,0.06903911299175687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,65536,16384,1.6083101696438258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,12288,3072,0.12366667058732773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,10240,1024,0.02197155521975623
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1536,2048,0.006493333313200209
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,4096,10240,0.14538844426472983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,4096,3072,0.007341333561473423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1536,4096,0.007982222570313348
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,7168,1536,0.006244444598754247
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,128,65536,0.1911688910590278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,8192,5120,0.013835555149449242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,256,128,0.00221422231859631
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1536,128,0.0035911111368073356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,128,3072,0.0069475554757648045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,32,10240,0.005942222144868638
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,1024,10240,0.01649777756796943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,768,10240,0.014610666367742749
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,6144,128,0.005559111220969095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1024,128,0.005571555760171678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,32,3072,0.00507377791735861
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,2048,7168,0.09593244393666585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,512,256,0.002566222308410539
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,256,1024,0.012812444733248817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1536,8192,0.06407466861936781
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,256,512,0.012075555821259817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,1536,2048,0.006528000036875407
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,3584,16384,0.021494223011864558
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3584,256,0.005994666781690385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,1024,3584,0.017870222528775532
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,5120,1536,0.018988443745507132
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,3584,12288,0.036537776390711464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,6144,3072,0.026428444517983332
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,512,10240,0.02158755560715993
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1536,256,0.004295999805132548
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,256,128,0.0029208889851967492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3072,3072,0.007391999993059371
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,6144,768,0.007656888829337225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,6144,3072,0.017650667164060805
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1024,6144,0.00981600003110038
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,5120,128,0.014838222000333997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,128,6144,0.05594311157862345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,256,10240,0.015941333439615037
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,6144,128,0.01991911066903008
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,256,16384,0.04626489016744825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,1024,16384,0.01626133256488376
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,16384,512,0.05539110965198941
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,10240,128,0.01792888840039571
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3072,16384,0.037019554111692644
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,768,3584,0.013169777890046438
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,6144,3072,0.010372444159454769
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,5120,2048,0.023064000738991633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,512,1536,0.0064693333374129395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,12288,256,0.0028417776856157514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,128,6144,0.017443555924627516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,10240,10240,0.02572622232966953
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,12288,1024,0.009688888986905416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,768,12288,0.02299822204642826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,8192,6144,0.018191999859280057
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,256,2048,0.00674222202764617
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2048,128,0.0021102221475707158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,768,6144,0.023356444305843774
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,10240,32,0.008917333351241218
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,256,3072,0.007735110819339752
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,5120,16384,0.27492978837754994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,65536,8192,0.14647111627790663
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3584,4096,0.008985777695973715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,7168,128,0.005610666755172942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,1536,256,0.01514400045077006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,3072,512,0.015075556106037565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,7168,2048,0.009504888620641496
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2560,12288,0.0347013341055976
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,5120,8192,0.014129777749379476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,6144,256,0.012829333543777466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,256,7168,0.025774222281244066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1536,8192,0.012989333106411828
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,32,16384,0.030850665436850652
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,16384,16384,0.06751200225618151
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,16384,3072,0.03271199928389655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,512,2560,0.014160888062583076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,7168,12288,0.024275556206703186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,3584,1536,0.009268444445398118
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,512,5120,0.008780444661776224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1536,4096,0.007981333467695449
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,256,12288,0.013447110851605734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,512,8192,0.027285334136750963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,1024,5120,0.016927111479971144
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,128,64,0.0018906665758954154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,3584,4096,0.028946667909622192
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,128,128,0.002181333386235767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,65536,2560,0.07988888687557645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,2048,1536,0.014204444156752693
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,512,5120,0.008380444513426887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,32,8192,0.00655733338660664
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,6144,256,0.005281777845488654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,3072,65536,0.09429066711001927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,12288,65536,0.15590043862660727
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,768,768,0.017056889004177518
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,1536,16384,0.031368000639809504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,65536,1024,0.07000444332758586
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,2560,5120,0.025546666648652818
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,12288,16384,0.05812888675265842
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,8192,10240,0.01831377711560991
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,10240,3072,0.03159644537501865
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,8192,8192,0.033486220571729876
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,4096,3072,0.01756533318095737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,12288,8192,0.026520000563727483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,10240,16384,0.04269599914550781
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,5120,128,0.004603555632962121
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,12288,3072,0.014067555467287699
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,768,1536,0.009683555199040307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,65536,2048,0.06306222412321302
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,7168,256,0.007923555870850882
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,65536,12288,0.5146729151407877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,65536,768,0.029476443926493328
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3584,6144,0.01519733336236742
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,65536,5120,0.07579822010464139
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,10240,65536,0.135352889696757
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3072,65536,2.2439742618136935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1536,128,0.011739555332395764
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,4096,5120,0.011063111325105032
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,10240,3584,0.01391999920209249
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,3584,3584,0.007677333222495184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,32,65536,0.126528885629442
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,768,768,0.0032462223122517266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,128,65536,0.08132888873418172
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2048,10240,0.015597333510716757
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1536,65536,0.06202310985989041
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,12288,256,0.006105777704053455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,5120,128,0.005582222094138463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3072,16384,0.020733333296246
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,256,12288,0.022673777408070032
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,10240,4096,0.014134221606784396
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,1536,2048,0.014442667365074158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,3072,3072,0.02377333409256405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,512,1024,0.006288888967699475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,16384,2560,0.023481777972645227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,64,6144,0.004572444491916233
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,4096,768,0.005932444499598608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,512,65536,0.062204447057512074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,64,64,0.007958222594526079
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2048,512,0.002578666640652551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,7168,1024,0.007717333734035492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,8192,128,0.0025537777692079544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,6144,3584,0.03158933255407546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2560,5120,0.010312000082598792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,8192,8192,0.024567999773555334
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,5120,768,0.0035911111368073356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1536,8192,0.019698666201697457
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,10240,3584,0.023013333479563396
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,8192,16384,0.048968887991375394
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3584,32,0.0021724444296624926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1536,10240,0.014800889624489678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,10240,1536,0.012121777567598553
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1536,7168,0.010306666294733683
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,6144,3072,0.028135110934575398
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2048,256,0.002536888958679305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,1536,65536,0.06294488906860352
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,16384,6144,0.03391822179158529
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,12288,3072,0.02814399864938524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,12288,12288,0.03140888942612542
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,2560,768,0.02830933199988471
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3072,10240,0.017455110947291057
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,2048,768,0.018180444836616516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,10240,2048,0.011023110813564725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,3072,512,0.005560889012283749
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,10240,10240,0.060812446806165904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,16384,2048,0.014452444182501899
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,2560,1536,0.011863999896579318
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,256,6144,0.023689778314696416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,256,512,0.009360888765917884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,768,3072,0.016552888684802584
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,32,1024,0.003581333491537306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,512,8192,0.01926933394538032
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,128,10240,0.020000000794728596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1024,65536,0.027230223019917805
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,1536,3072,0.015229332778188916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,768,2560,0.014843554960356818
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,768,3072,0.006999111009968652
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,4096,6144,0.009708444277445475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,1024,8192,0.019946666227446664
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,16384,16384,0.10416977935367161
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,6144,2048,0.01478666729397244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,1024,8192,0.15989332728915745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,256,1024,0.012416000167528788
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,3072,4096,0.011037333144081963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,65536,512,0.057137780719333224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,10240,6144,0.031585776143603854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,6144,10240,0.0325528879960378
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,7168,65536,0.0744204454951816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,256,65536,0.07941155301200019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,6144,16384,0.03671822283003066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,7168,12288,0.018221333622932434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3584,65536,0.07767822345097859
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,65536,12288,6.902875688340928
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,7168,3072,0.05436355537838406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,4096,12288,0.01716444392999013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,1024,128,0.01718755563100179
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,65536,1536,0.0507244434621599
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,6144,3072,0.14194044801923963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,4096,2048,0.06611822048823039
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,16384,2560,0.02884711159600152
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,8192,2048,0.021644444929228887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,6144,6144,0.01737777723206414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,2048,2048,0.006602666858169768
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,128,65536,0.11729333135816787
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,6144,256,0.0035182221068276297
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,128,128,0.005208000126812193
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,5120,512,0.017841777867741056
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,12288,7168,0.11060622003343369
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,512,512,0.007273777491516537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,256,768,0.00556355549229516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3072,6144,0.02643999954064687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,256,3584,0.015150222513410779
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,128,1024,0.012295111185974546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1536,6144,0.009326222042242685
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,768,1024,0.013806222213639153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,12288,2560,0.02386755579047733
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,768,3072,0.00794666674402025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1536,7168,0.01828977796766493
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,12288,1024,0.02108088963561588
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,8192,4096,0.12218488587273492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,768,512,0.006083555519580841
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,128,2048,0.005026666654480828
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,7168,7168,0.014004444082578024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,128,128,0.014179555906189812
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,64,512,0.006614222294754452
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,12288,8192,0.023200000325838726
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,3584,2560,0.01771111124091678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,32,12288,0.014183999763594734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,12288,3584,0.08477777904934353
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,6144,2048,0.020256888535287645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,128,8192,0.010762666662534079
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,6144,4096,0.2401804394192166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,5120,2560,0.024043555061022442
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,8192,12288,0.04963199959860908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,64,6144,0.03052622079849243
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,8192,512,0.019858666592174105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,4096,6144,0.02347466680738661
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1024,8192,0.01056977775361803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,2560,65536,0.07751288678910997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,6144,3584,0.03385688861211141
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,3584,16384,0.04387288954522875
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,5120,12288,0.02126666737927331
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,10240,8192,0.08205155531565349
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,1024,1536,0.04214577873547872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3584,2560,0.025157334076033697
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,12288,512,0.015475556254386902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,10240,1024,0.02161600026819441
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,1024,512,0.011723555624485016
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3584,3072,0.007252444823582967
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2048,16384,0.01978222197956509
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,12288,12288,0.4193582269880507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,10240,256,0.011025778121418424
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,10240,5120,0.04739377895991007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,12288,2560,0.01482222146458096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,768,16384,0.012226666841242047
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,10240,256,0.0035822221802340615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,16384,16384,0.0749102234840393
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,128,128,0.005327111317051782
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,256,32,0.0018800000349680583
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,768,5120,0.00795733349190818
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,65536,2560,0.03228711088498434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2048,64,0.0029164445069101122
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2560,128,0.0028853333658642242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,32,65536,0.04555022385385302
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,1536,7168,0.01889155473974016
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,256,3584,0.007672000262472365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,128,1536,0.0035253332720862497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,5120,64,0.0018479999982648427
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,8192,64,0.005581333405441708
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,768,128,0.014097778333557976
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,65536,4096,0.11530933115217422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,128,8192,0.006976000136799282
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,8192,128,0.008812444077597724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,12288,128,0.025343111819691126
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,5120,256,0.016554666890038382
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,12288,6144,0.021776000658671062
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,32,6144,0.009347555538018545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,6144,256,0.01683999929163191
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2048,6144,0.010294222169452244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,65536,3072,0.033372445238961115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,10240,7168,0.03901244534386529
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,10240,5120,0.01662133302953508
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,2048,1536,0.005903111149867375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,512,16384,0.015815999772813585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,10240,32,0.03206400076548258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,65536,8192,0.19416622320810953
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2560,1024,0.003232888877391815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,16384,5120,0.02680977847841051
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,7168,4096,0.02701777716477712
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,32,2048,0.006601777755551868
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,7168,8192,0.01571999986966451
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,65536,3584,0.07678311400943331
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,12288,256,0.01754399968518151
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,2560,3072,0.007661333514584436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,128,2560,0.004935111022657818
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2048,8192,0.04361244373851352
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,3584,1024,0.005919999960396025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,65536,256,0.3287466631995307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,256,128,0.012460444536474017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1024,768,0.005960888746711943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2048,5120,0.008665777742862701
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,7168,5120,0.01349688900841607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1024,2560,0.004939555707905027
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,12288,64,0.036883556180530123
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,2560,256,0.005224888937340842
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,6144,6144,0.011852444046073489
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,1024,7168,0.01904177831278907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,16384,8192,0.1361804405848185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,12288,65536,0.27108534177144367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,7168,65536,0.16428977913326687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,65536,128,0.02625777820746104
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,128,128,0.014479110638300577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3072,65536,0.06631022029452853
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,4096,2048,0.045999109745025635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2560,16384,0.019301333361201815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,10240,2048,0.012070222033394707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,1024,5120,0.018274666534529794
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3072,128,0.005628444254398346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,5120,3072,0.0075137780772315125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,10240,12288,0.0329786671532525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,128,12288,0.00565155554148886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3072,1536,0.008271111382378472
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,256,1024,0.0028453332682450614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,12288,1536,0.014839111102951897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,32,256,0.0021760000122918025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,10240,16384,0.04733599887953865
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,512,12288,0.010199999643696679
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,512,3072,0.01979022224744161
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,8192,2560,0.007596444752481248
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,4096,512,0.0380364457766215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1536,16384,0.012112889024946423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,7168,4096,0.012875555290116204
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,5120,768,0.006279110908508301
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3072,3584,0.012637333737479316
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,1024,4096,0.01754577789041731
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,4096,1024,0.006255110932721033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,12288,6144,0.04227644536230299
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,768,1024,0.013432000246312885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,8192,2560,0.010031999813185798
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,512,1024,0.0031973332580592898
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,65536,4096,0.050534221861097545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,2560,128,0.017774222625626456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,512,128,0.012783110969596438
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,7168,256,0.02014222244421641
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,768,512,0.002554666664865282
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,8192,128,0.01519733336236742
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,16384,65536,0.33092978265550405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,6144,16384,0.024313777685165405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,2560,10240,0.03652177916632758
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,16384,6144,0.04577689038382637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1536,32,0.005976888868543837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,256,512,0.011772444678677453
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1024,2048,0.00630044440428416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,8192,512,0.003167111012670729
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,32,512,0.008468444148699442
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,256,10240,0.012576888832781049
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,8192,512,0.01790488925245073
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,4096,3584,0.0075777777367168
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,12288,256,0.003403555601835251
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,32,512,0.005607111172543631
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2560,10240,0.01202488856183158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3072,3072,0.009022222624884712
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,10240,4096,0.04502044452561272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2560,16384,0.05340888765123156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,512,64,0.0059573331640826324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,5120,12288,0.048872000641293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,16384,2048,0.014936000108718872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,3584,256,0.015187554889255099
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,3072,768,0.0063235556913746735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,2048,256,0.005899555567238066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,4096,768,0.006246222390068903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,10240,5120,0.029545777373843726
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,16384,768,0.005956444475385878
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,5120,65536,0.10063378016153972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,10240,768,0.022879999544885423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,65536,12288,0.25907111167907715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2560,2560,0.009831111464235518
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,4096,12288,0.021204445097181533
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,512,512,0.0028355556229750314
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,512,3072,0.005212444398138259
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,768,7168,0.012712889247470431
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,12288,3072,0.020869332883093093
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,768,2048,0.01721155477894677
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,16384,8192,0.06929955879847209
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3584,1024,0.036751111348470054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,12288,6144,0.021243555678261652
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2048,2048,0.006568888823191325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,6144,2560,0.007908444437715743
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,5120,3584,0.04794666502210829
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,32,1536,0.003528888854715559
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3584,6144,0.012059555285506778
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,8192,65536,0.20997867319318983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,12288,3072,0.014688000082969666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,256,768,0.02699733277161916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,512,3584,0.008034666379292807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,2560,12288,0.01761866609255473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,256,10240,0.020638222495714825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,128,128,0.0053324442770746015
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,65536,256,0.013380444712109037
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,2560,2560,0.006907555378145642
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,1024,2048,0.018217777212460835
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,128,10240,0.1546257734298706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,4096,5120,0.01036266651418474
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,6144,65536,0.145468446943495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,16384,7168,0.021828444467650518
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,512,512,0.00564533347884814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2048,2560,0.0069511110583941145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,16384,768,0.007354666789372762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,256,3072,0.005624889085690181
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1024,16384,0.012108444339699216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,768,1024,0.00775022225247489
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,1024,768,0.008340444829728868
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,16384,2560,0.09766666756735908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,3072,8192,0.011075555450386472
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,8192,8192,0.019700444406933255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,256,1024,0.0035457776652442086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,4096,2048,0.00702311098575592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,5120,8192,0.030417776770061914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,8192,7168,0.014599111345079211
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,2048,256,0.014871110518773397
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,768,4096,0.016215110818545025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2048,64,0.0075902218619982404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1024,3072,0.006056889063782162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3584,16384,0.5069155693054199
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,16384,8192,0.03544888893763224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,65536,4096,0.841992007361518
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,1536,4096,0.007696888513035244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,12288,4096,0.03641600079006619
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,16384,2048,0.01465777721669939
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3072,12288,0.0199626667631997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,4096,2048,0.011145778000354767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,1024,512,0.005523555394675996
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,32,64,0.005566222386227713
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,5120,1536,0.07385066482755873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,128,512,0.010778666370444827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,256,6144,0.009682666924264696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,16384,3584,0.057378669579823814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,2560,3584,0.016484444340070088
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,8192,10240,0.13295643859439427
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,16384,1024,0.011750222080283694
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,3584,8192,0.04121333360671997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,768,6144,0.009744889206356471
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,128,7168,0.013190222283204397
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,2048,12288,0.016206221448050607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2560,5120,0.009019555317031013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,2560,4096,0.028608888387680054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,8192,1536,0.1170275608698527
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,768,5120,0.010417777631017896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1024,8192,0.013049778011110095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,768,32,0.008405333591832055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2048,2048,0.004563555535342959
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,3584,512,0.005675555517276128
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,768,5120,0.008980444735950893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1536,7168,0.018958222534921434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,1536,3584,0.015923556354310777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3584,64,0.005668444352017508
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,256,1536,0.016364443633291457
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,10240,12288,0.374908447265625
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,64,8192,0.004952888935804367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1536,1536,0.00361333332127995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,6144,1536,0.007677333222495184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,65536,6144,0.0891173349486457
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,1536,6144,0.008745777938101027
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,5120,8192,0.016169778174824186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,32,6144,0.013557333085272046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,10240,1024,0.007355555891990662
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1536,8192,0.011406222151385413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,512,32,0.0018568889548381169
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,16384,64,0.026141333911154006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,8192,65536,0.18349511093563506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,4096,2048,0.006719111154476802
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1536,2048,0.009369778136412302
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,16384,2560,0.012515555653307172
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,3584,16384,0.10638311174180771
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,128,2048,0.004451555510361989
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,4096,1024,0.005980444451173146
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,5120,8192,0.4011857774522569
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,256,512,0.006191111273235745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3584,8192,0.012716444830099741
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,8192,2048,0.24004178576999238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,10240,1024,0.007429333196745978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3584,1024,0.0063324446479479475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,65536,10240,0.19646222061581084
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,768,2048,0.026299556096394856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,65536,2048,0.049266666173934937
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,128,256,0.011728888584507836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2560,32,0.002537777854336633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,6144,1024,0.009754666851626502
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2048,768,0.005653333332803514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,4096,128,0.00254488881263468
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3072,7168,0.012895999683274163
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,768,256,0.01531555586391025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2560,65536,2.9925804138183594
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,32,256,0.016593777471118502
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,32,10240,0.008792888787057664
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,4096,6144,0.021340444684028625
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,3584,7168,0.012847999731699625
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,10240,32,0.006001777946949005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,2560,10240,0.019090667366981506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,16384,7168,0.05935821930567423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,65536,10240,0.0887884431415134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,4096,64,0.005583111196756363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,5120,16384,0.02365422248840332
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,3072,512,0.0052373334765434265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,16384,64,0.0021760000122918025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3072,12288,0.015837333268589444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,6144,4096,0.0880497760242886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,6144,16384,0.04804889029926724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,768,5120,0.03643111056751675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,12288,768,0.01926488843229082
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,65536,16384,9.47344716389974
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,5120,64,0.011047999891969891
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,8192,3584,0.030080888006422255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,4096,768,0.004907555464241239
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,768,2048,0.009523555636405945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,8192,16384,0.026985777748955622
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,1536,256,0.016905777984195285
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,65536,12288,6.502656724717881
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,1536,8192,0.010784000158309937
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,256,512,0.01202488856183158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,12288,65536,6.714979383680556
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,128,5120,0.011403555671374003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,128,512,0.006295111030340195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,1024,4096,0.008608889248636033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,7168,65536,0.12436000506083171
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,65536,16384,0.5952631102667915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1536,512,0.004597333156400257
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,128,6144,0.03421955638461643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,12288,1024,0.007320000065697565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,7168,6144,0.028361777464548748
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,5120,3584,0.008715555899673039
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,12288,5120,0.04342133469051785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,12288,6144,0.12642578283945718
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,7168,2560,0.00815822184085846
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,10240,4096,0.017877333694034152
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,65536,1024,0.02628177735540602
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,3584,128,0.005626666463083691
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,6144,2048,0.0075902218619982404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,12288,2560,0.19366044468349883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,256,4096,0.007651555869314406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1024,1536,0.004252444538805219
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3072,65536,0.06304088566038343
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,4096,1024,0.006739555547634761
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,5120,2048,0.006607111129495833
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,4096,5120,0.008794666992293464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2048,7168,0.01033155537313885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,32,4096,0.008371555142932469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,768,6144,0.009688888986905416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,10240,3584,0.029285334878497656
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,5120,12288,0.1658488909403483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2560,65536,0.06469688812891643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,512,10240,0.021327111456129286
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,256,256,0.012123555772834353
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,16384,6144,0.03305688831541274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,1024,128,0.015309333801269531
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,1536,3584,0.021710221966107685
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,12288,10240,0.10094133350584243
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,16384,2048,0.01421955558988783
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,10240,16384,0.042188445727030434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,1536,65536,0.06295555830001831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,16384,6144,0.04965155654483371
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1024,128,0.002177777803606457
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,5120,3072,0.008214222060309516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,65536,8192,0.1754115555021498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,128,4096,0.015156444576051501
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,12288,2560,0.03223289052645365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,512,512,0.006677333265542984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,5120,10240,0.01518133282661438
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,512,5120,0.008356444537639618
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3584,2560,0.007278222176763747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2048,10240,0.0169840008020401
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,16384,128,0.04005333450105455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3584,64,0.002179555594921112
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,64,512,0.002543111021320025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1536,5120,0.007686222592989604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,10240,12288,0.3603297869364421
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,6144,512,0.0031715554909573663
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,10240,8192,0.03158488869667053
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,32,2048,0.004544888933499654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,5120,4096,0.008352888955010308
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,256,1024,0.006292444550328785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,256,2048,0.010597333312034607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,64,5120,0.012849777936935425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,10240,2048,0.01921422282854716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,3584,2048,0.00629333323902554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,2560,7168,0.015455999308162265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,256,3072,0.007290666302045186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,256,5120,0.016234666109085083
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1536,128,0.0022239999638663395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,3584,128,0.014570666684044732
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,2048,3072,0.009691555466916826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,1024,65536,0.11117688814798991
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1536,5120,0.02397866712676154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,768,3584,0.007679111427730984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,16384,768,0.019976889093716938
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2048,3072,0.01981511049800449
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,16384,65536,0.8252479765150281
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,1536,1024,0.005970666805903117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,8192,3584,0.014006222287813822
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,16384,6144,0.15005244149102107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2560,65536,0.08114577664269342
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,6144,64,0.002191999927163124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,2048,768,0.013952000273598565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,2048,768,0.005656888915432825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,512,7168,0.018882667024930317
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,1024,3584,0.031501332918802895
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,256,1536,0.010420444111029306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,6144,256,0.005625777774386936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2048,3584,0.0074924445814556545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,1024,12288,0.025010666913456384
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,4096,2560,0.008040889269775814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3584,256,0.0032044444233179092
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2560,8192,0.08827377690209283
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,10240,2048,0.04088000125355191
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,256,12288,0.10354755322138469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1536,1536,0.015827554795477126
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,768,8192,0.010453333457310995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,768,10240,0.008338666624493068
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,7168,1536,0.006406222366624408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,3584,12288,0.017255110873116385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,5120,65536,0.09932177596622044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,10240,1024,0.005954666684071223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,6144,2048,0.006952889263629913
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,10240,768,0.008654221892356873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3584,3072,0.009778666827413771
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,1536,65536,0.06321689155366686
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,64,4096,0.004830222162935468
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,16384,3584,0.01605599953068627
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,8192,10240,0.01828266680240631
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,8192,5120,0.03670400049951341
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,10240,256,0.006151111175616582
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,7168,512,0.005665777872006099
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,1024,8192,0.010806221928861407
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,32,768,0.005962666538026597
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,65536,12288,0.5323893229166666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1536,64,0.00592533333433999
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,1536,256,0.015164444843928019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1536,7168,0.008810666700204214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,256,10240,0.02304800020323859
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,8192,256,0.034296890099843345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,10240,2560,0.011417778001891242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,65536,256,0.03458399905098809
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,512,512,0.005947555518812603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,12288,3584,0.03566844595803155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,65536,7168,0.1309973398844401
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,4096,3072,0.007506666911972894
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,256,6144,0.006712889091836081
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3584,64,0.0027137777457634607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,10240,8192,0.05298755566279093
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,2048,65536,0.06417777803209093
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,3584,8192,0.013502221968438892
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,7168,3584,0.029748443100187514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,12288,2560,0.023366222778956097
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,256,2560,0.013893332746293811
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,7168,8192,0.0491466654671563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,12288,16384,0.04044622182846069
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,3072,512,0.005588444570700328
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,256,1024,0.005932444499598608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,8192,128,0.022665777140193518
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,128,5120,0.008352888955010308
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,5120,1536,0.015301333533393012
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2560,7168,0.013405332962671915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,6144,128,0.00556355549229516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3584,2048,0.007000000112586551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2048,16384,0.020050666398472257
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2048,1536,0.004226666771703296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,32,1536,0.0035751110149754416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,8192,3072,0.023797333240509033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1024,7168,0.011742221812407175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3584,768,0.006322666588756774
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3072,512,0.005982222242487802
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,8192,6144,0.017465778523021273
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,5120,512,0.0029111111329661477
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,6144,2048,0.03421155611673991
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,1024,768,0.016302223006884258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,4096,256,0.005921777751710679
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,6144,65536,0.11685866779751247
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,512,3072,0.005669333454635408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,2048,4096,0.008035555481910706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,12288,4096,0.03746133380466037
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,7168,5120,0.027087999714745417
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,16384,768,0.02001066671477424
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,256,5120,0.015947555502255756
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,4096,2560,0.00720266666677263
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,3072,128,0.0053013331360287136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,8192,7168,0.015063999427689446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,16384,2048,0.015042665931913586
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,512,2560,0.05330488748020596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,16384,768,0.010064000056849586
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,16384,12288,0.06975911060969035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,5120,128,0.006305777778228124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,2560,3584,0.00869333330127928
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,2048,7168,0.036538667149013944
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,5120,16384,0.16869955592685273
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,2560,1536,0.0069395556218094295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,10240,2048,0.02643999954064687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,1024,7168,0.04919999837875366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,512,3584,0.007647111184067196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3072,4096,0.018410666121376883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,128,16384,0.016143111719025504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,1536,2048,0.006642666541867786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,12288,3072,0.012064889073371887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,3072,65536,0.06540355417463514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,10240,16384,0.03942311141226027
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,1536,512,0.006166222194830577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,2048,4096,0.007663999994595845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,4096,256,0.016576000385814242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,8192,2048,0.00903555585278405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,3584,16384,0.021320000290870667
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,5120,12288,0.04424711068471273
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,65536,32,0.16720888349745008
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,6144,8192,0.016142222616407607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,2560,16384,0.09056711196899414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,4096,1024,0.03357777661747403
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1536,64,0.00816533300611708
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2560,2048,0.004903111192915174
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1536,12288,0.1815342240863376
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3072,4096,0.010222222242090438
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,5120,16384,0.3100533220503065
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,6144,128,0.005836444596449534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1536,64,0.0021057778762446507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3072,256,0.005305777821275923
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1024,7168,0.010016889207892949
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,128,7168,0.010658666491508484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,512,3584,0.011016888750924004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1024,32,0.002442666639884313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,10240,16384,0.05553422371546427
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2560,7168,0.012796444197495779
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,32,1024,0.006331555545330048
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,6144,128,0.005853333406978183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,3072,256,0.015202666322390238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,8192,512,0.00740533322095871
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,3072,2560,0.01979200045267741
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2560,7168,0.013832000394662222
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3072,3072,0.00738666703303655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,512,128,0.012864888542228274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,16384,16384,0.16808088620503744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,6144,5120,0.03289244572321574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,6144,7168,0.013239999612172445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,4096,32,0.0038746665749284956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,12288,12288,0.0532497763633728
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,12288,10240,0.04323110977808634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2048,1024,0.004256000121434529
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,512,256,0.015586665934986539
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,512,768,0.015526221858130561
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,7168,65536,4.202805413140191
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,8192,2048,0.034684442811542086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1536,512,0.003909333298603694
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,768,2560,0.01662488944000668
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,5120,7168,0.013743110828929476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1536,8192,0.01423733267519209
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,6144,128,0.00573511090543535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,256,768,0.002871111035346985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,65536,5120,0.09247733487023248
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,7168,4096,0.09748978084988064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,4096,6144,0.023452444208992854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,768,768,0.005640000104904175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2560,8192,0.014934221903483072
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,10240,7168,0.03351199958059523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2048,65536,0.04484000139766269
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,6144,4096,0.03264088763131036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2560,12288,0.01867466668287913
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,64,512,0.00254844439526399
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,128,512,0.005601777798599667
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,4096,5120,0.010214221974213919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,64,10240,0.004574222283230888
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1536,4096,0.013745778136783175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,5120,12288,0.019865777757432725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,128,10240,0.02032800018787384
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,8192,5120,0.03990222348107232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,768,5120,0.009748444788985783
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,1536,3072,0.046410666571723096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,10240,2048,0.027152889304690894
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,5120,6144,0.010856889188289642
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1536,768,0.003645333151022593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,128,7168,0.011108444796668159
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,128,1024,0.005971555494599872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,32,128,0.0022897778285874263
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3072,7168,0.0873839987648858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,128,128,0.005679111099905438
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,7168,2048,0.0243511117166943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,768,3072,0.009748444788985783
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,512,10240,0.056644446320003934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1024,6144,0.009786666267448002
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,512,10240,0.014742222097184924
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,16384,65536,0.18376355701022676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,1536,3072,0.016911110944218107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,7168,64,0.02167911165290409
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,128,64,0.0018444444156355327
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,3072,4096,0.020654223031467862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,5120,65536,0.0703679985470242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,16384,256,0.030234667989942763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,1536,3072,0.031048887305789526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,2560,512,0.005576000031497743
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,16384,12288,0.24405156241522896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1024,1536,0.006776888751321369
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,12288,10240,0.02756888833310869
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1536,2048,0.00591466658645206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,6144,5120,0.013304000099500021
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,128,12288,0.01347733371787601
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,12288,7168,0.5021288659837511
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,16384,32,0.004487110922733943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,16384,1536,0.010088889135254754
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,512,7168,0.007652444144090016
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,6144,6144,0.01442399952146742
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,768,8192,0.011412444214026133
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,128,768,0.012773333324326409
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2048,65536,0.05959910816616482
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,4096,10240,0.027178666657871667
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3072,8192,0.014391111003028022
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,768,3072,0.005571555760171678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,6144,512,0.016175111134847004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,8192,1024,0.04758488800790575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,10240,64,0.008664888640244802
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,64,1536,0.006254222244024277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3072,12288,0.019719999697473314
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2048,6144,0.011730666789743634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,12288,4096,0.031063109636306763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,1536,4096,0.007647111184067196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,65536,12288,0.12394044134351943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,768,3584,0.008007999923494127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,512,256,0.002584888910253843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,65536,128,0.01564711166752709
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,128,256,0.005311111195219888
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,4096,2560,0.006704000135262807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2048,5120,0.009123555488056606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,65536,512,0.01367377738157908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,512,3072,0.006999111009968652
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2560,16384,0.021853332718213398
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,1536,1536,0.01421866648726993
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,2048,10240,0.012459555433856117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,768,2048,0.006679111056857639
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2560,2560,0.007328888608349695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2048,32,0.0069653333889113525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,5120,16384,0.10797155565685695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,65536,1536,0.6165431340535482
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,10240,1024,0.011100444528791638
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,128,65536,0.017280000779363845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,3072,2560,0.05764266517427233
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,256,16384,0.01995022263791826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,16384,16384,0.07078755564159818
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1024,1024,0.005659555395444234
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,2560,16384,0.02033866610791948
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3584,12288,0.018377777602937486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,2560,65536,0.19968177212609184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,7168,512,0.005882666756709416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,10240,64,0.002201777779393726
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,256,1024,0.0063279999627007385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,2560,8192,0.040892445378833346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,5120,10240,0.03214222192764282
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,64,3072,0.007653333246707916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,32,2048,0.005330666485759947
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,65536,2560,0.07914755741755168
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,4096,16384,0.025024000141355727
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1024,3072,0.015808888607554965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,128,3584,0.007352888584136963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,7168,512,0.005991111199061076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,3584,5120,0.09040266937679714
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,768,1536,0.006302222195598815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,2048,12288,0.04500977860556709
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,8192,3584,0.02682400080892775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,4096,4096,0.0144213338692983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,7168,768,0.01833511061138577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,7168,128,0.005020444591840108
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,65536,3072,0.03088800112406413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,2560,6144,0.030582222673628066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,5120,8192,0.018280888597170513
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,128,1024,0.0032479998966058097
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,4096,65536,0.0766506658660041
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,2560,16384,0.020346666375796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,3072,256,0.005392888767851724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3584,12288,0.024318221542570326
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,3072,65536,0.09313866827223037
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,3584,768,0.005929777605666055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,8192,1536,0.009579555855857002
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,768,4096,0.007605333295133378
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,65536,12288,0.5236533482869467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2560,16384,0.20084177123175728
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,5120,128,0.0032524443748924467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,6144,2048,0.009431111315886179
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,8192,4096,0.03370399938689338
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,16384,5120,0.048376000589794584
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,7168,12288,0.031101332770453557
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,7168,16384,0.04980355501174927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,512,1536,0.003605333467324575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,65536,2560,0.04573066698180305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,256,7168,0.023056889573733013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,256,12288,0.0052382221652401825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,10240,3584,0.013820444544156393
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,128,4096,0.004303110970391167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,256,128,0.011775111158688864
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,1024,512,0.014356444279352823
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,256,12288,0.025063110722435847
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,10240,16384,0.06284355454974704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,4096,10240,0.016037333342764113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,256,65536,0.05173066589567396
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,10240,12288,0.04109422365824381
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,128,3072,0.007312888900438945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1536,1024,0.003875555677546395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2560,4096,0.008488888541857401
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,2560,2048,0.006727111008432176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,5120,12288,0.017303110824690927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,12288,16384,0.04537333382500542
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3584,2560,0.007639110916190677
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,8192,16384,0.12414844830830891
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2560,65536,0.06504266791873507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,1536,6144,0.027521777484152053
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,16384,10240,0.028375112348132666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,6144,4096,0.010390222072601318
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,768,10240,0.013026666310098437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,6144,256,0.020076443751653034
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,128,3584,0.007667555577225155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,2560,7168,0.009825777676370408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,512,7168,0.028576887316173975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2560,16384,0.020404444800482858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,6144,1024,0.01651733285850949
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3072,512,0.005927111125654644
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,5120,4096,0.030108445220523413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,10240,5120,0.018145778112941317
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,1536,2048,0.006287999865081575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,512,8192,0.005958222266700532
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,768,2048,0.015141333142916361
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,256,8192,0.010660444696744284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,512,256,0.005219555563396878
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,7168,6144,0.013247110777431063
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,2560,7168,0.011297778122954898
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,4096,3584,0.010411555568377176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,5120,256,0.022691556149058874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,256,768,0.009016888837019602
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,4096,3072,0.011611555185582904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,1024,128,0.00526577772365676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,12288,1024,0.01017599966790941
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,2560,2048,0.01959022217326694
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,7168,65536,2.02547476026747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,5120,128,0.022535110513369244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,6144,1024,0.008062221937709385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,7168,3584,0.035382221142450966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,10240,1536,0.02366133365366194
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,12288,1024,0.007140444384680853
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,1536,1024,0.005951111101441913
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2048,3072,0.03233066532346938
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3584,5120,0.015564444992277356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,3584,128,0.016159110599093966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,128,2048,0.00461422238085005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,512,3072,0.007307555940416124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,7168,6144,0.0428746673795912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,8192,5120,0.10483022530873616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,768,7168,0.009649777577983009
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,16384,7168,0.07705955372916327
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,6144,7168,0.01978222197956509
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,8192,65536,0.16587911711798772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,6144,7168,0.018206222189797294
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,4096,1536,0.006251555350091722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,128,5120,0.008311111066076491
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,1024,1536,0.014162666267818876
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,3072,7168,0.02330577704641554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,6144,7168,0.057687997817993164
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,3072,512,0.005239999956554837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,16384,768,0.06284888585408528
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2048,4096,0.013952000273598565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,2048,7168,0.06778666708204481
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,12288,7168,0.03580355644226074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,65536,8192,0.07719288931952582
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,1536,2048,0.0069955554273393415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,1024,3584,0.01688799924320645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,7168,16384,0.024865777956114873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,4096,1536,0.008589333130253686
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,12288,7168,0.059446222252315946
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,65536,7168,0.07765333520041572
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1536,1024,0.003167111012670729
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1024,6144,0.009343110852771336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3584,5120,0.011183111204041375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1024,1536,0.005225777626037598
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,768,2560,0.016141333513789706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,16384,6144,0.023413333627912734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,1024,128,0.014440000057220459
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,7168,256,0.017014222012625802
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,65536,10240,0.09043288893169826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,8192,512,0.002834666727317704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,65536,5120,0.17873688538869223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,16384,128,0.025039111574490864
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,7168,6144,0.03611022233963013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,128,512,0.00591466658645206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3584,12288,0.04757066567738851
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,768,3072,0.0155404441886478
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,3072,1024,0.014431110686726041
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2560,3584,0.007674666742483775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,65536,10240,0.35076000955369735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,8192,3584,0.09839733441670735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,4096,256,0.0025191110455327565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,64,128,0.008022222254011366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,12288,256,0.017270222306251526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,16384,3584,0.02916355596648322
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,3584,2560,0.022687999738587275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,4096,12288,0.01956533392270406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,7168,3584,0.012132444315486483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,10240,768,0.017674666312005784
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,5120,3584,0.023112888137499493
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,12288,12288,0.05568799707624647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,8192,3584,0.03318399853176541
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,4096,768,0.005580444302823808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,12288,128,0.05180177754826016
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1024,6144,0.009704888694816166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,16384,12288,0.04126844472355313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3584,3072,0.010454222559928894
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,128,7168,0.0107351111041175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,128,10240,0.010965333216720156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,8192,128,0.005663110978073544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,16384,3584,0.02000711030430264
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,10240,10240,0.025898666845427618
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3072,64,0.006056889063782162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,7168,16384,1.6922133763631184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,16384,1536,0.017438222964604694
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,64,65536,0.20661511686113146
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,512,2560,0.014660444524553088
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2560,10240,0.01920088960064782
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,3584,12288,0.08255910873413086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2560,6144,0.08091466956668429
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,3584,10240,0.04615199897024366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,5120,65536,0.07569422324498494
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,8192,10240,0.027218666341569688
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,768,16384,0.015047111445003085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,2048,1024,0.005950222412745158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,12288,2048,0.021216000119845074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,512,6144,0.00979644474056032
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,512,65536,0.06172000037299263
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,6144,3584,0.025080889463424683
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1024,3072,0.007316444483068254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,5120,6144,0.013445333474212222
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,768,7168,0.01198933356338077
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,10240,64,0.010074666804737514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,16384,12288,0.03471111257870992
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,7168,7168,0.014495111174053617
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,8192,1024,0.020101333657900494
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1024,32,0.003908444609906939
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,12288,8192,0.09364355272716945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,65536,6144,0.19879644446902803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,6144,768,0.015439111325475903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,10240,2048,0.02792444494035509
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,64,65536,0.27658843994140625
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,1024,768,0.015830222103330824
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2560,32,0.005664000080691443
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2048,1536,0.006598222172922558
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,1024,7168,0.02015999952952067
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,8192,4096,0.013085333009560903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,768,128,0.002549333290921317
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,7168,8192,0.02225422196918064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,5120,65536,0.07191289133495755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,2560,5120,0.008338666624493068
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,2048,16384,0.038176000118255615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2048,5120,0.014476444986131458
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,1536,5120,0.013800000150998434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,10240,16384,0.05395199855168661
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,768,3072,0.007607111500369177
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,128,10240,0.005986666513813867
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,512,2048,0.004600888739029567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,16384,3072,0.013916444447305469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,1536,3072,0.030211554633246526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,4096,3072,0.024497777223587036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,32,65536,0.022242666946517095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,10240,65536,0.24151023228963217
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,7168,5120,0.02421422302722931
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,6144,8192,0.04890311095449659
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,4096,6144,0.04721866713629829
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,128,65536,0.06300710969501071
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,65536,3072,0.12076800399356419
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,1536,1024,0.013487999637921652
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,128,12288,0.03438755538728502
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,10240,768,0.01832177738348643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2560,512,0.002854222224818336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,128,8192,0.019259555472267997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2560,5120,0.016916443904240925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,6144,1024,0.016858667135238647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3072,768,0.006317333214812809
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,16384,10240,0.05804355276955498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,256,1024,0.005972444597217772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,512,4096,0.02604355580276913
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,4096,12288,0.04624355501598782
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,512,3584,0.007748444047239091
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,3584,10240,0.014510222607188754
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,4096,2048,0.04831644561555651
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,6144,8192,0.01687466601530711
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,1536,7168,0.010336000058386061
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,768,12288,0.025581333372328017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,8192,128,0.0028444443725877335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,256,65536,0.059506667984856494
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,3072,3072,0.012404444317022959
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,16384,768,0.005952888892756567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,8192,7168,0.031473779016070895
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3584,2560,0.006360000206364526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,256,512,0.0024826667375034755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1536,16384,0.12261244985792373
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,12288,5120,0.0318177772892846
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,7168,2048,0.02310488952530755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,16384,6144,0.02279022170437707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,65536,768,0.016753777861595154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,3584,1536,0.0062773335311147906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,16384,3584,0.33284176720513237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,5120,10240,0.03242311212751601
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,1536,65536,0.07654400004280938
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,7168,65536,0.08411644564734565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,4096,128,0.014953777194023132
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,2560,4096,0.02036711076895396
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,5120,4096,0.03230399886767069
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,4096,4096,0.010256888965765635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,10240,1024,0.005594666633341048
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,2560,8192,0.01499555508295695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,5120,32,0.005943999936183293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,768,1536,0.01479644411140018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1536,3584,0.005991111199061076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,1024,3072,0.015547555353906421
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,3584,12288,0.017024000485738117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2560,7168,0.01216533366176817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,3072,10240,0.03594133257865906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,65536,256,0.04328355524275038
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,3072,4096,0.02455733385350969
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,512,12288,0.006642666541867786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1024,128,0.00591111100382275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,7168,32,0.006096000058783426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,5120,256,0.0052355556852287715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,12288,3072,0.23730044894748262
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,12288,6144,0.02929511004024082
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,12288,4096,0.020269332660569083
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,512,65536,0.06396889024310641
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,512,128,0.01311644415060679
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1024,5120,0.008929777476522658
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,10240,4096,0.032617777585983276
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,128,512,0.007283555964628856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,256,256,0.005619555711746216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,2048,128,0.023383110761642456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,6144,64,0.002161777681774563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,3072,1536,0.02086488902568817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,16384,32,0.0069955554273393415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,8192,1024,0.006228444476922353
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,512,16384,0.05030044582155016
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,512,6144,0.03555466731389364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,16384,6144,0.3027422163221571
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,8192,65536,0.2440426614549425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,6144,12288,0.022444445225927565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,64,256,0.0027866665687825945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,2048,2560,0.030870222383075293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3584,1024,0.005979555762476391
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,4096,65536,0.0844142238299052
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,128,5120,0.049044443501366504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,7168,128,0.005352889084153705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,3072,65536,0.08666044473648071
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,256,1536,0.003597333199448056
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,512,12288,0.0196426659822464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,10240,128,0.006194666855865055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,65536,3072,0.15744622548421225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,5120,65536,0.07304977708392672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,2560,16384,0.0480275551478068
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,128,3072,0.0073626670572492815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,1536,1024,0.014762666490342883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,768,128,0.00555644432703654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,256,3072,0.009141333401203156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,128,1536,0.01238933371173011
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,5120,1024,0.0041653331783082755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,5120,16384,0.023762666516833838
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,7168,1536,0.017493334081437852
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,32,8192,0.010683555569913654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,512,6144,0.03644533289803399
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,16384,3584,0.020807110601001315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,4096,3584,0.007594666547245449
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,3072,7168,0.08649955855475532
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,5120,768,0.016003555721706815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,10240,12288,0.20215466287400988
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,8192,1024,0.018230222993426852
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1024,6144,0.0069644442862934535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,65536,512,0.3719715542263455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,256,6144,0.009949333137936061
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,16384,6144,0.052483555343416005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,128,512,0.009355555805895064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,10240,2560,0.02465155555142297
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,512,1536,0.012011555333932241
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,65536,3584,0.038206222984525896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,2048,512,0.005647111270162795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1536,128,0.002277333289384842
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,16384,7168,0.07192266649670072
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,65536,768,0.015490666031837463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,8192,7168,0.03639288743336996
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,65536,16384,0.14261066913604736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,256,12288,0.02274133265018463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,12288,8192,0.03815022110939026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,256,1024,0.012457778056462606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,512,8192,0.017550221747822232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,12288,8192,0.05543555484877693
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,5120,512,0.006423111177153057
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,128,128,0.0022151112142536375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,768,1536,0.006308444258239534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1024,4096,0.008750222623348236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,4096,2560,0.007487110793590546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3072,10240,0.015993777248594496
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,2560,8192,0.023452444208992854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,10240,5120,0.017814222309324477
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1024,3584,0.03975733452373081
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,65536,128,0.062310225433773465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,5120,8192,0.016183111402723525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,128,7168,0.018816888332366943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,5120,768,0.003168888803985384
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,7168,6144,0.020812445216708712
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,7168,1024,0.004827555682924059
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,65536,1024,0.024706666668256123
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,4096,12288,0.018523555662896898
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,65536,3072,0.03666755557060242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,4096,3584,0.0731946693526374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,1024,6144,0.018256889449225534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,768,3584,0.027815110153622095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,256,65536,0.07011466556125216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,32,16384,0.017014222012625802
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,4096,4096,0.009313777916961247
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1536,16384,0.014559111661381192
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,2048,12288,0.01576622161600325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,2560,2560,0.021372444099850122
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2048,512,0.005523555394675996
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,512,6144,0.0203493336836497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,7168,1024,0.0062773335311147906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3072,3072,0.007272000114123027
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,16384,64,0.047092444366878934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,6144,12288,0.02117777698569828
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,2560,2560,0.08002133501900567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,12288,4096,0.01735822194152408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,6144,16384,0.024887111451890733
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,8192,768,0.0653546651204427
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,16384,3072,0.017112889223628573
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,8192,12288,0.02548177705870734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,512,8192,0.010760889285140567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,512,3584,0.007979555262459649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,6144,3072,0.02553244431813558
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2560,2560,0.006863111009200414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,512,3584,0.008623110751310984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,65536,2048,0.07367466555701362
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,64,768,0.005952000204059813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2048,32,0.005640000104904175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,2048,16384,0.03250311149491204
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,4096,1536,0.0061928890645504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,5120,12288,0.03930222325854831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,6144,16384,0.024175999893082514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1024,1536,0.007019555403126611
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,256,1024,0.009324444664849175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,10240,7168,0.11234488752153184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,32,16384,0.024716445141368445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,512,5120,0.005584888988071018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3072,256,0.0024968888610601425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2048,3072,0.005960000058015187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3072,64,0.0036799998746977914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,65536,10240,0.27143287658691406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3072,12288,0.04874666531880697
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,8192,128,0.015167999598715039
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,5120,10240,0.19877866903940836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,6144,32,0.007682667010360294
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,768,8192,0.021672000487645466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1024,256,0.00254844439526399
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,7168,65536,0.2103466722700331
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,10240,4096,0.022865777214368183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,768,10240,0.016474665866957772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,7168,8192,0.015503111812803479
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,512,256,0.006990222467316522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,12288,768,0.007153777612580194
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,4096,768,0.008486222061845992
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,2048,1536,0.0061048890153567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,6144,256,0.0025600000388092464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,5120,4096,0.1915360026889377
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,32,12288,0.028424001402325098
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,256,8192,0.0081377774477005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,512,6144,0.017210667332013447
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,1536,768,0.005945777727497949
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,3584,3072,0.00693866651919153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,128,65536,0.16958844661712646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,12288,65536,0.34820622868008083
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,2560,16384,0.0315688881609175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,768,768,0.007272000114123027
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,768,3584,0.015800888339678448
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,7168,1024,0.0039057777159743835
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,1536,10240,0.024007999234729346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,1536,5120,0.11499199602339004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,128,2048,0.0069466663731469055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,12288,7168,0.02312711046801673
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3072,768,0.003915555361244413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,16384,128,0.0023253332409593794
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,768,256,0.00555644432703654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,8192,10240,0.035515556732813515
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,1024,65536,0.08213866419262357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1536,256,0.0029075555503368378
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,65536,1024,0.06505511204401652
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,1024,7168,0.01959288948112064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,1536,2048,0.0069120000633928515
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,32,256,0.002171555534005165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,768,1536,0.014847111370828418
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,4096,128,0.014779556128713818
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,4096,1024,0.007106666763623555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,7168,8192,0.033260444800059
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,256,12288,0.18852533234490287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,64,128,0.005608888963858287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2048,32,0.005632888939645555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,8192,2048,0.008581333690219456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,10240,768,0.004416000097990036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,768,4096,0.032256000571780734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,12288,128,0.006308444258239534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,65536,5120,0.0484808882077535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,7168,7168,0.03129244512981839
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,65536,6144,0.25278133816189235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,8192,256,0.04680000080002678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,5120,2048,0.008976000050703684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,3584,512,0.014505777094099255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,6144,2048,0.013767999907334646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,512,4096,0.015836444165971544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,7168,8192,0.016252444850073922
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,64,768,0.0031840000301599503
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,256,5120,0.006585777633719974
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1024,5120,0.009723555710580613
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,768,1536,0.004906666775544484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1024,1536,0.011731555892361535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,7168,1536,0.00655644428398874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,7168,10240,0.026406221919589575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,12288,3072,0.014937778313954672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,10240,6144,0.017830222845077515
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,5120,8192,0.01664799948533376
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1024,256,0.0027875554644399216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,10240,1536,0.0092622223827574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,16384,65536,0.30967821015252006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3584,1536,0.004411555412742826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,256,16384,0.056368887424468994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,3072,1024,0.04445155461629232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3584,10240,0.01716888944307963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,7168,1024,0.006284444282452266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,16384,12288,0.033088889386918806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,32,768,0.008907555705971187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,3584,128,0.005947555518812603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,65536,4096,0.6154915491739908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,256,65536,0.13829778300391302
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,10240,2048,0.009840889109505547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,16384,1536,0.010986666712496014
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,12288,768,0.050669332345326744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3072,1536,0.026684444811609056
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,5120,512,0.0158595558669832
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,256,1024,0.011801777614487542
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,10240,1536,0.024932444095611572
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,16384,3072,0.06025244130028618
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,12288,65536,0.13652977678510878
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,5120,1536,0.0344471103615231
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,512,6144,0.022686221533351477
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,256,65536,0.012813333008024426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,768,512,0.0032435556252797446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,16384,2048,0.03221155537499322
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,7168,2560,0.02288977801799774
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,5120,768,0.017512889371977914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,128,65536,0.01517688896920946
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3072,3584,0.01698933376206292
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,65536,64,0.003842666745185852
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,65536,512,0.014825777875052558
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,7168,7168,0.018657777044508193
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,7168,128,0.005655111124118169
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,10240,12288,0.029230223761664495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2560,1536,0.006783111227883234
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1024,2560,0.007284444239404466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,256,10240,0.022295110755496558
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,32,2048,0.004889777965015835
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,5120,4096,0.02742133372359806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,3072,1024,0.014763555592960782
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1024,768,0.0058959999846087555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,5120,2560,0.009400889277458191
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,65536,32,0.010341333018408881
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,3584,2560,0.016516443755891588
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,8192,12288,0.6573235723707411
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,12288,7168,0.036200000180138484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,16384,128,0.030340443054835003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,3072,16384,0.0660239987903171
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,256,1024,0.005896889087226655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,2048,1536,0.005976888868543837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,12288,256,0.006252444452709622
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2048,12288,0.017901332841979135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,16384,8192,0.046148445871141225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,1536,65536,0.0628382232454088
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,1024,512,0.013017777767446307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1024,512,0.005600000007285013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,128,1536,0.005922666854328579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,6144,7168,0.030159112479951646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3072,16384,0.19224266211191812
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,1536,512,0.005245333330498801
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,10240,7168,0.34687288602193195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,128,64,0.00869422240389718
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,7168,12288,0.04312444395489163
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,512,256,0.008995555341243744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,16384,5120,0.04669510987069872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,8192,5120,0.015419556034935845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,10240,5120,0.020966221888860066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,10240,16384,0.05046933227115207
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,256,3584,0.06259377797444661
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1536,10240,0.017129777206314933
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,768,16384,0.03126400046878391
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2048,3584,0.03936355643802219
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,10240,64,0.006259555617968242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,6144,1536,0.011039111349317761
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2560,5120,0.010646222366227044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3584,16384,0.044216887818442456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,5120,4096,0.01072355525361167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,7168,16384,0.08340444167455037
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,768,128,0.002165333264403873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,512,1024,0.013073777986897362
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,64,256,0.002141333288616604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,8192,3072,0.012749333348539142
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2048,768,0.01482311056719886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,5120,256,0.016928889685206942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,16384,16384,0.0846853322452969
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,8192,5120,0.018583110637134977
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,12288,10240,0.11352177460988362
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,1024,65536,0.06259466542137994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,64,768,0.00888177752494812
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,128,64,0.005605333381228977
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,7168,256,0.01684888866212633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,128,65536,0.009855999714798397
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,65536,5120,0.12174399693806966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3584,1536,0.0069653333889113525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,8192,128,0.015209777487648858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,32,65536,0.10304888751771714
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,7168,10240,0.026633777552180823
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,8192,512,0.006333333336644703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,3584,5120,0.018743111027611625
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2048,64,0.005628444254398346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,256,256,0.0021928888228204516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1536,2560,0.04289866818322075
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2048,8192,0.013087111214796701
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,10240,1536,0.04569688770506117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,6144,3072,0.029677334758970473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2048,6144,0.013213333156373767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,256,5120,0.016428444120619033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,10240,65536,0.2958479987250434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,3072,5120,0.030396444929970637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3072,512,0.005935110979610019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,8192,8192,0.08577155404620701
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,128,3584,0.011946666571829053
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1024,2048,0.006724444528420766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,8192,1024,0.006295111030340195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,3072,256,0.029161777761247423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,32,256,0.0022293333378103045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,5120,128,0.008088000118732452
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,2048,768,0.005616888817813661
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,10240,1536,0.0255022214518653
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,5120,32,0.0023457778410779103
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,65536,6144,0.06778933604558308
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,5120,1024,0.003608889049953885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,16384,8192,0.05149866806136238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,128,6144,0.009511111511124505
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,128,3072,0.009760888914267222
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3072,8192,0.011879111329714457
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,10240,2048,0.010703999963071613
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,7168,768,0.006071110980378256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,3072,2048,0.0070186663005087115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,256,64,0.005616888817813661
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,4096,512,0.01481244464715322
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,512,5120,0.00906399968597624
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2560,12288,0.019090667366981506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,256,12288,0.023431110713216994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,8192,512,0.02306933369901445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,32,5120,0.013854222165213691
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3072,768,0.0034906665484110513
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2560,10240,0.016508445143699646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3584,512,0.006183111005359226
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,2560,7168,0.011518222590287527
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,256,512,0.002587555597225825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3072,5120,0.040624890062544085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,1536,12288,0.04875555634498596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,8192,256,0.007369777394665613
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,3072,8192,0.013092444174819522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2048,3072,0.007448888487286038
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,65536,2560,0.2656737698449029
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,4096,64,0.0059164443777667145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,65536,10240,0.2508462270100911
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,256,512,0.003117333269781537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,1024,4096,0.021599110629823472
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,3584,65536,0.06647111309899224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1024,5120,0.008712000317043727
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,256,2048,0.04235466652446323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,256,256,0.005933333188295364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,8192,512,0.00590133335855272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,256,8192,0.010861333045694562
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,512,65536,0.1767946614159478
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,65536,12288,0.10802844497892593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,768,6144,0.025435555312368605
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,128,12288,0.013087111214796701
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,512,65536,0.01808444493346744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,64,12288,0.013091555900043912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,12288,10240,0.04821777674886915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,6144,3584,0.011471110913488599
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3072,768,0.004780444420046276
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,5120,10240,0.01760444376203749
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,65536,6144,0.14175111717647976
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3584,512,0.005953777581453323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,16384,2560,0.014327999618318347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2560,64,0.005948444621430502
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,1536,2048,0.0063288890653186375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,3072,3072,0.007338666253619724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,5120,10240,0.016875555117925007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,512,2048,0.006710222197903528
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,256,10240,0.02106222179200914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,12288,6144,0.04219733344184028
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,5120,512,0.00600000015563435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,128,5120,0.004319111092223061
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,8192,3072,0.04793333344989353
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,16384,1024,0.00755466686354743
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,3072,5120,0.021270222134060327
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,128,8192,0.01109333336353302
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,128,1536,0.006720888945791457
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3072,2048,0.0070266665683852295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,7168,6144,0.01826755536927117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,7168,3584,0.028206222587161597
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,768,3072,0.015909334023793537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1536,65536,0.03534311056137085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,65536,4096,0.09622133440441555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,128,3072,0.007664889097213745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,512,5120,0.022503111097547743
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,4096,32,0.01519733336236742
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,5120,7168,0.014896000425020853
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,7168,512,0.005616888817813661
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,12288,128,0.021374222305085924
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,2560,128,0.00563466673096021
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,1024,6144,0.008662222160233391
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,512,7168,0.0236799998415841
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,7168,7168,0.016168889072206285
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2048,6144,0.011186666786670685
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,512,1536,0.006633777585294511
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2048,6144,0.014631110760900708
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,8192,2560,0.01039644413524204
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,5120,768,0.017483555608325534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1024,3584,0.013242666920026144
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,8192,4096,0.01130488928821352
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,3072,1024,0.016951999730534024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,768,16384,0.029903110530641343
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,3584,4096,0.022877333892716303
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,128,64,0.005927111125654644
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,1024,8192,0.021913778450753953
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,10240,8192,0.02590222160021464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,6144,768,0.007342221836249034
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,10240,4096,0.029224889145957098
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,7168,32,0.00665155549844106
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,8192,3584,0.05565066470040215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3072,4096,0.03015022145377265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,12288,8192,0.254692448510064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,4096,2560,0.011065777805116443
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,6144,2560,0.009409777820110321
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,5120,6144,0.12860177622901067
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2048,10240,0.012807110945383707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,512,6144,0.007660444411966536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,6144,512,0.0075706665714581805
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3584,8192,0.03435555431577895
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,1024,3072,0.020969778299331665
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1024,32,0.0025457777082920074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,768,128,0.014454222387737699
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,768,16384,0.017312889297803242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3072,64,0.002855111120475663
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,16384,2560,0.05351110961702135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,12288,7168,0.030610667334662542
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1024,5120,0.01605866683853997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,3072,7168,0.02788177794880337
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,64,6144,0.012424888710180918
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,16384,65536,0.367098675833808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,65536,12288,0.24088001251220703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,4096,65536,0.0834346678521898
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,10240,10240,0.11031911108228896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2048,6144,0.018981332580248516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,1536,768,0.016520000166363187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,1024,3072,0.0073626670572492815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,7168,1536,0.055361777544021606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,12288,4096,0.034621334738201566
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,6144,3072,0.039988444911109075
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1024,3584,0.00629688882165485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,6144,7168,0.021556443638271753
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,3072,512,0.005607999861240387
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3072,4096,0.008983111215962304
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,128,10240,0.011815999945004782
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,3072,3584,0.007662222617202335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,512,5120,0.016947555873129103
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,1024,8192,0.015855999456511605
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1536,4096,0.007277333074145847
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,3584,3072,0.02791200081507365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,4096,3584,0.01720711092154185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,16384,16384,0.04637155599064297
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,2560,2048,0.016248888439602323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,768,1024,0.006292444550328785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2048,256,0.011008888483047485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,64,65536,0.014264000786675347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2048,1024,0.011482666763994427
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,8192,6144,0.15398489104376897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,12288,512,0.020044444335831534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,3584,768,0.016844444804721408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3072,3072,0.011049778097205691
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,128,768,0.008013333711359236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3584,2048,0.0358862214618259
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,10240,1536,0.030200001266267564
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,8192,2048,0.024350222614076402
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,65536,10240,0.20725599924723306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3584,128,0.0028391112056043413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1536,16384,0.03417866759830051
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2560,4096,0.05282844437493218
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,8192,2048,0.008348444269763099
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,8192,7168,0.015140444040298462
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,128,10240,0.02390044430891673
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,4096,768,0.003990222182538775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,512,65536,0.05632266733381483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,4096,6144,0.01163822246922387
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,8192,5120,0.012473777764373355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,128,2048,0.0124177775449223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,128,6144,0.008996444443861643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,65536,5120,0.10619378089904785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,4096,2560,0.007264888948864407
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,5120,2560,0.020375111036830477
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,6144,3584,0.012500444220172035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,32,6144,0.004877333425813251
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1024,64,0.002157333410448498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,2048,3584,0.01680266691578759
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,10240,8192,0.02551288902759552
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,256,256,0.007660444411966536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3072,3584,0.007688889073001013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,8192,16384,0.029151111841201782
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,3584,65536,0.08679288625717163
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2560,32,0.005894222193294101
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2048,768,0.006260444306664997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,6144,2560,0.011516444385051727
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,7168,10240,0.02792444494035509
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,768,512,0.013072888884279462
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,2048,768,0.017128888103697035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,5120,2560,0.007003555695215861
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,12288,1024,0.02299733294381036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3072,64,0.0021786666992637846
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,5120,7168,0.2697670989566379
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2048,768,0.0028231110837724474
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3584,768,0.0031831111345026228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,6144,65536,0.7890426847669813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,6144,10240,0.030215998490651447
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,512,1536,0.016419554750124615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,2560,6144,0.008994667066468133
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,3584,128,0.014527999692493014
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,12288,128,0.015480889214409722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,6144,1536,0.006596444381607904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,64,10240,0.004562666846646203
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,3584,12288,0.053130666414896645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,65536,768,0.3460222350226508
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,128,12288,0.009724444813198512
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,16384,7168,0.6272995736863878
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,5120,2048,0.025068443682458665
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,8192,7168,0.05535377727614509
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,16384,6144,0.28822755813598633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,8192,1536,0.009303111169073317
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2560,5120,0.008728889127572378
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,5120,128,0.01887999971707662
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,7168,65536,0.1525084442562527
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,3072,12288,0.0292871097723643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,2560,16384,0.02161688937081231
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,8192,16384,0.02534399926662445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1536,12288,0.016509332590632968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,3072,1536,0.006258666515350342
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,32,768,0.002790222151411904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,10240,2048,0.13835644721984863
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,4096,2560,0.006980444822046492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,64,2560,0.004949333353175057
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,1536,7168,0.022664889693260193
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,1536,12288,0.02613777750068241
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,2560,7168,0.06075377596749199
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,12288,256,0.03845422135459052
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,6144,12288,0.022414222359657288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,4096,8192,0.014857777290874057
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,768,128,0.014479110638300577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,64,1536,0.006236444330877728
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,1536,65536,0.25266843371921116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,768,16384,0.015820443630218506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,128,32,0.007954667011896769
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,1024,512,0.005197333378924264
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,10240,7168,0.04010488920741611
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3072,5120,0.03577155537075467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,16384,768,0.007501333124107785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,65536,7168,0.0610026650958591
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,128,8192,0.019581332802772522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,128,3584,0.007985778152942657
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,7168,8192,0.3600320021311442
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3584,128,0.007652444144090016
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1536,64,0.0058542220956749385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,8192,10240,0.016664889123704698
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,4096,512,0.007272000114123027
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,7168,1536,0.006339555399285422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,8192,65536,0.3752613332536485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2048,2560,0.007071110937330458
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,6144,10240,0.10713422298431396
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,10240,6144,0.03251377741495768
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,2048,12288,0.03923199905289544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,32,2560,0.004927111168702443
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,65536,10240,0.16543022791544595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,2560,4096,0.06622222397062513
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,1536,1024,0.005988444305128521
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,6144,1024,0.01831022236082289
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,65536,2560,0.03148711058828566
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2560,64,0.0063368889192740125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,10240,2560,0.01442488862408532
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,8192,1536,0.006702222343948152
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,7168,2048,0.007724444071451823
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,7168,6144,0.04016977879736159
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,768,32,0.008002666963471307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,128,32,0.0018933332628673976
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2560,4096,0.028692444165547688
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,65536,2048,0.07422044542100695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3072,512,0.002604444407754474
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3584,64,0.0035822221802340615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2048,7168,0.010442666709423065
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1024,6144,0.00868888861603207
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,512,7168,0.009732444253232744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,16384,12288,0.15052889453040227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,2560,2048,0.006638222270541721
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,7168,10240,0.08010133107503255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,7168,65536,0.07854577567842272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,16384,32,0.00221422231859631
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,256,4096,0.008368000388145447
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,7168,12288,0.019653333557976615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,256,16384,0.016932444439993966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3072,512,0.005952888892756567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,16384,32,0.007690666450394525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,768,512,0.012823999755912356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,768,10240,0.012144889268610211
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,64,32,0.0019582222319311565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,1024,512,0.01312888910373052
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,1024,768,0.015344889627562629
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3584,7168,0.10028266906738281
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,512,768,0.015211555692884656
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3584,768,0.0052897776994440295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,4096,3072,0.02236355510022905
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,5120,512,0.002579555536309878
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3584,6144,0.026208889153268602
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,32,65536,0.05000355508592394
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,6144,1536,0.006616000086069107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,7168,8192,0.03522488805982802
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,64,5120,0.004816888935036129
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,512,8192,0.009710222482681274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,256,16384,0.01605244477589925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,65536,12288,0.185860448413425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,32,65536,0.025761778155962627
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,16384,65536,0.1718053287929959
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,10240,2560,0.027772444817754958
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,6144,2048,0.007450666692521837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,12288,256,0.03167377909024557
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,4096,768,0.005616888817813661
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,128,12288,0.037780443827311196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,512,16384,0.023989333046807185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,6144,128,0.01477688882086012
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,6144,10240,0.03789955708715651
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,8192,12288,0.020968000094095867
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,768,32,0.005587555468082428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,4096,3072,0.021219554874632094
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,7168,7168,0.017927999297777813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,128,3584,0.007720000214046902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,1536,4096,0.01753777762254079
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,256,768,0.005196444276306364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,1024,1536,0.014143110977278816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,16384,4096,0.059327112303839795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,16384,4096,0.02715377840730879
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,16384,16384,0.043375111288494535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,4096,2560,0.009001777403884465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2560,1536,0.003858666867017746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,65536,128,0.04327911138534546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,8192,16384,0.043600890371534556
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3584,768,0.005584888988071018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,16384,2560,0.01552088889810774
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,768,7168,0.020608888732062448
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,256,2048,0.006628444625271692
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,128,7168,0.009654222263230218
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,8192,3584,0.012359999948077731
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,4096,65536,0.08057689004474215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,12288,1536,0.027217777238951787
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,12288,7168,0.41444354587131077
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,32,1024,0.0031786666562159858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,6144,3584,0.01349688900841607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,65536,7168,0.07688444190555148
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,768,1536,0.0064693333374129395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,768,3072,0.015284443895022074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,16384,2048,0.012172444827026792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1536,512,0.0026071110947264563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,10240,2048,0.01109155515829722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,12288,5120,0.016591999265882704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,2048,3072,0.007364444434642792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,3072,6144,0.033379554748535156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,512,512,0.002576888849337896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,65536,12288,0.3205475542280409
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,32,768,0.002945777856641345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,8192,512,0.0036231109665499795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1536,128,0.005688000056478713
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1024,128,0.005430222385459476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,10240,3584,0.11579644680023193
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3072,2048,0.008734222087595198
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,1536,256,0.024671999944580927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,1024,3584,0.028731554746627808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,7168,1024,0.019888000355826486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,768,512,0.002830222249031067
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2560,2048,0.004917333523432414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,16384,7168,0.053321778774261475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3584,32,0.005581333405441708
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2048,65536,0.12377866109212239
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,4096,2048,0.08109955655203925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,512,65536,0.1777902179294162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,16384,256,0.008450667063395182
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,12288,2560,0.012104888757069906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,4096,5120,0.03188177943229675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2048,768,0.00628977765639623
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,12288,16384,0.040467556979921125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,16384,5120,0.025822222232818604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,32,8192,0.005593777944644292
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1536,12288,0.014276444911956787
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,10240,65536,0.14413511753082275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,2048,4096,0.007983999947706858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,7168,1536,0.007334222396214803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2048,2048,0.004569777597983678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,6144,3072,0.00794933322403166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,5120,256,0.0028657778683635923
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,4096,1024,0.003855111284388436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,7168,8192,0.056193775600857206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,256,7168,0.019693333241674636
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1536,7168,0.01016711112525728
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,10240,256,0.02199111051029629
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,3584,256,0.0053048887186580235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,5120,10240,0.025077333052953083
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3072,128,0.0028248888750871024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,12288,12288,0.46134577857123477
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,1536,256,0.01809866726398468
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1024,12288,0.009397333694828881
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,5120,6144,0.032441778315438166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,256,10240,0.005686222265164058
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,10240,3584,0.03604977660708957
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3584,256,0.002543111021320025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2560,1024,0.005117333183685939
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,8192,768,0.006259555617968242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,4096,8192,0.014127999544143677
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3584,128,0.002141333288616604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,256,4096,0.007330666813585494
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2048,8192,0.009137777818573846
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,65536,1536,0.03949866692225138
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,8192,4096,0.01313244468635983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1024,16384,0.011772444678677453
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,5120,5120,0.022079111801253423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,4096,16384,0.022797332869635686
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,12288,1536,0.013292444248994192
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,8192,16384,0.04641155401865641
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,7168,3584,0.010856889188289642
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,6144,256,0.005605333381228977
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2560,1024,0.00591111100382275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2560,2048,0.0069644442862934535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,6144,768,0.005939555664857228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,512,10240,0.015560888581805758
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,7168,1024,0.018645332919226754
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2560,32,0.0069422221018208405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,32,768,0.003173333282272021
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,10240,3072,0.012705778082211813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,5120,64,0.0052595556610160405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,3072,6144,0.025736888249715168
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,5120,3584,0.008288000192907121
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,6144,1536,0.008279110822412703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,768,12288,0.024549333585633173
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3584,7168,0.05491822295718723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,12288,3072,0.0328693323665195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,256,2048,0.028889778587553237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,4096,2560,0.006754666566848755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,4096,2048,0.006642666541867786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,10240,256,0.02223288847340478
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,16384,8192,0.04044888748062982
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,5120,1024,0.005939555664857228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3072,512,0.005626666463083691
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,128,10240,0.02537155482504103
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,256,768,0.002899555489420891
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3584,6144,0.205441779560513
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,64,2048,0.004618666652176115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,65536,2048,0.07851999998092651
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,32,768,0.0062977779242727495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,2048,8192,0.010608889162540436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,65536,1536,0.021253334151373968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,65536,7168,0.15199111567603216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,12288,5120,0.05615555577807956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,6144,7168,0.020255999432669747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,7168,2048,0.00869777798652649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2048,512,0.0033928888539473214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,7168,64,0.002838222309947014
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,65536,128,0.02202933364444309
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,256,65536,0.2771208816104465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,1536,65536,0.07516000005933973
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,1024,2048,0.014864888456132678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,1536,1536,0.007267555428875818
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,16384,1024,0.025771554973390367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,10240,32,0.009527111219035255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,16384,65536,18.688500298394096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,65536,7168,1.6790044572618272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,5120,3584,0.007876444194051955
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,32,65536,0.09646311071183945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,3072,10240,0.04349244303173489
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,768,32,0.0069164443347189165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,512,2560,0.010473777850468954
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,1024,2048,0.016999999682108562
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,16384,10240,0.034635557068718806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,8192,10240,0.056815114286210805
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,2560,6144,0.025316443708207872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,8192,8192,0.02379200028048621
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,128,32,0.0021457777669032416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,65536,10240,0.25909866227044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,4096,256,0.007299555672539606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,4096,768,0.016650666793187458
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3072,2048,0.006983111302057902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3072,10240,0.023927999867333308
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,128,1024,0.012076444096035428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,8192,512,0.018331555856598746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,3584,16384,0.02233244478702545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,7168,2560,0.021838222940762837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,12288,16384,0.08834311034944321
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,128,65536,0.06107288599014282
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3584,1024,0.03754222061898973
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,768,32,0.0021342221233579847
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3584,3584,0.012121777567598553
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3072,32,0.0021991110924217436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,2560,4096,0.008350222474998897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,32,8192,0.004583111239804162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1536,3584,0.0064293332397937775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,5120,3584,0.007632888853549957
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,1536,1024,0.016175111134847004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,256,65536,0.06983377536137898
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,8192,4096,0.015352000792821249
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,16384,10240,0.05416355530420939
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,65536,5120,0.16502132680681017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,1024,8192,0.010500444306267632
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,12288,128,0.011754666765530905
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,8192,768,0.005913777897755305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,10240,12288,0.03038755390379164
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,128,2560,0.006649777707126405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,128,768,0.005595555735958948
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2560,7168,0.009671111073758867
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,4096,16384,0.025578666064474318
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,2560,6144,0.06880533032947116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,10240,10240,0.03560355636808608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,6144,12288,0.0378115541405148
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,256,128,0.0021671110557185277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,128,256,0.002521777732504739
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,256,256,0.005576888720194499
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,3584,8192,0.24857687950134277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,16384,64,0.0058906666106647914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,5120,5120,0.03048444456524319
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,16384,7168,0.04599644409285652
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,512,8192,0.011018666956159802
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,32,12288,0.01295555548535453
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,256,5120,0.00832711077398724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,768,1024,0.005902222047249476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,512,4096,0.007634667058785756
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,128,2048,0.02871911062134637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,65536,1536,0.09225777784983318
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,8192,128,0.04303288790914747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,2048,2048,0.02033244404527876
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,256,7168,0.01108266661564509
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,10240,512,0.019260444574885897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3584,10240,0.03624533282385932
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,7168,512,0.002833777831660377
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,4096,6144,0.01809599995613098
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3072,1024,0.005947555518812603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,16384,16384,0.04349244303173489
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3584,128,0.005930666708283954
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,128,512,0.005519111123349931
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,12288,1536,0.015663110547595553
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,16384,8192,0.03166310985883077
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,7168,1536,0.02843822373284234
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3584,3072,0.007653333246707916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,5120,3584,0.008438222110271454
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,16384,8192,0.06589955753750272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,16384,6144,0.05289511217011345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,768,16384,0.012438221938080259
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,65536,768,0.05004088746176826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,6144,7168,0.012434666355450949
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,4096,32,0.006660444455014334
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,128,256,0.0021893332401911416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,256,512,0.0069439998931354946
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,5120,768,0.018760000665982563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,16384,256,0.017567111386193168
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,6144,64,0.005896889087226655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,5120,5120,0.0248391115003162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,4096,65536,0.0833102199766371
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,128,6144,0.009047110875447592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,512,512,0.005960888746711943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,6144,3072,0.015224888920783997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1536,1536,0.003879111260175705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,2048,5120,0.008614222208658854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,16384,12288,0.1028168863720364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,2560,256,0.005507555686765247
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,12288,512,0.0046737777690092725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,128,768,0.005616888817813661
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,65536,7168,2.0602249569363065
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,12288,3584,0.048044443130493164
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2048,512,0.004223111189073986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1024,64,0.0028515555378463534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,128,4096,0.010392000277837118
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,5120,10240,0.01571911076704661
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,5120,768,0.004230222354332606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,16384,12288,0.08464888731638591
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,64,128,0.002165333264403873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,4096,1536,0.006320888797442119
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,10240,6144,0.01738755570517646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3072,1024,0.005975111077229182
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,128,2048,0.005936000082227919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,65536,2560,0.1142044464747111
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2048,5120,0.008962666822804345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,128,16384,0.015527110960748462
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,12288,512,0.010464888479974536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,3584,16384,0.06166399849785698
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,1024,128,0.006598222172922558
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2560,12288,0.018741332822375827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,3584,6144,0.021299555897712708
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,7168,6144,0.04177600145339966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,512,3584,0.0053084443012873335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,256,65536,0.11036977503034805
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,512,64,0.0022168887986077205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,12288,7168,0.020717332760492962
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2560,768,0.007374222079912822
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,65536,2048,0.036483556032180786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,4096,65536,0.08225866821077135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,3072,5120,0.008660444782839881
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2048,7168,0.018331555856598746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3072,2048,0.00490044429898262
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,3072,5120,0.009341333475377824
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1536,10240,0.02222577730814616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,64,1536,0.005949333310127258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,2048,6144,0.009407111340098912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,3584,12288,0.034985777404573225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,4096,3584,0.009894222021102905
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,7168,2048,0.013580444786283704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,2560,512,0.011384000380833944
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,12288,256,0.006229333165619109
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2560,4096,0.00793599999613232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,65536,12288,0.10797511206732856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1024,65536,0.02573866645495097
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1024,6144,0.03953244288762411
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,1024,768,0.00565155554148886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,2560,8192,0.024102222588327196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,128,1024,0.005643555687533484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,32,512,0.0025822222232818604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2560,3072,0.016935111747847665
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,65536,3584,1.4361520343356664
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,32,4096,0.014215999179416232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,12288,12288,1.0151706271701388
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,10240,256,0.02607466611597273
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,3584,1024,0.02815199891726176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,7168,5120,0.027140445179409448
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,8192,1536,0.05625066492292616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,2560,3072,0.007781333393520779
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,2048,1024,0.013497778111033969
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3584,3072,0.007387555307812161
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,512,128,0.005345777918895085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,7168,32,0.00565155554148886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,10240,2048,0.01257511145538754
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3584,65536,0.06632533338334826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,2560,2560,0.007303111255168915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,8192,6144,0.015616888801256815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,32,1536,0.006301333506902059
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1024,7168,0.008344000412358178
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,7168,1024,0.01831644442346361
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,4096,7168,0.012131555212868584
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,10240,64,0.004503111044565837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,768,5120,0.005311111195219888
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,128,10240,0.023230221536424425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,12288,2560,0.0121760004096561
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,1536,6144,0.020610666937298246
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,8192,1024,0.006181333214044571
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,32,65536,0.04697155621316698
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,65536,8192,0.1805164416631063
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,128,1536,0.012488889197508493
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,12288,64,0.0069715554515520734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2560,3072,0.007369777394665613
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,12288,256,0.004973333328962326
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2560,1024,0.01811733345190684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,4096,10240,0.018770666586028207
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,32,256,0.0063635553750726916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,12288,768,0.008126222425036961
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,32,768,0.003906666818592283
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,6144,512,0.016491555505328707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,256,5120,0.00832177781396442
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3072,7168,0.01385244478782018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,12288,128,0.015235554840829639
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,5120,3584,0.00870311094654931
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,256,64,0.0019493332753578823
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,65536,1024,0.015099555253982544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,3072,4096,0.02334044377009074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3072,64,0.0021600000974204806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,12288,7168,0.032051556640201144
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,768,768,0.005950222412745158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,65536,12288,0.12335733572642009
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,10240,4096,0.028311110205120508
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,512,512,0.0028106667515304354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,768,4096,0.011740444435013665
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,3072,512,0.005274666680230035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2560,64,0.005554666535721884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,1536,768,0.01752355529202355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,7168,1024,0.07074399789174397
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3072,3584,0.007608888877762689
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,128,512,0.005564444594913059
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,6144,32,0.009034666750166152
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,5120,3584,0.026001777913835313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,12288,2048,0.16153866714901394
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,4096,4096,0.014969777729776172
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,16384,512,0.022319111559126113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,256,5120,0.008965333302815756
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,5120,7168,0.014173333843549093
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,5120,7168,0.014223110344674854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,5120,3584,0.007986666427718269
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,8192,10240,0.03003822101487054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,512,16384,0.015867556134859722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2048,10240,0.21325600147247314
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,512,12288,0.013438222308953604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,1024,512,0.013438222308953604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,65536,64,0.010047111246320937
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,8192,10240,0.026007999976476032
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,3072,768,0.019609777463806998
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,8192,8192,0.40069156222873265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,3072,5120,0.009007111191749573
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,2048,512,0.007631999750932057
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,64,2048,0.00904444439543618
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2048,512,0.005912000106440649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,2560,16384,0.03395911057790121
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3584,2048,0.013781333135233985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,6144,4096,0.03257866700490316
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,16384,65536,9.280570136176214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,7168,512,0.0063075555695427795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,5120,8192,0.03827022181616889
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,4096,256,0.002569777684079276
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2560,32,0.0033111110743549135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,128,5120,0.005016889009210798
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,256,12288,0.01313511116637124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,768,32,0.00566133318675889
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,5120,6144,0.017871111631393433
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,256,512,0.0069475554757648045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,3584,1024,0.014817777607176038
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2560,32,0.0025555555605226094
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,128,10240,0.005679111099905438
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,256,4096,0.008226667013433244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,2048,5120,0.018730666902330186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,65536,16384,0.22922221819559732
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2048,768,0.007708444363541073
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,5120,65536,0.07255999909506904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,32,6144,0.017665778597195942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,512,12288,0.013252444565296173
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,3072,5120,0.03436444534195794
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3584,10240,0.019350222415394254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3072,2560,0.007301333049933116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,16384,3584,0.01901955571439531
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,512,256,0.005271111097600725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,128,12288,0.016692444682121277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,10240,3072,0.033182223637898765
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,5120,1024,0.0059368887709246735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,8192,3584,0.029211554262373183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,256,2048,0.013771555489963956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,768,128,0.005601777798599667
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,3072,7168,0.009787555370065901
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1024,768,0.005615111026499007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2048,2560,0.00490488898422983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,8192,256,0.006030222194062338
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1536,512,0.002846222163902389
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,256,1024,0.003136000078585413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,4096,65536,0.28052621417575413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,8192,3072,0.11762844191657172
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3584,7168,0.01484444406297472
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,16384,4096,0.04710577925046285
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3072,7168,0.015423110789722867
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,8192,7168,0.019354666272799175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,7168,12288,0.02975999977853563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,256,10240,0.02499466637770335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,2560,8192,0.028632889191309612
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,512,768,0.009552000297440423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,2560,3072,0.007308444215191736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,768,3584,0.016879111528396606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2560,7168,0.011358222199810876
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,128,3072,0.015170666906568738
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,10240,8192,0.02861600120862325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,12288,1024,0.008238222036096785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1024,1024,0.005950222412745158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3072,16384,0.02400355537732442
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,256,512,0.005718222094906702
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3584,2048,0.0070222218831380205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,5120,5120,0.011790222591824003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,4096,3072,0.009521777431170145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,6144,8192,0.029842668109469946
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,7168,768,0.006099555641412735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,1024,2560,0.02365600069363912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,65536,8192,0.14810310469733343
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,128,1536,0.0063697778516345555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,8192,2560,0.08513422144783868
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,65536,5120,0.16410666041904026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,5120,4096,0.011001778145631155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,2048,768,0.005640000104904175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,7168,256,0.005949333310127258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,3072,256,0.02309600015481313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,10240,512,0.006240889016124938
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,128,256,0.005256888767083486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,2048,2048,0.016188444362746347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,1536,4096,0.03648533423741659
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,512,1024,0.005956444475385878
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1024,10240,0.011750222080283694
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,3072,8192,0.024728000164031982
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3072,4096,0.008740444150235917
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3072,2560,0.0090524446633127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,3584,4096,0.007678222325113084
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,3072,2048,0.0062871111763848206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,4096,7168,0.013393777940008374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,5120,10240,0.04397777716318766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,4096,10240,0.04714222086800469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,6144,3072,0.008346666892369589
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,16384,4096,0.20555733309851754
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3584,65536,0.0648133357365926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,5120,1536,0.022728888524903193
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1536,3584,0.0063279999627007385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,5120,1024,0.009356444080670675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,3584,6144,0.01033866653839747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,2560,512,0.005568888866239124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3072,6144,0.01810933318403032
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,12288,4096,0.028611554039849177
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,2048,2048,0.008104888929261101
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,5120,256,0.0052355556852287715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,8192,12288,0.24117067125108507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,2560,1536,0.006030222194062338
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,256,2048,0.012633778154850006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,6144,6144,0.013941332697868347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,3584,768,0.016561778055297006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,512,2560,0.007287999822033777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,7168,512,0.01739377776781718
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,4096,7168,0.02962933315171136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,65536,2560,0.10530044635136922
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,2560,65536,0.43675555123223203
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,6144,2048,0.006605333338181178
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,6144,10240,0.017294221454196505
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,2560,8192,0.04073333409097459
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,6144,3072,0.006969778074158563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,16384,64,0.006635555376609166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,1536,512,0.013082666529549493
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,512,1024,0.007285333342022366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,512,2560,0.0099973330895106
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,2560,3072,0.015855999456511605
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,512,256,0.0023173333870040043
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,8192,8192,0.05070933368470934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,512,16384,0.015934222274356417
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,32,4096,0.005234666582610872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,1024,12288,0.013164444102181328
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3072,6144,0.07490755452050103
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,8192,1024,0.01924977699915568
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,8192,1536,0.006934222247865465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,256,1536,0.003912888881233004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,2560,8192,0.0251528885629442
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,10240,3072,0.014739554789331226
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,2048,7168,0.00943199959066179
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1536,3072,0.007304888632562425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1024,3072,0.010332444475756751
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,16384,10240,0.05284622311592102
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,10240,7168,0.034055110481050276
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,8192,2560,0.13560799757639566
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,768,12288,0.04321066538492838
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1024,10240,0.01237422227859497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,768,7168,0.020301333732075162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,4096,768,0.00620266670982043
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3584,5120,0.02276444435119629
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,16384,16384,0.0652737749947442
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,8192,3584,0.042783111333847046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3584,2560,0.007257777783605788
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,6144,768,0.0084915558497111
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,12288,256,0.002560888934466574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,3072,128,0.023573334018389385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,4096,3584,0.007986666427718269
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,256,1536,0.01442488862408532
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,768,10240,0.028331554598278467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,128,1536,0.006231110956933763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,12288,12288,0.031011554929945204
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,10240,3072,0.04545066754023234
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,5120,16384,0.022815111610624526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1536,3072,0.007257777783605788
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,32,32,0.0018453333112928604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,256,65536,0.01575377749072181
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,10240,512,0.006690666493442323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,12288,768,0.018564444449212816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,2560,5120,0.017650667164060805
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,5120,3072,0.012465777496496836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,16384,768,0.0069955554273393415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,10240,768,0.007844444778230455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,6144,1024,0.03016977839999729
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,768,7168,0.01000355515215132
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,3584,5120,0.021985777550273474
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,8192,7168,0.06557866599824694
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,3584,1536,0.007304000357786815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,5120,5120,0.010305778019958073
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,256,128,0.0052693333062860705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,10240,32,0.004576000074545543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,16384,256,0.0069413334131240845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,65536,1024,0.06334222025341459
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,10240,7168,0.04707555638419258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,65536,1536,0.07695377535290189
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,65536,64,0.00629688882165485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,32,2048,0.023694222172101338
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1024,3072,0.007617777420414819
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,5120,768,0.002854222224818336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,12288,32,0.0032097777972618737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,5120,10240,0.018582221534517076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,768,65536,0.06213866339789497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,768,3072,0.01716088917520311
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,3584,3072,0.019243554936514962
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,512,10240,0.009100444614887238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,7168,128,0.010317333042621613
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,768,12288,0.028183999988767836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,1536,12288,0.25286756621466744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,6144,6144,0.029494222667482164
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,65536,3584,0.04212355613708496
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,8192,768,0.005976888868543837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,3584,5120,0.024998222788174946
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,512,10240,0.006917333437336816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1024,10240,0.012704888979593912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,4096,8192,0.030477331744299993
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,8192,128,0.0021751111166344746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,10240,128,0.015492444237073263
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1536,256,0.005592000153329637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,1536,1024,0.0059164443777667145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,2560,2048,0.0422364440229204
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2048,4096,0.00870488915178511
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,7168,16384,0.027013333307372198
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,128,768,0.006627555522653792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,256,12288,0.022634666826989915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,3072,512,0.014783999986118741
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,4096,2560,0.09649155537287395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,3584,512,0.005635555419656966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,128,8192,0.01037333326207267
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,16384,2048,0.024877332978778418
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,7168,12288,0.5662373436821831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,32,512,0.005583999885453119
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,8192,1536,0.006287999865081575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,65536,4096,0.11681332853105332
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,4096,65536,0.1798177825080024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,5120,10240,0.04347999890645345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,3584,1024,0.008013333711359236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,4096,1536,0.018192888961897958
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,2048,128,0.014829332629839579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,1024,5120,0.00814488861295912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,6144,256,0.002463999928699599
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3072,7168,0.011903111305501727
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,4096,5120,0.00903733323017756
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1536,32,0.0022088889446523455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,1024,8192,0.02107288936773936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,5120,12288,0.01793333391348521
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2560,2048,0.0052639999323421055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,65536,4096,0.10276089111963908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,768,2048,0.006253333141406377
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,768,768,0.005568000177542369
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,2048,1536,0.00666844430896971
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,65536,3072,0.03288088904486762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,8192,64,0.0069751110341813825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,64,768,0.002936888900068071
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,32,3584,0.007654222349325816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,1024,3072,0.012298666768603854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,4096,64,0.005626666463083691
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,16384,64,0.006291555447710886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,128,7168,0.024840000602934096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,32,5120,0.00905777762333552
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,2560,10240,0.14334577984280056
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,6144,6144,0.018415111634466384
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2048,128,0.0032711111836963226
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,16384,5120,0.039657778210110135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2560,5120,0.009578666753239101
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,7168,65536,0.09570399920145671
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1536,12288,0.012517333030700684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,12288,2560,0.01349955548842748
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,65536,3072,0.05053333441416422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,512,2048,0.013807999591032663
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,64,128,0.005588444570700328
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3584,512,0.005255110975768831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,7168,768,0.006009777800904379
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,128,16384,0.02944799926545885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,768,256,0.014828445182906257
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,128,256,0.005639111002286275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,4096,3584,0.013015111287434896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,10240,7168,0.02158755560715993
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,512,512,0.012436444560686747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,7168,6144,0.012080888781282636
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,10240,1024,0.01962311069170634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,32,12288,0.004978666702906291
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,6144,5120,0.016929777132140267
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,4096,1024,0.005959110955397288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,64,6144,0.005236444373925527
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,256,3584,0.015052444405025907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,10240,3584,0.014767999450365702
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,512,128,0.012813333008024426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,128,3072,0.007336888876226213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,512,8192,0.007755555212497711
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,16384,10240,0.0332951115237342
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,16384,65536,0.2177635563744439
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3072,3072,0.014752000570297241
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,3072,2560,0.0069573331210348345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,8192,4096,0.015772443678643968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,2560,256,0.005990222096443176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,12288,2560,0.03029511041111416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2560,3072,0.007324444750944774
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,2048,7168,0.02869066596031189
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,5120,1024,0.0073902226156658595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,5120,512,0.006983111302057902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1536,256,0.005525333185990651
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,16384,1024,0.027090667022599116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,1536,5120,0.023746665981080797
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,5120,256,0.005568888866239124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,8192,5120,0.017699556218253244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3584,512,0.006274666637182236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,10240,2560,0.0114720000161065
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,32,128,0.002193777718477779
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,5120,16384,0.03645599881807963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3072,4096,0.008739555875460306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,4096,256,0.0028355556229750314
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,32,3584,0.010686222049925061
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,65536,3072,0.4548373222351074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,7168,8192,0.01564711166752709
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,1024,768,0.012562666502263812
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1024,16384,0.07516000005933973
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,2048,65536,0.1456364393234253
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,3584,128,0.024728889266649883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,512,128,0.0021866667601797315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,8192,65536,0.09379822015762329
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,10240,5120,0.033850666549470686
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,768,6144,0.056186669402652316
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,3584,1024,0.015503111812803479
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,5120,4096,0.023380445109473333
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,1536,65536,0.07574488719304402
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,768,16384,0.016152888536453247
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,16384,6144,0.0303146673573388
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,512,1536,0.013556444810496436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,512,10240,0.02095111045572493
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,256,768,0.007333333293596904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,4096,10240,0.10562755664189656
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,128,128,0.0069822221994400024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,65536,4096,0.3669777711232503
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,512,3584,0.005682666682534748
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,256,7168,0.040258665879567467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,2560,2048,0.015239111251301236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,6144,32,0.005961777849329843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,10240,8192,0.0363395545217726
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,10240,512,0.004173333446184794
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1536,768,0.0062773335311147906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,256,1024,0.0032257777121331957
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,10240,10240,0.02587288949224684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3072,1024,0.007968000239796108
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2560,6144,0.00996977753109402
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,128,128,0.005658666706747479
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,128,6144,0.016527111331621807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,32,5120,0.004596444467703502
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,256,1024,0.012173333101802401
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,6144,1536,0.005403555515739653
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2560,128,0.005979555762476391
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3584,2560,0.007330666813585494
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,10240,12288,0.06578844123416476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,16384,5120,0.026550221774313185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,1536,8192,0.040575110250049166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,128,128,0.011728000309732227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,2048,3584,0.016374222106403775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,6144,12288,0.01942044496536255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,256,10240,0.012410666379663678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,7168,6144,0.14105599456363255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,12288,2048,0.025260445144441392
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,128,256,0.006641777853171031
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,512,256,0.014471999473041959
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,512,6144,0.031425777408811785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,5120,8192,0.01462222304609087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,7168,512,0.02106488909986284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,32,8192,0.011050666371981302
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,12288,7168,0.14260000652737087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,128,256,0.002144888871245914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,1024,3072,0.015985776980717976
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,512,16384,0.015823110938072205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,16384,2048,0.03289422061708238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,512,6144,0.0081395556529363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,512,512,0.01033688916100396
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,256,2560,0.0064177778032090925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,32,1024,0.0028408887899584244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,7168,1024,0.005983110931184556
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,16384,4096,0.04321155614323086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,8192,64,0.002190222135848469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,16384,1536,0.03016266557905409
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,4096,65536,0.08176622125837538
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,32,16384,0.0169493340783649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,64,128,0.0052728888889153796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1536,65536,0.04079733292261759
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,2560,10240,0.038889778984917536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,1024,3584,0.01927911076280806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,12288,256,0.01718933383623759
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,12288,16384,0.2803013324737549
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3072,2560,0.006235555642180973
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,64,4096,0.008000888758235509
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,8192,3584,0.19944000244140625
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2048,1536,0.006272000157170826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,32,64,0.0052435555391841466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,2048,2560,0.006968888971540663
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,65536,256,0.008266666697131263
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,8192,3584,0.013807999591032663
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,6144,65536,0.08047911193635729
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,1536,10240,0.017271111408869427
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,5120,768,0.0058266665372583605
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,7168,6144,0.012641777594884237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,10240,2560,0.013926222920417786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,4096,8192,0.014283556077215405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,64,3072,0.005591111050711737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,16384,1536,0.010699555277824402
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,64,2048,0.006686222222116258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,7168,5120,0.04829600122239855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,4096,128,0.016215999921162922
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,64,1024,0.007325333025720384
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,65536,6144,0.1470293336444431
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,256,65536,0.06262044774161445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2560,64,0.005296000176005893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,2560,3072,0.007252444823582967
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,7168,6144,0.015079110860824585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,8192,6144,0.17161067326863608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,5120,4096,0.10319199826982285
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,7168,1536,0.012323555847009024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1024,2048,0.006616000086069107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,8192,3072,0.02955377764172024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,1536,1536,0.0063324446479479475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,65536,8192,0.07624177800284491
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,12288,128,0.008937777744399177
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,10240,10240,0.03631733192337884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,8192,256,0.002817777709828483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,2560,1024,0.005967111223273807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2560,64,0.0052675555149714155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,16384,1536,0.010409778190983666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,32,5120,0.004870222260554631
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,2048,16384,0.05426222085952759
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,1536,12288,0.027591110931502447
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,32,3072,0.006986666884687211
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,65536,10240,3.0231511857774525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,1024,16384,0.022797332869635686
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,12288,256,0.012144889268610211
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2560,1536,0.006483555667930179
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,32,256,0.002234666711754269
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,12288,6144,0.12062844965193008
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,16384,3584,0.019896888070636325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,5120,256,0.008772444393899705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,64,512,0.005718222094906702
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,32,512,0.005745777653323279
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1024,4096,0.013247999880048962
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,4096,2560,0.0074924445814556545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,8192,2560,0.008338666624493068
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2560,768,0.0032142222755485107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,512,768,0.005581333405441708
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,16384,2048,0.029225776592890423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,8192,2048,0.007407111426194509
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3072,128,0.002977777686383989
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,6144,256,0.03144355614980062
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,256,128,0.00508355556262864
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,6144,65536,0.07884622282452054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,1024,512,0.005955555372767978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,10240,2560,0.01146311147345437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,7168,10240,0.034663110971450806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,128,1024,0.012123555772834353
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,4096,10240,0.015653333730167814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,7168,4096,0.014871999621391296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,12288,4096,0.020015110572179157
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,7168,2560,0.031025777260462444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,16384,10240,0.1042791141404046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,128,4096,0.008012444608741337
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,7168,2560,0.04396177662743462
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,5120,4096,0.013886222408877479
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,65536,3072,0.036405334870020546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,128,768,0.010055999788973067
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2048,512,0.005599110904667113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,3072,3584,0.11049066649542914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,1024,768,0.015538665983412
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3584,1024,0.00647555540005366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,7168,128,0.02693422304259406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2048,2560,0.007256000406212277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,5120,6144,0.02532533307870229
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,512,7168,0.00959911114639706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3584,512,0.005608888963858287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,32,7168,0.010738666686746808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,32,5120,0.012667555775907306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,65536,128,0.009775111244784461
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,7168,10240,0.05540800094604492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,128,10240,0.013557333085272046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,7168,2048,0.023432888918452795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,4096,5120,0.01242844429281023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,4096,6144,0.06504799922307332
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,128,512,0.005520888914664586
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,1536,12288,0.0138586668504609
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,7168,1536,0.008665777742862701
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,8192,1536,0.006984000404675801
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,7168,8192,0.022816888160175745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,16384,8192,0.026058667235904272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,512,64,0.0021671110557185277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,256,1536,0.006294222341643439
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,128,8192,0.011045333411958484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,512,32,0.002149333349532551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,8192,512,0.03343822227583991
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,65536,1024,0.01534844438234965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,32,8192,0.010447999669445885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,7168,512,0.01717866626050737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,256,3584,0.006204444501135085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1024,65536,0.6353128751118978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,128,2048,0.0254737784465154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,7168,128,0.03501688771777683
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,1536,12288,0.02833066715134515
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,12288,1024,0.02160888910293579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,2560,6144,0.009681777821646797
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,3072,5120,0.008300444318188561
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,3584,8192,0.01388088862101237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,32,65536,0.044347554445266724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,1536,2560,0.015925332903862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,128,4096,0.007649777664078607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,4096,10240,0.016541333662139047
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,512,3072,0.007306666837798224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,6144,8192,0.0333262218369378
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,3584,6144,0.011360888679822287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,256,12288,0.013528889252079858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,512,8192,0.005250666704442766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,768,1536,0.004592888885074192
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,8192,3584,0.03281955586539374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,2048,10240,0.03432622220781114
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,512,5120,0.008360000120268928
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,6144,5120,0.046718223227394946
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,8192,3584,0.03161155515246921
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,12288,768,0.007481777833567725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,128,256,0.00700800038046307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1536,1024,0.012000888586044312
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,6144,512,0.005671999934646819
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,8192,4096,0.013154666456911298
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1536,8192,0.03580800029966567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3584,256,0.00795377790927887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,128,3584,0.02589688864019182
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,5120,5120,0.009173333644866943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,16384,128,0.015476443701320224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,768,768,0.016540444559521146
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,32,65536,0.08425511254204644
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,16384,128,0.002536000063021978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,3584,6144,0.02900444467862447
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,1536,4096,0.016173332929611206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,256,2560,0.007298666569921706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,32,12288,0.004953777624501122
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,32,2048,0.004585777719815572
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,65536,4096,0.08715644147660996
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1024,3584,0.007864889171388414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,2048,1024,0.015815110670195687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,128,512,0.005599110904667113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,16384,5120,0.03830844495031569
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,3584,128,0.0053342220683892565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,7168,2048,0.024511999554104272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2560,65536,0.06421244144439697
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,5120,8192,0.21482843822903105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2048,65536,0.06786488824420504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,256,768,0.009395555489593083
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,5120,512,0.005561777700980504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,12288,256,0.004295111116435793
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,8192,768,0.006320000108745363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,16384,16384,0.3832533359527588
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,8192,512,0.01146044416560067
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3584,2560,0.01590577761332194
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1024,3584,0.008330666356616551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3072,512,0.005782222168313132
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,5120,65536,0.13892443974812826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,32,3072,0.009404444032245213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1024,16384,0.16718933317396376
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,1536,65536,0.06339466571807861
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,3584,6144,0.023595554961098567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1536,1024,0.003945777813593547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3072,2048,0.054120000865724355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2560,1024,0.03247288862864176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,2560,2560,0.017519111434618633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,4096,1536,0.007114667031500075
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,4096,3072,0.007287999822033777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,7168,5120,0.017050666941536795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,7168,4096,0.020326221982638042
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,2048,6144,0.02646222213904063
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3072,32,0.005555555638339784
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,1536,5120,0.008546666966544257
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3072,6144,0.011047999891969891
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2560,6144,0.009667555491129557
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2048,2560,0.006933333145247565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,1024,4096,0.011425777441925473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,5120,2560,0.007954667011896769
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2560,2560,0.01517955627706316
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3072,1536,0.009165333376990424
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,16384,3584,0.015871110889646742
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1024,5120,0.0059262220230367445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,7168,4096,0.012358222570684222
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,10240,2048,0.01570755574438307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,3584,3072,0.007320889168315464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,7168,32,0.009019555317031013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,10240,1536,0.007616889145639207
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,512,65536,0.34193510479397243
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2048,3584,0.01034755590889189
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,256,3584,0.01796266602145301
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,6144,2560,0.009861333502663506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,2048,768,0.020299555526839364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3072,16384,0.02180888917711046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,2048,1536,0.007937777373525832
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,1536,10240,0.012326222327020435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,256,1536,0.00830488900343577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,128,2048,0.01939733326435089
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,4096,8192,0.017550221747822232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,256,256,0.01239911135700014
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,6144,3072,0.07037333647410075
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2048,1024,0.005960000058015187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,32,7168,0.010693333215183683
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,1024,2048,0.014089778065681458
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3072,2560,0.007276444799370236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,10240,8192,0.04519644379615784
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,5120,128,0.014833778142929077
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,8192,256,0.0025128889828920364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,5120,65536,0.20188444190555146
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,5120,1536,0.006132444573773279
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,4096,512,0.005341333233647876
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,2048,1024,0.005919111271699269
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,128,8192,0.018886221779717337
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,2048,16384,0.021107556091414556
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,3584,65536,0.19757155577341715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,768,12288,0.06621599859661527
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,6144,1536,0.09305155277252197
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,5120,2048,0.02088711162408193
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,7168,128,0.007073777417341868
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,512,6144,0.018239999810854595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,8192,2048,0.014264000786675347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,10240,12288,0.030055999755859375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,65536,7168,0.07670133643680148
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,6144,256,0.005542222410440445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,5120,1536,0.006607999818192587
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,256,8192,0.027263111538357202
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,4096,512,0.015810666812790763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3072,16384,0.02386844489309523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,5120,4096,0.009692444569534725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2560,8192,0.013914666242069669
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,6144,512,0.005612444546487596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1536,768,0.006077333456940121
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,1024,3072,0.019115555617544387
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,6144,768,0.009713778065310584
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,3584,256,0.010580444501505958
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,1536,2048,0.018351111147138808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,1024,5120,0.008436444732877944
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,64,12288,0.01351199961370892
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,7168,12288,0.2946888870663113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,768,128,0.013839110732078552
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,8192,7168,0.04169599877463447
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,16384,4096,0.06143289142184787
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,10240,5120,0.018248889181349013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2048,4096,0.008651555412345463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,10240,16384,0.06498933500713773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,6144,3584,0.01071911139620675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,256,2048,0.004548444516128964
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,7168,3584,0.02401955591307746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,768,6144,0.009376889301670922
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3072,3584,0.11026755306455825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,1536,1536,0.00657155571712388
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3072,3584,0.02755822241306305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,768,768,0.015220445063379077
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,512,7168,0.01899111105336083
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3584,64,0.0021831111775504216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,7168,2560,0.008797333472304875
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,5120,3072,0.02606311109330919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,65536,5120,0.046451555358039014
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2560,16384,0.032835556401146784
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1536,512,0.005966222120655908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,6144,65536,0.0732755528555976
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,1024,128,0.014510222607188754
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,3584,3072,0.007320889168315464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,3072,65536,0.0894355575243632
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,12288,6144,0.0207004447778066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,512,1024,0.0032222221295038858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1024,10240,0.022016000416543748
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1024,12288,0.011343999869293638
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,10240,6144,0.017550221747822232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,64,32,0.0018613334331247543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,3584,3072,0.024156444602542456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,3072,768,0.022237333986494277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,6144,8192,0.05381333496835497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,4096,3072,0.023357333408461675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,1536,1024,0.005940444353553984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,8192,3584,0.024375999967257183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,256,1536,0.006339555399285422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2048,65536,0.045999109745025635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,12288,2560,0.044333332114749484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1024,2048,0.006648889018429651
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,2560,12288,0.0950488911734687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3584,2048,0.005118222286303838
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1536,512,0.006694222076071634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,512,512,0.002736888825893402
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,6144,16384,0.05733422438303629
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,512,512,0.0025742221623659134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,512,16384,0.009371555513805812
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,65536,3072,0.06615644693374634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,5120,6144,0.01722133325205909
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,3072,3072,0.016527111331621807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,5120,768,0.04399111204677158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,256,512,0.0026142222599850762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,7168,12288,0.03074044320318434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,12288,512,0.020428443948427837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,8192,3584,0.01088177743885252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,65536,128,0.010122666756312052
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,256,10240,0.023079110516442194
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,10240,10240,0.029341333442264136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,64,3584,0.005048000150256687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1024,3584,0.009747555686367882
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,4096,256,0.01648977730009291
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,768,5120,0.009777777724795872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,6144,4096,0.02862488892343309
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,128,2560,0.006621333460013072
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,3072,512,0.014177777700954013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,8192,32,0.002130666747689247
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,10240,3072,0.013331555657916598
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3584,32,0.006343110981914733
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,65536,5120,0.07178755601247151
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,32,6144,0.004583111239804162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,16384,3072,0.014811555544535318
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,12288,12288,0.1161395576265123
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,6144,32,0.00222044438123703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,10240,10240,0.039656000004874334
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,16384,5120,0.04686577783690559
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3072,32,0.01276000009642707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,256,768,0.010057777994208867
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,12288,64,0.0021626665774318907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,768,512,0.0028773333049482773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,10240,768,0.01923111081123352
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1024,65536,0.05444799860318502
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,3072,2048,0.006535999890830781
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3072,5120,0.011360000405046674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,10240,3072,0.030438221163219873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,64,7168,0.005583111196756363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,12288,5120,0.034112887250052557
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,7168,4096,0.01272266689274046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,256,128,0.005607999861240387
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,4096,16384,0.02125599980354309
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,32,5120,0.04048000110520257
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,64,3072,0.004926222066084544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,2048,8192,0.010779555473062726
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,12288,5120,0.021920889616012573
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,256,10240,0.005936000082227919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,4096,256,0.0063973334100511335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,256,768,0.005723555468850666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,8192,1024,0.024567999773555334
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,10240,1536,0.007821333077218797
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,4096,1024,0.0058346668051348785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,768,4096,0.00664533343580034
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,512,3584,0.017253332667880587
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,4096,4096,0.023670223024156358
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,4096,768,0.005904888941182031
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,6144,3072,0.009507555928495195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,2560,2048,0.007299555672539606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,6144,12288,0.03065155612097846
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,32,32,0.0058257778485616045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,768,5120,0.00833422193924586
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,12288,1024,0.006919999917348226
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,1536,512,0.01408266690042284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,5120,128,0.024041778511471216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,2560,10240,0.013829333086808523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1536,5120,0.00871733327706655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,64,512,0.006657777975002925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,5120,6144,0.013379555609491138
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3072,3584,0.008560000194443597
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,3072,2560,0.020662221643659804
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,32,16384,0.0069848886794514125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,6144,512,0.007341333561473423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,5120,7168,0.014646222194035849
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,512,16384,0.05313333206706577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,512,128,0.0053280000057485365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,16384,1536,0.04309155543645223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,512,128,0.005980444451173146
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2048,32,0.002542222125662698
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,256,65536,0.05467377768622505
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,5120,512,0.0032151111712058387
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1536,6144,0.010491555763615502
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,7168,3072,0.04370489054256015
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,512,3584,0.006310222049554189
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,10240,7168,0.024507555696699355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2560,5120,0.012488000094890594
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1536,12288,0.01998044384850396
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,2560,5120,0.008416888614495596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,65536,4096,0.10857155587938096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,256,1536,0.006231110956933763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,128,1024,0.005943999936183293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,3584,1536,0.006346666564544042
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,1024,768,0.006313777632183499
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,12288,5120,0.03165422214402093
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,32,65536,0.05908799833721585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,512,256,0.002238222294383579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,4096,2560,0.024064888556798298
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,8192,6144,0.08532000250286526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,4096,128,0.005568000177542369
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1536,5120,0.009569777382744683
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,7168,768,0.023916444844669763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,65536,1024,0.01702311138312022
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3072,3072,0.0417155557208591
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,16384,7168,0.02223644488387638
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1024,512,0.005280888742870755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,5120,1536,0.020257777637905545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,6144,16384,0.04623022344377306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1536,3584,0.006370666540331311
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,16384,3584,0.019350222415394254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,10240,128,0.006270222365856171
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,8192,32,0.006222222414281633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,6144,512,0.005998222364319696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,6144,16384,0.024805333879258897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,3584,1024,0.024576000041431848
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,10240,2048,0.0087013335691558
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,256,3072,0.024455999334653217
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,8192,12288,0.06423022349675496
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,7168,3584,0.017669333351982962
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,768,7168,0.02295733326011234
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,1536,6144,0.02941688895225525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2048,5120,0.012610666453838348
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,768,65536,0.020611556039916146
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,6144,3584,0.01108444482088089
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,8192,256,0.005913777897755305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3584,16384,0.02165599995189243
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,2560,12288,0.02862488892343309
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,4096,65536,3.0722630818684897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,7168,2560,0.023008000519540574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,65536,3584,0.04751288890838623
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,10240,3584,0.026447110705905493
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,1024,7168,0.02032711108525594
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,2048,3584,0.007642666498819987
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3072,2560,0.0069777775141927935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3584,8192,0.02379111117786831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,12288,5120,0.9099004533555773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,768,256,0.0052355556852287715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,128,6144,0.009020444419648912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,7168,3584,0.018040888839297824
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2560,128,0.005782222168313132
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,4096,4096,0.007977777885066139
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1536,5120,0.008922666311264038
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,1024,3072,0.007334222396214803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,64,256,0.0038728887836138406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,4096,32,0.002228444442152977
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3072,4096,0.008008889026112026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,32,2560,0.006976889239417181
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2560,3072,0.007300444775157505
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,65536,256,0.011722666521867117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,1536,4096,0.009192888935407003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,8192,4096,0.045520000987582736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,4096,4096,0.02012266715367635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,768,128,0.013779555757840475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,256,64,0.00554044461912579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2560,4096,0.0081386665503184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,8192,7168,0.017625777257813346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,8192,12288,0.09824355443318684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,10240,5120,0.07151733504401313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,12288,64,0.004063999901215236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,10240,5120,0.015855110353893705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1536,6144,0.011143999795118967
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2560,12288,0.02570044497648875
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,2560,256,0.014849777022997538
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,6144,768,0.01792888840039571
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,7168,256,0.007976888782448238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,5120,5120,0.023911999331580266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1024,32,0.006615110983451207
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,8192,2560,0.04199377695719401
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,64,1536,0.008336889247099558
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,65536,768,0.02187555531660716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3584,1536,0.00628355559375551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,256,12288,0.013227555486891003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,512,65536,0.05679644478691948
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,32,16384,0.015169777803950839
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,256,65536,0.09077688720491196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,6144,128,0.02509333358870612
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,2048,5120,0.008421333299742805
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,10240,8192,0.024876443876160517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,10240,12288,0.04050666756100125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,1536,4096,0.007987555530336168
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,3072,256,0.0052782222628593445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,3072,768,0.006241777704821692
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,12288,1024,0.12027111318376328
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3584,65536,0.06503555509779188
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,5120,65536,0.06993777884377374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,7168,1024,0.03252266512976752
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,12288,10240,0.030065778228971694
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,6144,3072,0.016650666793187458
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,16384,256,0.006480888773997624
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,6144,12288,0.039192001024881996
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,32,3072,0.004924444274769889
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2048,1536,0.006554666492674086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,3584,6144,0.02311733365058899
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,4096,7168,0.02366133365366194
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3584,256,0.005896889087226655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,512,7168,0.016742222838931613
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3072,1536,0.004567999806669024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,5120,512,0.005952000204059813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,7168,1536,0.005764444255166584
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,2048,5120,0.031118220753139917
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3072,10240,0.016561778055297006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,1536,65536,0.06332000096638997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,6144,5120,0.03552444444762336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,10240,10240,0.07208622164196439
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,2560,2560,0.015839111473825242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,6144,128,0.006434666613737742
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,3072,128,0.005928888916969299
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,32,2048,0.009448889229032729
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,16384,5120,0.04356888930002848
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,512,16384,0.020441777176327176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,1024,65536,0.06261599726147123
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,256,256,0.00526311124364535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,32,10240,0.012728888955381183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,10240,128,0.006075555665625467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,2048,3584,0.023471999499532912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,65536,128,0.010040000081062317
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3584,3584,0.011110222174061669
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,65536,7168,3.673870934380425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,8192,6144,0.02111111084620158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,5120,1024,0.0336097776889801
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,65536,512,0.014614222778214348
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,65536,12288,0.13281688425276014
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,4096,12288,0.03058755397796631
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2048,16384,0.021114667256673176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,65536,64,0.16714844438764784
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,4096,768,0.01825066738658481
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2560,2048,0.004635555462704765
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,768,3072,0.007678222325113084
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,32,256,0.00656355544924736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3072,6144,0.012544000314341651
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,7168,12288,0.06404799885219999
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,5120,3584,0.010071111222108206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,768,7168,0.01938488913906945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,3584,512,0.00564177789621883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,64,12288,0.013487999637921652
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,32,64,0.0019493332753578823
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,256,3584,0.010633777413103314
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,4096,2048,0.016227554943826463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,8192,4096,0.03411022159788344
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,512,3072,0.02959733208020528
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2048,1024,0.008365333080291748
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,128,6144,0.009039999710188972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1536,2560,0.010440888504187265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,4096,12288,0.0584755539894104
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1024,256,0.0070266665683852295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,8192,128,0.007963555554548899
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,5120,5120,0.03403110967742072
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,16384,2048,0.026199110680156287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,7168,2560,0.027637332677841187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,2560,1536,0.017631111873520743
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,2048,2560,0.019724445210562814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,128,4096,0.014865777558750577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,1536,16384,0.0661262207561069
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,7168,4096,0.025143999192449782
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,768,65536,0.04174133472972446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,65536,3072,0.2962453365325928
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,12288,1024,0.02236888971593645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,1024,8192,0.08803821934594049
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,1536,128,0.0053386667536364655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,10240,65536,0.41452889972262913
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,6144,2048,0.0063697778516345555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,5120,4096,0.028220444917678833
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,1536,2560,0.014166222678290473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,8192,5120,0.035911109712388777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,1536,128,0.014492443866199918
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,65536,1024,0.07746666669845581
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,256,128,0.011454222102959951
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,4096,10240,0.016199999385409884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,128,6144,0.009359111388524374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,1024,65536,0.07288799683252971
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,3072,4096,0.017197334104114108
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,256,256,0.00527022240890397
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,256,12288,0.013526221944226159
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,65536,7168,0.10652266608344184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,6144,2560,0.07672533061769274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,7168,256,0.04166133205095927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,768,2048,0.0069751110341813825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,4096,128,0.015574221809705099
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,128,7168,0.010060444474220276
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,256,3584,0.00798044436507755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3584,6144,0.033534222178988986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,128,512,0.011419555379284753
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1024,256,0.002543111021320025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,3584,2048,0.006589333216349284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,64,8192,0.011066666907734342
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,1536,5120,0.00907466643386417
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1536,64,0.0021964444054497615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,7168,7168,0.013971555564138623
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,65536,4096,0.06318133407168917
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,768,4096,0.006731555693679386
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,8192,256,0.005608888963858287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,512,512,0.0052791109515561005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,4096,5120,0.010500444306267632
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,3584,8192,0.031110223796632554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,12288,512,0.020644444558355544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,3584,512,0.007327111230956183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,65536,512,0.05037777622540792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,2560,16384,0.05807377894719442
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,32,3584,0.007674666742483775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2560,3072,0.0846666693687439
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,16384,65536,0.1439359982808431
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,6144,768,0.00590133335855272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,256,4096,0.008325333396593729
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,32,3072,0.010059555371602377
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,512,1024,0.005949333310127258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,2048,1024,0.02052088909678989
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,7168,4096,0.01531377765867445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,65536,512,0.03029511041111416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,65536,8192,0.18887733088599312
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1024,1024,0.00654755574133661
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,8192,256,0.004282666577233208
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2048,2560,0.005250666704442766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,256,1024,0.005691555639108022
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,64,1536,0.003540444291300244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,4096,1536,0.006245333287451003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2560,64,0.002451555596457587
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1536,3072,0.0070088886552386815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,6144,1024,0.0038657776183552216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,128,10240,0.014846222268210517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,65536,1536,0.0688355565071106
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,10240,3584,0.012859555582205454
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1024,128,0.005623999983072281
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,1024,1536,0.017645332548353408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,8192,2048,0.03100444542037116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,256,1024,0.005920889063013925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,4096,1536,0.006296000132958095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,8192,768,0.003546666767862108
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,7168,8192,0.015203555425008139
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,1536,4096,0.024031110935741003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,65536,6144,0.15995555453830296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,65536,2560,0.036399109496010676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,6144,2560,0.007520889242490132
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,10240,6144,0.06881866852442424
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,5120,1024,0.006238222122192383
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,512,12288,0.013192888763215808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,768,8192,0.018213333355055917
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,128,32,0.0053048887186580235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,256,32,0.005588444570700328
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,16384,5120,0.01754755609565311
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,256,1024,0.003244444520937072
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,128,768,0.012708444562223224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,2560,512,0.005577777822812398
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3072,8192,0.014449778530332776
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,2048,512,0.013761777844693927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3584,10240,0.022818666365411546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,768,512,0.013097777962684631
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,3584,7168,0.04286488890647888
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,7168,512,0.018178666631380718
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,5120,2560,0.011295999917719098
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,32,2048,0.011352000137170156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,6144,8192,0.030551999807357788
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,768,2560,0.007273777491516537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,12288,3584,0.04330844349331326
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,2560,8192,0.02478222217824724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,768,65536,0.07571911149554782
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,4096,256,0.00564177789621883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,64,8192,0.0053324442770746015
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,2048,16384,0.03048622277047899
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,32,32,0.006372444331645966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,8192,3072,0.04831555485725403
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,512,32,0.005959110955397288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3584,7168,0.013617777989970313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3584,768,0.029650668303171795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,3584,2048,0.03226577904489305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,16384,65536,0.4695306883917914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,3072,256,0.00562755556570159
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,4096,256,0.01018488903840383
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,8192,7168,0.01700088878472646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,256,7168,0.005655111124118169
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,256,7168,0.009768000079525841
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,6144,3584,0.009400889277458191
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,1536,7168,0.05970755550596449
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,12288,256,0.05039199855592516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,6144,6144,0.012107555237081317
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3584,5120,0.011075555450386472
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,65536,7168,0.05815822548336453
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,7168,7168,0.051190220647388034
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,5120,256,0.005311111195219888
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,7168,2560,0.038204444779290095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,4096,7168,0.02421066661675771
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3072,8192,0.013916444447305469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1536,128,0.005895111295912001
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,64,7168,0.006205333189831839
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,1024,16384,0.02905155552758111
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,768,3072,0.006983111302057902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,4096,12288,0.021185777253574792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,10240,7168,0.02549955579969618
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,6144,12288,0.0717137787077162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,128,10240,0.0209840006298489
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,8192,512,0.009520889156394536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,6144,2560,0.03484800126817491
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1024,5120,0.021431111627154883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,65536,16384,0.3991519874996609
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,16384,256,0.014160888062583076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,2048,8192,0.014488000008794995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2560,512,0.004600000050332812
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,1536,3584,0.018550222118695576
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,6144,3584,0.01094133324093289
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,256,8192,0.011749332977665795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,128,16384,0.011795555551846823
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,64,8192,0.021457778082953557
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,7168,768,0.0059057776298787855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,64,32,0.005618666609128316
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,12288,12288,0.0711324413617452
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,3584,3584,0.007634667058785756
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,2048,10240,0.044496887260013156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,1536,6144,0.009406222237481011
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,4096,8192,0.03434222274356418
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3072,768,0.00600355573826366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1536,10240,0.022060443957646687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,65536,6144,0.08895555469724868
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,128,4096,0.004307555655638377
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1536,4096,0.00798044436507755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,8192,128,0.002536888958679305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,3584,8192,0.025439111722840205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2560,256,0.006664888726340399
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,6144,2560,0.025989333788553875
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,256,4096,0.007341333561473423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,2560,128,0.021915555000305176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,32,8192,0.004388444539573458
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,2560,6144,0.024853333830833435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,256,16384,0.010135999984211391
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,65536,6144,1.5360986921522353
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,6144,2560,0.007728888756699032
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3072,12288,0.021523555119832356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,2560,512,0.005559999909665849
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,16384,4096,0.6265181965298122
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,128,32,0.0018648888087934917
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,10240,128,0.02683111031850179
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2048,32,0.0032053333189752367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,8192,6144,0.029002666473388672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,8192,12288,0.020589333441522386
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,256,6144,0.021640888518757288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,8192,8192,0.019917334119478863
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,12288,1024,0.008079999850855933
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,7168,3072,0.011517333487669626
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1024,6144,0.009378666679064432
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,65536,1536,0.030255112383100722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3584,3584,0.007624000310897827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,64,3584,0.00490488898422983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,512,12288,0.0075395554304122925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,2560,16384,0.020502222908867728
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,3584,7168,0.010905777414639791
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1024,2048,0.006768888897365994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,32,4096,0.004642666627963384
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,256,6144,0.05754933092329237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,3584,8192,0.022679110368092854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,3584,4096,0.13210844993591309
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,8192,8192,0.01677600045998891
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,7168,8192,0.06868088907665677
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,4096,4096,0.007799111306667328
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,128,2560,0.004939555707905027
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,7168,5120,0.08614133463965522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,8192,6144,0.013632888595263163
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3584,3584,0.014269333746698169
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,6144,7168,0.012746666868527731
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,32,3584,0.022621333599090576
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,256,768,0.0028444443725877335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,12288,12288,0.030766222212049697
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1536,32,0.009024889104896123
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1536,768,0.0031964443624019623
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,12288,1536,0.007965332931942409
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,16384,16384,0.07959377765655518
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,512,6144,0.015006222658687167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,65536,10240,0.10747822125752766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,10240,768,0.004902222090297275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,512,512,0.012460444536474017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,256,6144,0.016912889149453905
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,4096,12288,0.037794666157828435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3584,7168,0.013931555880440606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,32,128,0.005960888746711943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,64,6144,0.00941066692272822
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3584,7168,0.013392888837390475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,4096,6144,0.020807999703619216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,256,1536,0.005615111026499007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,8192,2560,0.015190222197108798
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2560,7168,0.012784889174832238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,7168,16384,0.10220444202423096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,7168,8192,0.01573155489232805
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,7168,2048,0.0076142218377855085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,10240,4096,0.01813333398765988
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,6144,3072,0.008408000071843466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,2048,65536,0.08523644341362847
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,512,5120,0.008552889029184977
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,8192,768,0.00902133352226681
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3584,16384,0.021776000658671062
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,2048,256,0.015200000670221118
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,12288,1024,0.007347555624114142
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,8192,4096,0.013783999615245394
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3584,7168,0.014012444350454541
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,12288,8192,0.12577777438693577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,12288,16384,0.07144888904359606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,6144,8192,0.019727999965349834
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3584,5120,0.01145155562294854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,512,1536,0.006000888844331105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,65536,3072,0.08067999945746528
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,7168,10240,0.21823467148674858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,1024,256,0.01481244464715322
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,64,512,0.002495999965402815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,65536,512,0.012524444195959302
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,2048,5120,0.008980444735950893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,4096,1024,0.006635555376609166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,256,128,0.005643555687533484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,6144,10240,0.22139378388722739
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,512,65536,0.055776887469821505
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,1536,65536,0.12471022870805527
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,6144,1024,0.01926399932967292
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2560,16384,0.020436444216304354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,6144,2048,0.027493332823117573
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,5120,12288,0.05663022067811754
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3072,5120,0.013546666337384118
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3584,3584,0.009837333526876237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3584,4096,0.013976888524161445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,2048,2560,0.008017777568764156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,6144,6144,0.038331554995642766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,2048,12288,0.027239110734727647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,768,1536,0.014793778459231058
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,12288,12288,0.03488711184925503
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,768,7168,0.006639110959238476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,3584,8192,0.03321866525544061
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,2560,1024,0.006572444405820634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,6144,16384,0.034833778937657676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,12288,12288,2.314460542466905
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,768,65536,0.06215644545025296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,6144,7168,0.04563733273082309
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,2560,7168,0.03669510947333442
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,1536,4096,0.009751111268997192
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,512,12288,0.018022222651375663
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,7168,3584,0.009815110928482479
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,64,65536,0.0607315566804674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,12288,10240,0.04558488726615906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,1536,768,0.016908443636364408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,768,10240,0.02164622147878011
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1024,512,0.0032542221662071017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,32,1024,0.007629333270920648
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,1024,7168,0.009690666364298927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,8192,8192,0.023249778482649062
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,3584,1024,0.005976000179847081
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,128,3072,0.0075937774446275495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,5120,1024,0.013236444857385425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,2560,6144,0.020925333102544148
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2048,3072,0.03513155712021722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,1024,1024,0.014109333356221518
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,768,5120,0.012999110751681857
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3584,8192,0.26626043849521214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,64,4096,0.007756444315115611
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,7168,1024,0.006350222147173352
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,768,6144,0.006232888748248418
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,128,4096,0.028451555305057104
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,5120,64,0.005557333429654439
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3072,65536,0.0700275566842821
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,512,3584,0.007947555846638149
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,16384,3584,0.13269689348008898
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,65536,10240,0.45667468176947695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,12288,2560,0.0262435558769438
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,2048,1024,0.006259555617968242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1536,512,0.0028595555987623003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,8192,128,0.0021724444296624926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,8192,12288,0.0632408857345581
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,5120,512,0.0037973332736227247
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,1024,3584,0.007718222008811102
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,32,32,0.002136888810329967
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,3072,1024,0.016795555750528973
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,3584,512,0.01548266741964552
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2048,2560,0.0069395556218094295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,512,2048,0.0069475554757648045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,8192,7168,0.017104888955752056
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,64,2560,0.01315199997689989
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2560,1536,0.016175111134847004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,256,4096,0.007899555895063613
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,32,4096,0.00796888851457172
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2560,512,0.008289777570300633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,768,5120,0.01865600049495697
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,2048,2048,0.016951999730534024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,10240,7168,0.027085334062576294
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,2560,4096,0.007991111112965478
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,6144,5120,0.016711999972661335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,128,3072,0.01447111037042406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,5120,8192,0.014545778433481852
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,16384,32,0.005943111247486538
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,12288,6144,0.018277333842383493
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,512,768,0.015207111835479736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,10240,4096,0.015670221712854173
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3072,256,0.0059573331640826324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,6144,8192,0.022410665949185688
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,65536,10240,0.27056532435946995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,512,10240,0.012093333734406365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,7168,7168,0.04116800096299913
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,3584,1536,0.030386666456858318
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,5120,1536,0.01688888834582435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,128,7168,0.006608888920810487
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,2048,1536,0.015469332536061605
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,65536,2560,0.02793777651256985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,32,64,0.002026666576663653
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,65536,2560,0.4684284528096517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,2048,1024,0.009000000026490953
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,4096,2560,0.009104889300134447
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3072,3072,0.009334222310119206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3584,512,0.01276000009642707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,6144,3072,0.02884977724817064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,768,4096,0.009680888719028896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3072,4096,0.009961778091059791
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,32,512,0.003206222214632564
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,16384,512,0.006659555352396435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,8192,256,0.016857778032620747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,4096,512,0.006311999840868845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,65536,32,0.04301599992646111
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,256,16384,0.006591111007663939
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,10240,10240,0.02833066715134515
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,12288,256,0.04815466536415947
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,16384,768,0.006144888699054718
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,65536,1024,0.06419022215737237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,256,12288,0.008408000071843466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,256,512,0.005591111050711737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,1536,1024,0.005998222364319696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,3584,3072,0.02247200078434414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,7168,768,0.015040889382362366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,6144,32,0.0021786666992637846
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,256,7168,0.011153777440388998
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,256,1536,0.003638222399685118
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,10240,12288,0.044482668240865074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,7168,1536,0.006295111030340195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,8192,3584,0.013520888984203339
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,16384,5120,0.058280891842312284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,5120,3072,0.02216800053914388
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,4096,32,0.005913777897755305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,12288,7168,0.020322667227851022
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,8192,12288,0.02994577752219306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,256,128,0.012127999630239276
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2048,64,0.005568000177542369
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,64,6144,0.010018666585286459
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,768,768,0.00564533347884814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,128,768,0.013647110925780402
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,2560,1536,0.006372444331645966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,65536,12288,0.1287013292312622
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2048,768,0.01090311093462838
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3072,1024,0.004501333253251182
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,8192,768,0.039834668238957725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,128,65536,0.04896266592873467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,7168,1024,0.005597333527273602
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,16384,2560,0.07098666826883952
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,10240,4096,0.039664000272750854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,3072,128,0.005311111195219888
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,768,65536,0.08156889014773898
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3072,3584,0.009099555512269339
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,2048,6144,0.009006222089131674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,16384,2048,0.03155288762516446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,10240,10240,0.0478942228688134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1536,2560,0.010934222075674268
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2048,4096,0.009036444127559662
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,6144,64,0.00778399987353219
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1024,6144,0.06854844755596586
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,16384,256,0.006732444382376141
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,3584,256,0.005369777894682354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,5120,3584,0.01459022197458479
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,512,768,0.003168888803985384
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,1536,4096,0.026893332600593567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,8192,6144,0.029622223642137315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2048,3072,0.007618666523032718
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,4096,8192,0.016024889217482675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,64,7168,0.015328889091809591
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,128,16384,0.07734222544564141
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,6144,1024,0.005939555664857228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,10240,1024,0.009310222334331935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2560,2560,0.005727111051479976
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,12288,128,0.006272888845867581
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,256,65536,0.08684000041749741
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,8192,512,0.005961777849329843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,512,65536,0.027836445305082533
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,12288,5120,0.01902755598227183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,1024,7168,0.014202665951516895
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,2560,1536,0.01572800013754103
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,65536,1024,0.4633644421895345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,10240,12288,0.029300444655948218
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,64,3072,0.004970666848950916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3584,6144,0.01313511116637124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,12288,2560,0.01201866649919086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,512,5120,0.008767999708652496
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,3072,1536,0.016153777639071148
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1536,65536,0.14587644735972086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,1536,5120,0.017145777742067974
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,128,1536,0.006627555522653792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,12288,12288,0.04576711191071404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,3584,4096,0.07986844248241849
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,1024,6144,0.02628177735540602
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,16384,5120,0.03461866577466329
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1536,64,0.0059262220230367445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,512,65536,0.07187200254864164
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,16384,3584,0.01759199963675605
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,768,7168,0.00960355583164427
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,1024,1024,0.007631999750932057
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1536,768,0.0035217776894569397
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,64,7168,0.009701333112186855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,256,7168,0.01720711092154185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,64,2560,0.016152888536453247
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,768,2048,0.006244444598754247
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,1536,6144,0.031713777118259005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1024,10240,0.016178665889634024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,1536,3072,0.016196444630622864
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,65536,5120,0.05779288874732124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,768,6144,0.009389333426952362
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,3584,256,0.005559111220969095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,12288,12288,0.03711466656790839
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,65536,64,0.09367999765608047
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,7168,6144,0.054449776808420815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,768,2560,0.004927111168702443
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,12288,256,0.006375999914275275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,128,1024,0.006033777776691649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,6144,65536,0.07711644305123223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,128,12288,0.005520888914664586
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,256,4096,0.015140444040298462
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2048,16384,0.020264888803164165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,12288,16384,2.2207120259602866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1536,1536,0.00905511114332411
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,512,6144,0.017456889152526855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2048,64,0.019810666640599568
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,6144,65536,0.23613421122233072
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,16384,8192,0.025616889198621113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3072,3072,0.01733688844574822
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,12288,12288,0.08540888627370198
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,7168,10240,0.025759999950726826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2560,512,0.002586666701568498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,65536,3584,0.0463048881954617
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,5120,768,0.008143111235565608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,12288,8192,0.023936889237827722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,3072,256,0.015807110402319167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,16384,512,0.007089777953094906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,256,32,0.007961778177155389
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,3584,16384,0.033547557062572904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,12288,256,0.006269333263238271
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,1536,128,0.0052373334765434265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,128,32,0.005632888939645555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,4096,12288,0.030329777134789362
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3072,10240,0.016137777103318107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,768,2560,0.014783110883500842
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1024,1024,0.0032355555643637977
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,10240,10240,0.05549066596561008
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,4096,8192,0.0481697784529792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,512,3072,0.01477688882086012
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,5120,8192,0.044863998889923096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,2560,1024,0.020530665914217632
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,7168,3584,0.01108266661564509
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3072,5120,0.01567555632856157
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,12288,8192,0.03902933332655165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,4096,1024,0.015159999330838522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,768,8192,0.011060444845093621
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1024,512,0.0029324444217814338
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,512,65536,0.04483555422888862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,512,3584,0.03403822249836392
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2048,256,0.005550222264395819
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,8192,10240,0.018312000566058688
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,256,1024,0.012446222205956778
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,8192,3072,0.008667555948098501
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3072,64,0.005550222264395819
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,10240,3584,0.03444266650411818
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,8192,512,0.01793333391348521
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,128,768,0.0028568889117903183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,3584,65536,0.06559999783833821
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3072,10240,0.01603200038274129
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,1024,256,0.005247999810510212
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,32,1536,0.006612444503439798
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,4096,64,0.013471999930010902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,512,65536,0.018744000130229525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2560,8192,0.013635555903116861
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,256,8192,0.045941332976023354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,8192,5120,0.030993776188956365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,6144,1536,0.00664444433318244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,5120,12288,0.036474668317370944
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,512,12288,0.013069333301650153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,256,64,0.0021911110315057966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,5120,256,0.016528000434239704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,6144,65536,0.11920532915327285
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,1536,5120,0.01796444422668881
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,128,3584,0.007600000335110559
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,5120,3072,0.007287110719415877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2048,65536,0.04564088914129469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,10240,3072,0.034083555142084755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,3072,768,0.005606222069925732
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1536,128,0.008192000289758047
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,1024,512,0.005622222191757626
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2048,3072,0.008425777984990014
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,12288,256,0.01745244363943736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2560,4096,0.009277332988050249
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,5120,6144,0.013354666531085968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,6144,1024,0.024549333585633173
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,16384,1536,0.010819555156760745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2560,512,0.005920889063013925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,12288,3584,0.028208888239330713
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1536,1536,0.005048888838953442
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,5120,7168,0.06538489129808214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2560,2048,0.0057297779454125305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,3584,4096,0.01512977812025282
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,5120,1024,0.006234666539563074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,768,4096,0.005913777897755305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,12288,32,0.005920889063013925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,256,6144,0.01650311052799225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,2048,6144,0.011755555868148804
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,6144,256,0.00700800038046307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,65536,8192,0.18542310926649305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,7168,65536,0.4037333329518636
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,6144,5120,0.011412444214026133
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,2560,16384,0.0817413330078125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,256,768,0.027875555886162653
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,6144,6144,0.027843554814656574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,12288,12288,0.03088888857099745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3584,64,0.005615111026499007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,128,65536,0.015634665886561077
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,2560,5120,0.02787999974356757
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,512,4096,0.007784888976150089
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,8192,8192,0.02349688940578037
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,1024,7168,0.009744000103738572
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1536,5120,0.009094222552246517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,4096,32,0.002887111157178879
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,6144,32,0.005618666609128316
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,65536,1024,0.0628933310508728
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,1024,512,0.013166222307417126
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,7168,3072,0.026864889595243666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3584,32,0.0022444443570242985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,2048,3584,0.025824889540672302
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,5120,2048,0.007727999654081132
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,2048,2560,0.009791110952695211
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,12288,256,0.0032586666444937387
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,2560,768,0.006365333166387346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,65536,3584,0.04231199953291151
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,65536,5120,0.051391111479865186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,6144,10240,0.01866222255759769
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,5120,4096,0.023012444376945496
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,256,2560,0.0059368887709246735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1024,6144,0.010739555789364709
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,12288,5120,0.023056000471115112
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,64,8192,0.01075111081202825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,768,2560,0.014848889576064216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,16384,256,0.034479998879962497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,5120,8192,0.030889779329299927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1024,4096,0.01465777721669939
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,128,6144,0.008346666892369589
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,512,8192,0.019254222512245178
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,16384,12288,0.0999644464916653
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,1536,6144,0.029770665698581274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,5120,3072,0.007711110843552484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,3072,8192,0.03271199928389655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,3584,12288,0.029380444023344252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,6144,16384,0.023753777146339417
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,8192,6144,0.29998932944403756
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,1536,128,0.007048000064161088
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,4096,7168,0.1989662249883016
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,65536,1024,0.0652186671892802
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,512,16384,0.02123466630776723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,768,4096,0.019688000281651814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,1024,512,0.014105778601434497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,32,7168,0.004909333255555895
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,1024,8192,0.019215111931165058
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,768,2048,0.0042133331298828125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,12288,16384,0.04536177714665731
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,2560,4096,0.021669333179791767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,32,3584,0.004231111043029361
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,32,768,0.002941333378354708
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,3072,10240,0.013416888813177744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2048,10240,0.023039999935362074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,4096,4096,0.027123555541038513
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,16384,5120,1.1870187123616536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,12288,1024,0.012607999973826937
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,5120,1536,0.016537777251667447
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,10240,1536,0.008389333056079017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,16384,4096,0.04636177751753065
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,10240,7168,0.020029332902696397
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,128,7168,0.05771999888949924
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,3584,3072,0.007350222104125553
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,2048,2048,0.017934223016103108
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,7168,3584,0.023636443747414485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3584,7168,0.02141422198878394
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,65536,3584,0.2885813448164198
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,3072,1024,0.007376888559924231
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,64,10240,0.008760000268618265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,7168,1536,0.0181048893266254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,128,2048,0.006684444430801604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,6144,1536,0.04960711134804619
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,256,32,0.0019271110908852683
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,1536,6144,0.033024887243906655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,3072,2560,0.007366221812036302
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3072,16384,0.030743112166722614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,7168,2560,0.010013333625263637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3072,3584,0.00794666674402025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,2560,3072,0.007368000017272101
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,6144,2560,0.008371555142932469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,5120,768,0.017628444565667045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,6144,7168,0.04306311077541775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,32,128,0.0022151112142536375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,4096,3584,0.007600000335110559
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,1536,8192,0.010727110836240979
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,7168,7168,0.04564177658822802
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,64,5120,0.004957333207130432
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,7168,768,0.01872177753183577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,16384,2560,0.013352000051074557
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,256,10240,0.02036977807680766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1536,64,0.009395555489593083
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,7168,10240,0.020289777053727042
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,10240,2560,0.010760000182522668
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,4096,1536,0.005693333430422678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,2560,4096,0.019838222199016146
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,768,1536,0.017482666505707633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,4096,4096,0.020465777979956735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,65536,2560,0.07975288894441393
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,8192,10240,0.029331554969151814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,65536,256,0.014178666803571912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,6144,8192,0.0328764451874627
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,256,12288,0.005653333332803514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2048,3072,0.06391555733150907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,12288,1536,0.008616888688670265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,10240,1536,0.007703999678293864
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,10240,1536,0.011653333074516721
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,6144,1536,0.006641777853171031
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,64,3072,0.007275555696752336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,512,12288,0.014175110393100314
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,128,5120,0.004990222139490975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,4096,6144,0.012740444805887012
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3072,64,0.005622222191757626
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,32,3072,0.004324444466167026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2048,8192,0.012820444173283048
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,512,12288,0.007518221934636434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,65536,32,0.005295111073387994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,5120,2560,0.007856888903511895
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,6144,16384,0.07870311207241483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,1536,3072,0.007359111474619971
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,768,1024,0.014279999666743808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,32,1024,0.00351555562681622
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,6144,512,0.005669333454635408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2560,256,0.005579555614127054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,32,5120,0.008919999831252629
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,65536,7168,0.4593200153774685
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,8192,4096,0.5017155541314019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,16384,128,0.022364444202846948
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1536,6144,0.01587199999226464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,512,1024,0.014560000763999091
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,512,256,0.006435555716355641
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2048,8192,0.01238222254647149
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2048,65536,0.06436444653405084
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,10240,7168,0.023471110396915015
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3072,64,0.0025519999778932994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3584,512,0.00665511108107037
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,10240,128,0.015499555402331881
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,6144,10240,0.018746667438083224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,5120,3072,0.02474844455718994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,12288,1536,0.010460444622569613
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,256,10240,0.011765333513418833
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,3584,6144,0.0618399977684021
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,12288,1536,0.01588888963063558
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,2048,128,0.015444444285498725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,64,7168,0.01070844464831882
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,128,2560,0.013311111264758639
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,3072,4096,0.018562666243977018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,10240,8192,0.03742400142881606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,8192,64,0.0022168887986077205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2048,16384,0.14674754937489828
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,256,4096,0.00589155571328269
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3584,256,0.006616888774765863
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,6144,128,0.0021395554973019496
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,5120,1024,0.007308444215191736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,1536,256,0.01511111193233066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,1024,5120,0.008993777963850234
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,256,64,0.005941333456171884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,7168,768,0.018601778480741713
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3584,8192,0.01296622223324246
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,128,1536,0.003913777983850903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,4096,16384,0.07333066728379992
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,7168,65536,0.13701777988009983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,10240,32,0.004967999955018361
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,5120,4096,0.011688888900809817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,16384,65536,0.16692622502644858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,1024,7168,0.010416000253624385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,2048,5120,0.025238222546047632
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,3584,1024,0.017681777477264404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,65536,2560,0.028492444091373022
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,1024,8192,0.010819555156760745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,65536,128,0.012448888685968188
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,2048,3584,0.009020444419648912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,65536,10240,5.163720872667101
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,1024,512,0.020340444313155282
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,5120,512,0.01687111126052009
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,128,768,0.01519377695189582
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,65536,12288,0.09492977460225423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,65536,16384,0.7504942152235242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,10240,7168,0.03262844350602891
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,768,65536,0.16374844974941677
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,16384,12288,0.7746231291029188
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,4096,128,0.016536000702116225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,7168,768,0.010797333386209277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,6144,4096,0.012832889126406776
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,512,10240,0.05610399776034885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,768,7168,0.01685599982738495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,16384,32,0.0025173332542181015
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,12288,128,0.0028719999310043124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,2560,128,0.009319999979601966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1536,768,0.005639111002286275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,5120,1024,0.007374222079912822
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,1536,16384,0.029061334000693426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1024,1024,0.007268444531493717
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,10240,10240,0.028957333829667833
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,12288,8192,0.1444524394141303
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,8192,65536,0.13426933023664686
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,512,12288,0.02442933287885454
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,10240,65536,0.40434932708740234
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,4096,256,0.0052666668262746595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,2560,128,0.02032888929049174
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,8192,5120,0.017246221502621967
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3072,3584,0.009872888525327047
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,7168,7168,0.028142223755518597
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,1024,2048,0.014800889624489678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,5120,768,0.005648889061477449
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2048,3072,0.007994666695594788
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,12288,8192,0.035649776458740234
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,4096,1024,0.005922666854328579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,1536,5120,0.008325333396593729
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2048,2048,0.009498666558000777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,10240,2560,0.01180355581972334
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,768,16384,0.016509332590632968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,8192,12288,0.08474133412043254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2048,65536,0.06721778048409356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,7168,16384,0.37483821974860293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,256,3584,0.03965333435270522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,768,2048,0.0069982219073507525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,5120,256,0.005550222264395819
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,4096,5120,0.010679110884666443
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,128,3584,0.016173332929611206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,128,3072,0.014135110709402295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,12288,4096,0.017436444759368896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,5120,768,0.00868800034125646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,7168,2560,0.007087111473083496
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,5120,3072,0.08099822203318278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2048,768,0.0032239999208185407
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,16384,2048,0.03446400165557861
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,12288,6144,0.22772089640299478
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,768,2560,0.0069644442862934535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3072,10240,0.023030221462249756
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,8192,16384,1.7988026936848958
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,65536,16384,0.3604542149437799
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,10240,512,0.003933333274390963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,5120,6144,0.04240533378389147
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,768,3584,0.016606221596399944
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,768,7168,0.011113777756690979
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,65536,1024,0.026755554808510676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,32,768,0.0029084444459941653
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1536,4096,0.008384889198674096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,1536,2048,0.016727111405796476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,12288,64,0.005271111097600725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,2048,768,0.005647999958859549
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2048,12288,0.01976711054642995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1024,5120,0.007675555845101674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,4096,4096,0.02030933399995168
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,7168,3072,0.010264889233642155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,4096,65536,0.3044968975914849
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2048,7168,0.01185155577129788
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,7168,3584,0.008007110820876228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,1024,16384,0.032396445671717324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,512,6144,0.00600000015563435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,65536,2560,0.08557599782943726
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,128,4096,0.0052728888889153796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,128,12288,0.013527111046844058
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,2048,8192,0.010446222292052375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,7168,16384,0.07984444167878893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,10240,768,0.011509333219793109
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3584,3072,0.010215111076831818
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,768,2048,0.004484444442722532
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,512,3584,0.017941332525677152
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1024,128,0.0021955555097924336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,128,16384,0.005621333503060871
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,12288,256,0.0064897777305708984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,10240,2048,0.00998488896422916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,256,6144,0.0052915554907586836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,8192,10240,0.029975109630160864
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,64,2560,0.0073520003093613525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,1536,768,0.00563466673096021
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,4096,12288,0.06141333447562324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,32,128,0.00563822231358952
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,1024,7168,0.018930666976504855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1024,8192,0.011075555450386472
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,32,10240,0.012186666329701742
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2560,16384,0.01793511046303643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1536,128,0.005972444597217772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,2560,768,0.014852444330851236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3584,32,0.002624888800912433
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,10240,16384,0.2294853263431125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,1536,4096,0.05591111050711738
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,3072,65536,0.08518489201863606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,256,256,0.005720888988839255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3584,65536,0.07506222195095487
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,2560,2560,0.006958222223652734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,1536,3584,0.024988444315062627
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,1024,65536,0.07469689183764987
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3584,32,0.005931555396980709
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3072,10240,0.015645333462291293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,65536,8192,0.09365955326292251
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,3072,2048,0.0069288888739215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3584,256,0.006335999816656113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,12288,6144,0.4711315366956923
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,6144,2560,0.01144266708029641
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2560,1024,0.00618488879667388
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,768,65536,1.078153822157118
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,32,6144,0.009719111025333405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,768,2048,0.00630666646692488
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,5120,2048,0.00999111102686988
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,65536,256,0.026547556122144062
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,10240,7168,0.040241777896881104
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,12288,10240,0.04550310969352722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,4096,1024,0.006043555421961679
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,5120,1536,0.006314666734801398
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1024,4096,0.008373333348168267
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,65536,256,0.009930666950013902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,65536,768,0.030863109562132094
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,12288,8192,0.026259554757012263
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,512,7168,0.018516444497638278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,12288,4096,0.026844445202085707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,65536,1536,0.023219555616378784
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3584,128,0.002209777840309673
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,7168,10240,0.020568889048364427
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,8192,10240,0.12851733631557888
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,1024,6144,0.06998311148749457
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,10240,1024,0.019983111156357657
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,512,8192,0.008640000389681922
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,5120,7168,0.027879112296634253
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,10240,12288,0.04182755615976122
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,7168,10240,0.018160889546076458
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2560,3072,0.008859555754396651
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,768,6144,0.009378666679064432
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,10240,10240,0.03563377923435635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,10240,7168,0.03324088785383437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,1536,10240,0.01185955521133211
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2560,512,0.007106666763623555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,8192,5120,0.011744000017642975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,8192,4096,0.07032266590330336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,3072,8192,0.0970311098628574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2048,1024,0.003588444242874781
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,512,7168,0.018584888842370775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,10240,2048,0.011187555889288584
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,5120,12288,0.44812891218397355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,10240,128,0.0028844444702068963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,1024,16384,0.031367109881507024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,6144,3584,0.029712888929578993
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,7168,7168,0.014382221632533602
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,512,3072,0.008808000220192803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,128,65536,0.06153244442409939
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,1024,12288,0.019332443674405415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,6144,3584,0.013448889056841532
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,12288,768,0.007284444239404466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,2048,2048,0.019951999187469482
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,768,256,0.014865777558750577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,768,32,0.0022808888720141514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,64,1536,0.008175111479229396
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1536,12288,0.018509333332379658
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,1024,7168,0.010586666564146677
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,65536,256,0.09873511393864949
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,7168,2048,0.008656889200210571
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,8192,1024,0.006280888699822956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,4096,7168,0.015303999185562134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1536,256,0.002858666703104973
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,16384,6144,0.05177866750293308
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,16384,7168,0.036066667901145086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,3072,1024,0.017370666066805523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,2048,2048,0.007720888488822513
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,5120,128,0.0034862222770849862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,32,8192,0.008249777886602614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,512,2048,0.01626311077011956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,512,32,0.0052888890107472735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,1536,3584,0.007600000335110559
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,6144,32,0.0026142222599850762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,6144,16384,0.31296179029676646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,4096,6144,0.2074071168899536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,12288,8192,0.04810400141610039
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,65536,10240,0.1058808896276686
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,3072,10240,0.016030222177505493
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3072,6144,0.09466666645473903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,10240,2560,0.012701333396964602
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,3072,2048,0.007315555380450354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1536,768,0.01072266697883606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,768,64,0.005333333379692501
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1024,512,0.0029786665820413162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,5120,65536,0.07169689072502984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,2560,2048,0.007643555601437886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,16384,8192,0.03092355529467265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,8192,10240,0.018231110440360177
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,128,2048,0.004598222259018156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1024,10240,0.009759110709031424
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,65536,8192,0.11263288391960992
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3584,1024,0.006020444548792309
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,4096,65536,0.06791022088792589
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,256,5120,0.01980799933274587
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,256,8192,0.005583111196756363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,10240,65536,0.12981066438886854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,7168,3584,0.010558221903112201
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,16384,6144,0.030432889858881634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2560,64,0.0023022221608294383
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,4096,7168,0.23976977666219076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1024,512,0.0138568886452251
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,7168,2048,0.10417955451541477
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,2560,65536,0.06444266769621107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,4096,5120,0.02348711093266805
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,16384,1024,0.009187555147541894
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,768,768,0.015978667471143935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2560,2560,0.007368000017272101
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,12288,65536,0.6493706703186035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,128,128,0.011794666449228922
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1536,32,0.00254488881263468
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,12288,16384,0.08039022154278226
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,64,128,0.0058613332609335584
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,3072,10240,0.04161244299676683
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,16384,6144,0.0637271139356825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2560,128,0.005583999885453119
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,4096,5120,0.013927999469969006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,2048,6144,0.02703911066055298
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,2048,3584,0.007638221813572778
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,65536,512,0.01329422245422999
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,65536,8192,1.990655051337348
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,6144,65536,0.8360035684373645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,1024,1024,0.020908445119857788
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,7168,1024,0.009370666411187913
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,768,5120,0.014504889647165934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,4096,1536,0.03232266505559286
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2560,12288,0.016894222961531747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,256,6144,0.00941422250535753
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2048,32,0.003964444415436851
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,128,7168,0.010072888599501716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,768,2560,0.004964444372389051
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,16384,2048,0.011252444651391772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,32,10240,0.005695111221737332
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,4096,7168,0.038021332687801786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,3072,1536,0.0074622225430276655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,12288,6144,0.01847199930085076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,10240,4096,0.031834665271970965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,6144,4096,0.01834844383928511
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,128,1024,0.012721777790122561
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3584,64,0.0023697778168651792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2048,3072,0.005819555371999741
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,128,128,0.002159111201763153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,8192,128,0.005591111050711737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,12288,4096,0.15294399526384142
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,6144,512,0.008188444707128737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,10240,6144,0.017789334058761597
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,1024,8192,0.021980444590250652
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1024,256,0.002481777841846148
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,8192,5120,0.014442667365074158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,6144,10240,0.07682400279574923
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,65536,2048,0.04420888755056593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,64,512,0.005594666633341048
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,4096,4096,0.007991111112965478
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,64,256,0.005323555734422471
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,3072,5120,0.009016888837019602
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,5120,2560,0.011039111349317761
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,6144,7168,0.04935377836227417
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,65536,2048,0.0270115551021364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2560,256,0.0026737778551048706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,64,4096,0.011434666812419891
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,4096,65536,0.06712977753745185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,256,8192,0.018924444913864136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,65536,4096,0.09712177515029907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1024,256,0.0025235555238193935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,128,16384,0.01588000026014116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,10240,12288,0.05430488785107931
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,12288,4096,0.022429333792792425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2560,128,0.009477333062224919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,16384,6144,0.024167999625205994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,3072,4096,0.020604444874657523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2048,8192,0.02553333342075348
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,1024,768,0.005692444327804778
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1024,16384,0.013049778011110095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,16384,2560,0.013632000320487551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1536,6144,0.015947555502255756
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2048,12288,0.011877333124478659
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,12288,4096,0.021066667305098638
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2560,5120,0.00886933339966668
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,12288,2560,0.011415999796655444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,4096,128,0.005582222094138463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1024,1536,0.004825777891609403
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,1536,6144,0.009009777671760982
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,256,16384,0.029595557186338637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,12288,2048,0.011758222348160215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,65536,256,0.007259555160999298
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,8192,3072,0.03571733170085483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,4096,10240,0.049560887946022876
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,10240,32,0.002534222271707323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,12288,10240,0.04282222191492716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,768,65536,0.11542133490244548
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,12288,64,0.008752000000741746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,4096,5120,0.01072533345884747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,3072,512,0.014150222142537435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,6144,1024,0.003842666745185852
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,65536,1536,0.01998222205373976
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1536,128,0.005555555638339784
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,12288,512,0.004060444318585926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,16384,10240,0.05356533328692118
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2560,6144,0.01312622262371911
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,768,2560,0.0069475554757648045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,768,128,0.017242666747834947
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,7168,3584,0.02641777859793769
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,128,3584,0.01441511180665758
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2048,6144,0.009575111170609793
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,8192,65536,0.13110844294230142
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,16384,5120,0.4691040250990126
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,256,3072,0.007439999944633908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,5120,256,0.007692444655630324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,512,5120,0.03539377782079909
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,256,768,0.007679111427730984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1536,2048,0.006502222269773483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,128,2048,0.013145777914259167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2048,1536,0.006625777731339137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,512,1536,0.006295111030340195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,128,3072,0.014803555276658801
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,8192,65536,0.25958577791849774
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,65536,512,0.013090666797426013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,128,16384,0.00582311095462905
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,2048,2048,0.01551999979548984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,512,2048,0.014120888378885059
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,3072,8192,0.017295110556814406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,512,12288,0.026701332794295415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,512,2560,0.014463111758232117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,8192,5120,0.03630222214592828
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,1024,2048,0.020819556381967332
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,128,65536,0.12754578060574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,6144,2560,0.04424711068471273
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1536,2560,0.004936888813972473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,16384,1536,0.010477333433098264
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,12288,10240,0.030116445488399927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,256,2560,0.014501333236694336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,3584,3584,0.01793155570824941
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,3072,5120,0.03582933213975694
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,3072,12288,0.030771556827757094
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,256,2560,0.006643555644485686
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,5120,512,0.005648889061477449
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,256,4096,0.016960889101028442
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,64,65536,0.045498665836122304
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,8192,8192,0.03612000081274245
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,5120,128,0.005617777920431561
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,128,4096,0.008022222254011366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,1536,65536,0.1991342173682319
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,65536,3584,0.10430577728483413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,8192,128,0.0069982219073507525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,5120,1536,0.004905777672926585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,6144,1024,0.006237333433495627
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,512,16384,0.03506133291456435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,768,2048,0.017779555585649278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,1536,4096,0.016932444439993966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,128,32,0.0019315555691719055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,256,12288,0.026022222306993272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,12288,16384,0.05510400070084465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,3072,12288,0.040039112170537315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3584,10240,0.07658577627605863
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,5120,256,0.016523554921150208
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,65536,256,0.030006223254733618
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,512,1536,0.00408266650305854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,1536,4096,0.016963554753197562
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3072,6144,0.011952000359694162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,7168,128,0.0058373332851462895
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,1024,1024,0.006014222072230444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3072,3584,0.008055999875068665
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,4096,3584,0.011399110986126794
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2560,4096,0.009010666774378883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,8192,5120,0.014353776971499125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,32,2560,0.0069982219073507525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3584,7168,0.012101333174440594
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,12288,768,0.09142133262422349
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,768,768,0.030322667625215318
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,16384,10240,0.05873955620659722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,6144,3584,0.008349333372380998
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2560,5120,0.009470221896966299
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,4096,4096,0.009925333162148794
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,3584,8192,0.04675377739800347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,8192,768,0.009485333330101436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,6144,8192,0.022473777333895367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,12288,512,0.006631111105283101
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1024,10240,0.023014222582181294
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,768,10240,0.008669333325492011
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,16384,2560,0.012180444267061023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,7168,12288,0.06428177489174737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,7168,512,0.00599644457300504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,3584,4096,0.031735112269719444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,2560,10240,0.04335644510057238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,32,2048,0.004809777769777509
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,64,768,0.0028773333049482773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1024,1536,0.004621333132187526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,256,64,0.002191999927163124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,4096,2048,0.03335288829273648
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,128,1024,0.011843555503421359
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,16384,128,0.006583111153708563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,1536,3584,0.007681777907742395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,65536,768,0.07290310992134942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,2048,3072,0.023392000132136877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1024,65536,0.054024888409508594
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,2048,1536,0.006304889089531369
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,512,5120,0.00905511114332411
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,256,512,0.014222222897741528
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3072,4096,0.008797333472304875
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,128,1024,0.015493333339691162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1536,12288,0.03014311194419861
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,12288,16384,0.04526133338610331
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,4096,4096,0.012111110819710625
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2048,7168,0.013443555268976422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,512,256,0.005858666780922148
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1024,512,0.002847111059559716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,4096,8192,0.016040000650617812
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,768,10240,0.009452444811662039
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,6144,6144,0.01980799933274587
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,1536,5120,0.008435555630260045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,32,2048,0.00850311087237464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,32,3584,0.010693333215183683
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,256,65536,0.08230044444402058
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,768,10240,0.024135111106766596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,12288,12288,0.050847109821107656
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,32,4096,0.022470222579108343
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,2560,1536,0.01791111131509145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2048,4096,0.014831999937693277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,768,1536,0.013978666729397245
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,16384,1536,0.02951644526587592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2048,256,0.002556444456179937
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,512,1024,0.013764444324705334
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,2560,2560,0.01684799955950843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,3072,65536,0.06469866964552137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,12288,4096,0.06388711267047457
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1024,5120,0.028853333658642236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,4096,1536,0.006256000035338932
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,4096,3584,0.008676444490750631
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,1024,3072,0.015248000621795654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,1536,5120,0.02440533373090956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,3072,65536,0.1653777758280436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,10240,8192,0.24611910184224448
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,8192,7168,0.023333332604832117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,5120,3072,0.011805333197116852
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,6144,6144,0.013388444152143268
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,7168,4096,0.011081777513027191
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,16384,65536,0.16525510946909586
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,4096,512,0.005340444544951121
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,256,7168,0.017687999539905123
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,2048,1024,0.005615111026499007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,256,128,0.005594666633341048
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,4096,512,0.02179466684659322
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,128,65536,0.10887377791934544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,10240,12288,0.03391022152370877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,128,128,0.005902222047249476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,4096,10240,0.020445333586798776
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,2048,7168,0.05827466646830241
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,7168,10240,0.01792533364560869
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,1536,65536,0.06451821989483304
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2560,7168,0.013377777404255338
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,512,256,0.014161777165200977
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,512,7168,0.010008888940016428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,8192,512,0.0059279998143513995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,64,8192,0.005592000153329637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3584,64,0.006284444282452266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,1536,5120,0.01754666699303521
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,512,768,0.0148026661740409
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2048,2048,0.006566222343179915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2560,512,0.006270222365856171
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,4096,12288,0.020289777053727042
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,2048,3072,0.016586666305859882
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,5120,8192,0.03018222252527873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1536,2560,0.004898666507667965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,65536,2048,0.023231999741660223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1024,7168,0.009956444303194681
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,4096,5120,0.037286221981048584
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,256,65536,0.05204622281922234
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,6144,3072,0.009482666850090027
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,768,10240,0.01205244412024816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,10240,8192,0.029824889368481104
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,12288,3584,0.016605334149466622
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,65536,512,0.05922933419545492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,16384,128,0.006521777974234686
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,2048,3072,0.016167999969588388
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,5120,5120,0.01604977746804555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,128,16384,0.015463110473420886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,4096,5120,0.053568889697392784
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1536,512,0.005566222386227713
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,32,256,0.002297777682542801
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,768,1024,0.0059004442559348206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,512,64,0.005612444546487596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,512,2048,0.011770666473441653
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,10240,256,0.006163555714819167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,65536,5120,0.1322702169418335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,3584,256,0.0052933332820733385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2560,5120,0.0350471105840471
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,256,4096,0.007649777664078607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3072,7168,0.010764444039927589
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,1024,3584,0.018370666437678866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,1024,4096,0.0221368885702557
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,10240,128,0.006657777975002925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,4096,4096,0.00907911111911138
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,2560,5120,0.07992266946368747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,128,5120,0.016833777228991192
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2048,16384,0.019708444674809773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,5120,10240,0.03334755698839823
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,7168,256,0.002507555608948072
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,5120,64,0.0025946667624844448
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,12288,6144,0.024899555577172175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,128,768,0.007324444750944774
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,7168,3584,0.010210666391584609
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,16384,2560,0.03357244531313578
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,2048,1536,0.014861333701345654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1024,2048,0.00519822206762102
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,4096,6144,0.02698755595419142
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,5120,768,0.006007110906971826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,7168,128,0.005590222362014983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,32,64,0.0020942222326993942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,3584,256,0.030646221505271062
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,32,512,0.002838222309947014
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,4096,3584,0.018400889303949144
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,65536,12288,0.18485332859887016
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,5120,6144,0.013163555827405719
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,10240,10240,0.02566577825281355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,5120,65536,0.14017332924736872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,10240,16384,0.05353066656324598
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,12288,2560,0.014041778114106921
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,512,8192,0.007294221884674496
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2048,6144,0.009825777676370408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,6144,2560,0.007698666718271043
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1024,6144,0.010083555347389644
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,8192,2048,0.021593777669800654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,6144,7168,0.01552088889810774
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,32,6144,0.00905600024594201
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2560,1024,0.016159110599093966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2560,64,0.006026666611433029
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,8192,512,0.017232000827789307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3584,128,0.0024213333510690266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,32,8192,0.010432889064153036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,768,8192,0.008929777476522658
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,65536,16384,0.20852088928222656
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,16384,8192,0.026288888520664636
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,4096,8192,0.027247111002604168
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,7168,7168,0.0426862206723955
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,6144,768,0.017975111802419026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,6144,1536,0.012448000411192576
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,3072,1536,0.017937777770890128
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,4096,7168,0.014230221509933472
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,8192,256,0.005609777652555042
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,64,5120,0.01311288856797748
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,128,6144,0.01942488882276747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,3072,128,0.014837332897716098
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,768,4096,0.008276444342401292
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2048,128,0.005241777747869492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2048,3584,0.008358221915033128
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,768,768,0.005983110931184556
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3072,16384,0.019658666517999437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,7168,5120,0.03714933329158359
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,8192,8192,0.01665244499842326
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,65536,256,0.009684444301658207
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2560,3584,0.007306666837798224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,3072,128,0.0052791109515561005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3584,1024,0.0037120001183615792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,4096,16384,0.07141155666775174
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,7168,4096,0.012123555772834353
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,10240,2048,0.02735022207101186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,16384,16384,0.10778755611843532
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,3072,3072,0.0999022192425198
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,1536,3584,0.00869511150651508
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,10240,2048,0.020615110794703167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,256,12288,0.011136000355084738
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,768,768,0.005587555468082428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,6144,6144,0.014936889211336771
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,5120,8192,0.04443555408053928
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3072,256,0.005933333188295364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,5120,256,0.01684711045689053
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2560,768,0.017125333348910015
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,5120,128,0.014833778142929077
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2048,16384,0.017456000049908955
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,512,10240,0.02369155486424764
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2048,2048,0.006599999964237213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,256,10240,0.005981333139869902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,8192,768,0.007623111208279927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,6144,1536,0.08378666639328003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,7168,3584,0.014436443646748861
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1536,512,0.002537777854336633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1536,7168,0.009664889011118147
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,6144,10240,0.023769777682092454
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,7168,6144,0.02715911136733161
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,256,8192,0.06654844681421916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3072,65536,0.10819555653466119
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,65536,4096,0.12007555696699355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,5120,2048,0.017645332548353408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,2048,768,0.005623111294375525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,64,128,0.002151111140847206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,5120,3584,0.0086666668454806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,5120,5120,0.012076444096035428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,768,128,0.014035556051466199
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,3584,2560,0.01687111126052009
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,768,65536,0.08339911037021214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,512,8192,0.006286222073766921
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,7168,32,0.0021262222694026097
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2048,8192,0.011965333587593503
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,256,512,0.005562666803598404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,1536,512,0.01346666696998808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,7168,1024,0.009396444592210982
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,3584,2560,0.006934222247865465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,128,65536,0.0490551127327813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,6144,7168,0.13944978184170193
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,512,65536,0.12259644932217068
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1536,12288,0.012690666649076672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1536,7168,0.01830666595035129
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2048,2048,0.02421599957678053
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,16384,16384,0.11214844385782878
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,32,16384,0.01586577792962392
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,10240,65536,0.1391271087858412
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,2560,8192,0.010718222293588849
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,512,1024,0.005959110955397288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,768,256,0.00527555536892679
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3584,7168,0.029142220815022785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,16384,3584,0.020690666304694284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,32,16384,0.005688888745175467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,2560,12288,0.03542666633923849
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,128,2048,0.006353777729802662
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,128,32,0.0052977779673205475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,768,512,0.0025920000755124623
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1536,7168,0.02015911042690277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,5120,12288,0.5970817671881782
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,768,16384,0.01630577776167128
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2560,3072,0.011140444212489657
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3072,64,0.010785777535703448
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,8192,65536,0.0921635561519199
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,2048,10240,0.028824888997607764
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,7168,3584,0.008091555701361762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3072,6144,0.011492444409264458
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,2048,5120,0.010804444551467896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,768,256,0.005967111223273807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,16384,1536,0.01277599980433782
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,128,6144,0.019511999355422128
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,16384,8192,0.0852924452887641
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,12288,5120,0.019235556324323017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3584,4096,0.009504888620641496
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1024,3584,0.005960000058015187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,2048,4096,0.01015466699997584
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,768,65536,0.04878933231035868
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,6144,7168,0.4045093324449327
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1024,16384,0.04686311218473646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,256,2048,0.009015111459626092
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2560,3584,0.00941955546538035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,512,8192,0.014491554763582019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,6144,2048,0.020662221643659804
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,6144,2560,0.01789155602455139
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,64,3072,0.007256888680987888
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,3072,6144,0.009304888546466827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,7168,2560,0.007332444190979004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,64,10240,0.012096888489193387
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,256,768,0.0069448889957533936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,2560,1536,0.0062773335311147906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,64,1024,0.005658666706747479
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,10240,8192,0.02502577834659153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,1536,10240,0.03421155611673991
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,768,10240,0.011049778097205691
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,10240,4096,0.07051199674606323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1024,8192,0.012134222520722283
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,32,5120,0.005647999958859549
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3584,2048,0.0052177777720822235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,4096,1024,0.008360889222886827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,5120,2560,0.014888000157144336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,256,768,0.0029084444459941653
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1536,32,0.002508444504605399
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,8192,65536,0.15402311748928493
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2048,256,0.01016533292002148
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,64,2560,0.007009777757856581
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,16384,3584,0.18823822339375815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3072,10240,0.015283554792404175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,512,2048,0.004527111020353106
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,768,2560,0.009072889056470659
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,768,4096,0.006909333169460297
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,8192,6144,0.05164177881346809
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2560,32,0.0032426667296224167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,128,8192,0.04694755540953743
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,16384,128,0.0038124442928367188
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,7168,768,0.018631110588709515
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,512,768,0.014878221684032016
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,5120,768,0.04201777776082357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,4096,3584,0.007902222375075022
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2048,6144,0.011403555671374003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,8192,16384,0.028287112712860107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,8192,3072,0.09966133038202922
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,12288,1536,0.03070844544304742
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,12288,128,0.006291555447710886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,6144,1536,0.0069715554515520734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,1536,128,0.01551555593808492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3072,768,0.0046728890803125174
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,7168,3072,0.007845333053006066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,256,1024,0.003192000091075897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,4096,768,0.03394666645261977
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,2048,12288,0.05254844493336148
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,256,256,0.005636444522274866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3584,5120,0.011219555305110084
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,512,6144,0.012108444339699216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,5120,8192,0.014342221948835584
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,128,10240,0.024840000602934096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2048,2560,0.007690666450394525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,6144,4096,0.011869333684444427
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,768,16384,0.01941333380010393
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,16384,16384,0.04560533497068617
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,8192,64,0.003927999900446998
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1024,3584,0.012181333369678922
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,1536,4096,0.008018666671382057
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,5120,256,0.002305777743458748
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2048,5120,0.00870666652917862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2048,10240,0.09520444605085586
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,16384,7168,0.04380888740221659
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3072,1536,0.006346666564544042
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,64,3584,0.004337777694066365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,6144,3072,0.007987555530336168
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,12288,10240,0.5909244219462076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,4096,1024,0.006646222124497096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,6144,7168,0.02308355602953169
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,16384,65536,0.353021330303616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,16384,10240,0.03037422233157688
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3584,2048,0.009576000273227692
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,768,512,0.005616000129116907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,512,1536,0.006654222392373615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,32,512,0.0026151111556424033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,7168,128,0.015032889114485847
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,3072,6144,0.01144444445768992
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,16384,12288,0.036567111810048424
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,16384,1024,0.009291555318567488
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,3072,2560,0.006909333169460297
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3072,7168,0.020462221569485135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,6144,7168,0.031192888816197712
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,8192,5120,0.04210755560133192
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,4096,2048,0.0063031112982167145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,12288,3072,0.021702221698231165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,5120,10240,0.03159733282195197
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3072,2048,0.0069884442620807225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3072,10240,0.06122844749026828
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,512,12288,0.01520088811715444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,128,3072,0.004664888812435998
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,6144,6144,0.016352888610627916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,32,2560,0.009349333743254343
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,12288,3072,0.03240266773435805
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,512,6144,0.00907022257645925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,4096,1024,0.005256888767083486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,256,64,0.0028293333533737394
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,1024,1536,0.029345777299669053
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,2048,65536,0.08219022221035428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,32,6144,0.005245333330498801
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2560,64,0.004549333204825719
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,1536,3584,0.016157334049542744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2560,4096,0.04752622379197014
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,3584,10240,0.026597332623269822
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,12288,4096,0.036066667901145086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,6144,10240,0.42928534083896214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,16384,1536,0.01036888857682546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,3072,10240,0.027082666754722595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,10240,2048,0.01073333372672399
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3072,16384,0.018717333674430847
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,8192,10240,1.0600915484958224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,256,1024,0.017529777354664274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,65536,16384,0.20957956049177381
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,10240,768,0.006587555425034628
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,10240,768,0.042853332228130765
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,768,16384,0.02436711059676276
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,128,256,0.012127111355463663
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,64,1024,0.007663110891977946
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,8192,12288,0.02064622276359134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3072,1536,0.004898666507667965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,4096,5120,0.042605333858066134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,64,256,0.0021333334346612296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,5120,5120,0.031459556685553655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,2560,3072,0.053584890233145825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,7168,12288,0.020026667250527274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1024,256,0.005745777653323279
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,512,16384,0.011571555501884885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,64,256,0.005932444499598608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1024,2560,0.0052728888889153796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,128,2048,0.00490399988161193
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,768,12288,0.01351466692156262
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,16384,128,0.006558222075303395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,3584,128,0.005292444593376583
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,5120,3584,0.14493866761525473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,768,4096,0.008357333640257517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,128,256,0.012440889245933957
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,64,2560,0.004618666652176115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,3072,12288,0.056067552831437856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,12288,1024,0.06454488966200086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,4096,512,0.03106400039460924
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,768,512,0.013532444834709167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,2048,128,0.014496000276671516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,768,1536,0.0038399998512532976
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,256,12288,0.005601777798599667
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,10240,3072,0.014594667487674289
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,10240,256,0.005970666805903117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1024,512,0.0032968889507982465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,10240,3584,0.02696711156103346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1024,2048,0.028836445675955877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,12288,1536,0.04976888828807407
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,256,12288,0.06374755832884046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,65536,768,0.013621333572599622
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1536,16384,0.06723999977111816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,512,10240,0.02276355524857839
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,7168,128,0.0028693332440323303
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,65536,2560,0.029118223322762385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,8192,128,0.01556711064444648
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3584,65536,0.926338619656033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,16384,1024,0.04233066572083367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,8192,256,0.027801778581407335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,1024,1536,0.025472000241279602
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,7168,10240,0.025224000215530396
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,12288,8192,0.04160977734459771
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,2048,6144,0.030093332131703694
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,7168,2560,0.06523111131456164
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,768,16384,0.039275556802749634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,8192,12288,0.02497600018978119
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,10240,1024,0.02720977697107527
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,1024,10240,0.0214675549003813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,8192,5120,0.1306648916668362
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,256,16384,0.029626667499542236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,12288,10240,0.07146222061581083
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,6144,128,0.01588800052801768
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,2048,1536,0.014608888162506951
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,128,3072,0.005635555419656966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1536,7168,0.11184799671173096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2048,6144,0.11024977763493855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,3072,256,0.015817777978049386
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,10240,768,0.007536888950400882
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,16384,8192,0.046352889802720815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,10240,1536,0.008488888541857401
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,1536,16384,0.18294666873084175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,10240,2560,0.015542222393883599
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,256,2048,0.004622222234805425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,768,16384,0.02754755483733283
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,10240,7168,0.23091289732191297
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1024,128,0.0022275555464956495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,128,128,0.006664000037643645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,256,6144,0.009681777821646797
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3072,1536,0.025754666990704004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,128,65536,0.06853422191407946
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,8192,1024,0.02645422187116411
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,4096,3584,0.009204444785912832
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2048,1024,0.006214222146405115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,5120,768,0.007308444215191736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,8192,7168,0.030228445927302044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,8192,256,0.005960000058015187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,5120,64,0.005400889035728242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,16384,12288,0.06220177809397379
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,1024,12288,0.024459555745124817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,768,2560,0.007661333514584436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,1024,3584,0.009741333623727163
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,8192,6144,0.02089333368672265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,16384,12288,0.5502150853474935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,512,4096,0.007687110867765214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,5120,10240,0.020475554797384474
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,512,10240,0.02273599969016181
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,8192,7168,0.014380445082982382
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,2048,12288,0.03796622157096863
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1536,10240,0.010762666662534079
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,6144,65536,0.11845689349704319
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,1024,12288,0.03071199854214986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,256,5120,0.008924444516499838
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,32,2560,0.007495111061467065
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,7168,4096,0.03208355440033807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,16384,128,0.006024888820118374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,512,2560,0.017663111289342243
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,6144,16384,0.04673333300484551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,8192,7168,0.027290667096773785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,4096,256,0.005610666755172942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,7168,10240,0.017704889178276062
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2048,512,0.0028417776856157514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,512,128,0.0022151112142536375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,5120,7168,0.042697777350743614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,64,3072,0.004962666581074397
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,7168,6144,0.016898666818936665
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,768,64,0.0025759999536805684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,65536,512,0.012153777811262341
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3584,3072,0.04766222172313266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1024,65536,0.05732444259855482
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2048,512,0.0028195555011431375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,768,128,0.0052737775776121355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3584,7168,0.013217777841620974
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,12288,8192,0.060765332645840116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,2048,16384,0.20134399996863472
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1536,1536,0.026622222529517278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,10240,4096,0.042635556724336415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,64,128,0.002193777718477779
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3584,32,0.011800000237094032
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,256,7168,0.009720889230569204
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,128,512,0.0026151111556424033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,512,3072,0.007651555869314406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,16384,3072,0.017536888519922893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,12288,10240,0.3776880105336507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,6144,4096,0.01975644462638431
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,512,7168,0.005552000055710475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2048,4096,0.012200000385443369
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,512,16384,0.026787555880016748
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,16384,8192,0.054515557156668774
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,10240,3584,0.027447111076778833
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,64,65536,0.04603111081653171
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,2048,5120,0.017656889226701524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,512,1024,0.013053333593739403
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1024,2048,0.00868888861603207
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,16384,512,0.0069617778062820435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1024,1024,0.006312888943486744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,65536,768,0.015593777100245157
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,2560,2560,0.007664889097213745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,8192,4096,0.011401777466138205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,8192,1024,0.0063777777055899305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,768,768,0.015857777661747403
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,65536,128,0.024488000406159297
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,2560,6144,0.02016799979739719
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,8192,12288,0.08686933252546523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,768,5120,0.05546844336721632
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3584,16384,0.021487111846605938
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,7168,4096,0.1361075507269965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,10240,128,0.006294222341643439
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,6144,4096,0.12455644872453477
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,128,3072,0.013872000078360239
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2560,6144,0.023038221730126276
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,12288,7168,0.04426311122046577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,512,64,0.009729777773221334
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1024,256,0.009688888986905416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,4096,2560,0.006963555183675554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,256,16384,0.025916443930731878
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,7168,768,0.003944000022278892
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,10240,2560,0.06005066633224487
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,32,12288,0.05962844689687093
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1536,3584,0.006241777704821692
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,5120,12288,0.017343110508388944
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,256,3584,0.004641777939266629
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,10240,1024,0.0074933336840735535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,5120,1024,0.02641333308484819
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,7168,4096,0.015200000670221118
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,7168,3072,0.01053244454993142
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,128,1024,0.0029191111938820947
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,2560,768,0.01755466726091173
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,6144,8192,0.03151288959715102
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,768,7168,0.007910221815109253
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,5120,1024,0.015831111205948722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,7168,8192,0.03494933247566223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,8192,128,0.00556355549229516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,65536,10240,0.15985777642991808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,12288,2048,0.013589333328935834
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,512,12288,0.013185777597957186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,5120,256,0.005704888867007361
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,16384,4096,0.040885332557890154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,8192,12288,0.04535733328925239
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2048,3584,0.008096888661384583
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,4096,7168,0.02780088782310486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,6144,7168,0.027600000301996868
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,8192,32,0.0022044444663657085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,32,5120,0.004584889031118817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,12288,16384,0.2579004499647352
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,8192,4096,0.013168888787428537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1024,7168,0.010807111031479306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,3072,10240,0.01536799967288971
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,3072,2560,0.016190222567982145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2048,128,0.005950222412745158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,256,64,0.001998222122589747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,12288,1024,0.007711110843552484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,6144,10240,0.024297777149412367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,512,16384,0.015864888827006023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2048,1024,0.006267555471923616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,8192,256,0.010071111222108206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,10240,768,0.07827733622656928
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1536,3584,0.008293333152929941
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,6144,12288,0.027606222364637587
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,2560,1024,0.014649778604507446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,128,3584,0.06178310844633314
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,7168,6144,0.013588444226317935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,256,4096,0.03003022074699402
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,7168,768,0.03947733177079095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,256,12288,0.062231997648874916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1024,6144,0.009552000297440423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,5120,12288,0.020848888489935134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,32,16384,0.006623111251327727
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,2560,10240,0.027003554834259882
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,3584,7168,0.0696586635377672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,5120,7168,0.021534222695562575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,512,12288,0.01870133313867781
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,2048,65536,0.06335822078916761
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2560,2048,0.006911110960774952
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,512,2560,0.006934222247865465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,3072,512,0.014482667048772177
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,3072,16384,0.021944888763957556
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,8192,65536,2.005420472886827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,128,16384,0.032891554964913264
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,65536,7168,0.10676888624827068
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2048,768,0.0038151111867692736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,32,4096,0.004554666578769684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1024,6144,0.006887110984987683
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,16384,65536,4.5806354946560335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,65536,128,0.01864266726705763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,128,512,0.011004444625642566
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,5120,2560,0.010082667072614035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,6144,2048,0.00944622192117903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,1536,8192,0.010716444088353051
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,10240,64,0.006623111251327727
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,1536,768,0.018568888306617737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,256,2560,0.005219555563396878
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,1024,12288,0.040101332796944514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,128,2560,0.006931555353932911
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,768,1024,0.0070266665683852295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,4096,128,0.014058666096793281
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,256,16384,0.02610844373703003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,7168,32,0.0023093333260880574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,3072,3072,0.03864977757136027
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,32,768,0.006065777606434292
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,12288,12288,0.06493600209554036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,10240,32,0.002895999906791581
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,128,2560,0.0069715554515520734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,128,128,0.011377777490350934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,128,10240,0.012888888518015543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,64,5120,0.009304888546466827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2048,10240,0.015406222807036506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,3584,128,0.010606221854686737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,10240,65536,0.14109244611528185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,256,3072,0.05537511242760552
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,16384,512,0.009335111412737105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,5120,8192,0.01665155589580536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,65536,16384,0.23226488961113823
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,4096,7168,0.013428444663683573
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,3072,12288,0.13913599650065103
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,10240,1024,0.014286221729384528
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,16384,6144,0.06021777788798014
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1024,65536,0.05953688753975762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2048,64,0.005921777751710679
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3584,2048,0.0222053329149882
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,2560,65536,0.08366577492819892
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,1024,65536,1.171335114373101
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,5120,2560,0.06477155288060506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,12288,3584,0.015992000699043274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,5120,65536,0.5403137736850315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,12288,16384,0.05707644753985935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,128,1536,0.006339555399285422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,512,3584,0.007400888535711501
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3072,65536,0.06293777624766032
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,512,4096,0.00797333319981893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,16384,3072,0.01773155563407474
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,10240,2048,0.011160888605647616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,64,7168,0.0052728888889153796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,5120,3584,0.009701333112186855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,128,3072,0.013907555076811047
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,2560,3072,0.01699911057949066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,12288,1536,0.07405777772267659
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,16384,768,0.0663822227054172
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,256,1536,0.0036159998012913596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,12288,5120,0.023408000667889912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,8192,32,0.01827911039193471
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,8192,65536,0.1926844384935167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,2048,16384,0.0661626656850179
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,256,7168,0.01945955554644267
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2560,7168,0.009862222605281407
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2048,1536,0.0038702223036024305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,512,2048,0.00702133360836241
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,2048,65536,0.1944204436408149
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3072,1024,0.00498844434817632
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,6144,7168,0.041767110427220665
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,2560,7168,0.02193244463867611
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3584,768,0.0028568889117903183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,12288,3584,0.14819733301798502
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,6144,2560,0.024068444967269897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2048,8192,0.01421511173248291
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3584,2048,0.005691555639108022
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,12288,3584,0.14047733942667642
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,8192,768,0.0059057776298787855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,65536,256,0.018996444013383653
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,128,8192,0.01826133330663045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,7168,16384,0.025566221939192876
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,5120,12288,0.02120977805720435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,12288,2560,0.013911111487282647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,65536,8192,0.23257690005832246
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2048,10240,0.047276443905300565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,7168,10240,0.01814311080508762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,64,256,0.006706666615274217
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,3584,512,0.005678222411208683
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,256,2048,0.012110222544935016
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,512,7168,0.023708444502618577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,2048,3584,0.017324444320466783
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3072,256,0.014178666803571912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,256,256,0.013204444613721637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2560,5120,0.1024986637963189
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,4096,32,0.006000888844331105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,2048,2048,0.006684444430801604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,65536,16384,18.78930494520399
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,2048,7168,0.010478222535716163
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,768,8192,0.00977422214216656
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,2048,128,0.026100445124838088
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,7168,12288,0.03163911236657037
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,2560,256,0.005542222410440445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,2048,128,0.005281777845488654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,6144,1536,0.023000889354281958
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1024,5120,0.006598222172922558
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,768,1024,0.006248888870080312
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,512,512,0.003399111123548614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,65536,16384,0.14191822210947672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,64,65536,0.012797333300113678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,4096,16384,0.02176711128817664
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1536,5120,0.00905866672595342
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,10240,2048,0.011139555937714048
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,4096,128,0.005629333357016246
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,10240,128,0.015948444604873657
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,65536,5120,0.05656800005171034
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,16384,512,0.009733333355850643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,8192,32,0.003979555434650845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,4096,5120,0.010763555765151978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,512,32,0.0021662221600612006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,5120,16384,0.025038222471872967
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3072,2560,0.006600889066855113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,768,2048,0.012401777837011548
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1536,7168,0.010015111002657149
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,16384,768,0.015609777635998197
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1024,6144,0.009368889033794403
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,512,256,0.002493333278430833
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,6144,3072,0.023616888456874426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,128,2560,0.013818666338920593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,12288,64,0.0021404443929592767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,2048,12288,0.01869066721863217
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,12288,128,0.00630044440428416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,5120,6144,0.024824000067181055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3584,2048,0.007102222078376346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,4096,5120,0.027087999714745417
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,10240,3072,0.015007999208238391
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1536,12288,0.018019555343521964
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,768,1536,0.005145777844720417
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,5120,2048,0.006706666615274217
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,7168,6144,0.03044711218939887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,3072,768,0.005621333503060871
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1536,8192,0.12708266576131186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,3584,1536,0.008775999976529015
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2560,8192,0.01387733303838306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,65536,8192,0.36378222041659886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,256,65536,0.030309332741631403
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,65536,2048,0.12592000431484646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,128,3584,0.025528889563348558
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3072,128,0.003208000005947219
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,768,7168,0.044686221414142184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,768,12288,0.11582844787173802
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,16384,512,0.007328000333574083
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,768,64,0.005576000031497743
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1024,10240,0.008887111312813228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,10240,3584,0.013944889108339945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,10240,768,0.02095111045572493
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,10240,65536,0.3003137906392415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,1536,12288,0.14007200135125056
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,65536,6144,7.160891215006511
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,12288,6144,0.05438489052984449
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,512,6144,0.018886221779717337
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,12288,64,0.0022515555222829184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,7168,6144,0.02791555391417609
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,2048,768,0.017235555582576327
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,1024,1536,0.00591555568906996
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,512,10240,0.021624000536070928
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,768,64,0.006310222049554189
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,7168,1024,0.01921866668595208
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,512,512,0.006090666684839461
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,7168,64,0.002524444419476721
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,6144,4096,0.16170310974121094
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,5120,16384,0.021967111362351313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,10240,3584,0.01367200000418557
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,128,12288,0.00851733320289188
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,64,128,0.0021315556433465746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,8192,4096,0.023518222901556227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,6144,1536,0.027439999911520217
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,7168,3584,0.010018666585286459
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,10240,3072,0.01313333296113544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,512,6144,0.00869599978129069
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,8192,768,0.018199111024538677
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,65536,3584,0.3323040008544922
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,10240,64,0.005896889087226655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,2048,6144,0.018984888990720112
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,16384,64,0.006233777850866318
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,7168,10240,0.11929599444071452
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,128,10240,0.02197066611713833
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,12288,10240,0.037652442852656044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,65536,6144,0.6464088757832845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,10240,2560,0.041514668199751116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,10240,512,0.004880889008442561
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,2048,10240,0.025460445218616064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1536,32,0.007118222614129384
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,3072,10240,0.048473778698179454
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,256,4096,0.012123555772834353
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,512,256,0.019581332802772522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,6144,2048,0.01036444471942054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2048,32,0.005995555470387141
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,6144,1024,0.007337777978844113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3072,64,0.005927111125654644
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,65536,8192,0.19580177466074625
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,768,1024,0.014434667097197639
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,12288,128,0.007286222444640265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,4096,1024,0.005127111242877113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,12288,3584,0.013215111361609565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,768,256,0.002206222257680363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,8192,16384,0.04723199870851305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,10240,8192,0.023057777020666335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,6144,7168,0.020090666082170274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,8192,7168,0.10768266518910725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1536,3072,0.007299555672539606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,10240,512,0.018703111343913607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2560,64,0.00562755556570159
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,512,3584,0.029575109481811523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,5120,7168,0.01882577770286136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,512,3584,0.005921777751710679
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,512,2048,0.006976000136799282
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,1536,8192,0.03056266572740343
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,65536,12288,0.16240889496273464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,512,512,0.005964444329341252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,2560,1024,0.017290666699409485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,32,32,0.005577777822812398
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3072,7168,0.023413333627912734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,3584,6144,0.0222382214334276
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,65536,2048,0.10970933569802178
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,3584,16384,0.0343537794219123
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,2048,768,0.01625688870747884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2048,4096,0.007987555530336168
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,6144,256,0.005305777821275923
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,128,4096,0.008303110798199972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,2048,10240,0.026118222210142348
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,32,8192,0.005548444473081165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,768,3072,0.007251555720965068
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3072,5120,0.011393778026103973
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,2560,1536,0.015797333584891427
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,16384,1024,0.02682222260369195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,512,32,0.001964444501532449
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,8192,2560,0.03356711069742838
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,16384,6144,0.024350222614076402
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2048,6144,0.0096124443742964
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,10240,2560,0.011679111255539788
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,10240,1024,0.008642666869693333
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,5120,12288,0.02453600035773383
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,10240,256,0.007610666255156199
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,256,3072,0.008011555506123437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,10240,64,0.002499555548032125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,7168,4096,0.01074222226937612
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,128,3072,0.007335111498832703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,3072,2560,0.0069271110826068455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2560,65536,0.0524213347170088
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,1024,2560,0.01614755557643043
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,768,8192,0.020976000361972388
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,1536,2560,0.01862844493654039
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,65536,7168,0.13892443974812826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,2048,4096,0.01687022215790219
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,512,128,0.0021724444296624926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,16384,65536,2.3341954549153647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1536,768,0.004013333469629288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1536,2048,0.006599999964237213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,128,16384,0.04020177655749851
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,65536,16384,0.23222043779161242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,6144,4096,0.011972444752852121
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1024,12288,0.017838221457269456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,8192,7168,0.014560888210932413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,1024,65536,0.0758062203725179
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3584,32,0.003510222252872255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2560,2560,0.03167111012670729
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,4096,12288,0.03786044319470724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,2048,1024,0.039860443936453924
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,4096,2560,0.0791964464717441
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,512,256,0.006603555546866522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,768,6144,0.010135110881593492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,32,12288,0.013191111385822296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,3584,512,0.005689777847793367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,256,3584,0.008113777471913232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,6144,3072,0.02288355595535702
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,512,12288,0.03160355488459269
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,512,768,0.007690666450394525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,512,256,0.0052871112194326185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,6144,5120,0.027641778190930683
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3584,12288,0.034796443250444196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3584,768,0.009386666946940953
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,8192,1536,0.022395556171735127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,16384,5120,0.01778133379088508
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,768,6144,0.010139555566840703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,6144,10240,0.019573332534896005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,512,6144,0.005974222388532426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,128,2560,0.04819644490877787
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,768,128,0.013786666923099093
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,5120,2560,0.007294221884674496
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,128,65536,0.1181902223163181
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,2560,4096,0.022100443641344707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,2560,2048,0.006493333313200209
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,3072,12288,0.06034666962093777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,2048,768,0.018610666195551556
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,12288,256,0.003576888806290097
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3584,128,0.00629688882165485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3072,12288,0.037978665696250066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,3584,1536,0.01925511161486308
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,128,8192,0.010609777437316047
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,1536,128,0.014908444550302295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,8192,2560,0.008023111356629265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,7168,12288,0.05446933375464546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2048,256,0.006504888749784893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,32,1024,0.003105777833196852
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,128,65536,0.08454400300979614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,128,1536,0.023177777727444965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,256,512,0.006313777632183499
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,128,6144,0.01685066686736213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,5120,65536,3.016183005438911
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,12288,3584,0.02198222279548645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,32,10240,0.026141333911154006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,2048,1536,0.006284444282452266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3584,32,0.007994666695594788
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,12288,12288,0.03575733304023743
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,4096,4096,0.0285502208603753
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,7168,3072,0.026678222748968337
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,8192,768,0.02004888819323646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,65536,3584,0.6506471104092068
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,256,2048,0.007068444457319047
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2048,3584,0.00903555585278405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,512,64,0.0022239999638663395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,768,65536,0.06351555718315972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,12288,8192,0.027775999572541978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,8192,32,0.006933333145247565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,256,2048,0.006640000061856375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,64,768,0.006248000181383557
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,64,64,0.0021768889079491296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,1536,12288,0.01570933394961887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,512,1024,0.006260444306664997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,8192,6144,0.05218755536609226
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,65536,10240,3.0269510481092663
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,3584,1536,0.028453333510292902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,4096,512,0.02201866606871287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,65536,512,0.020376000139448378
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,16384,512,0.04984800020853678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3072,128,0.0022373333987262514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,65536,12288,3.3433475494384766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,512,3072,0.012470222181744046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,3072,6144,0.04096355703141954
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,12288,2048,0.01885777711868286
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,4096,16384,0.022466666168636743
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,2048,512,0.014840889308187695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,512,768,0.005948444621430502
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,7168,8192,0.027466666367318895
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1024,3584,0.026183999247021143
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,7168,3584,0.023834667272037927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,128,3584,0.01423911088042789
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,4096,512,0.007053333024183909
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,64,3584,0.007324444750944774
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,16384,1024,0.01573155489232805
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,256,3072,0.0144533332851198
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,65536,16384,0.227088901731703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,3072,3584,0.02268444498380025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3584,6144,0.01850044396188524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,512,1536,0.01626755628320906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,4096,512,0.0029537777105967202
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,6144,512,0.027266666293144226
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,768,65536,0.07398577531178792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,7168,10240,0.06445155541102092
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3072,1024,0.00599644457300504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,4096,256,0.016557332542207506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,65536,5120,0.23073156674702963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,16384,256,0.032103111346562706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,512,3584,0.007353777686754863
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,2048,4096,0.028152889675564233
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,5120,6144,0.046999110115898975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,7168,768,0.018254222141371835
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3584,256,0.002509333400262727
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,4096,10240,0.21199022399054634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1536,8192,0.01365600029627482
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,10240,64,0.008579555484983657
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,10240,8192,0.07426489061779447
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,1536,8192,0.021927111678653296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,12288,768,0.007668444679843054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,256,32,0.0021199999998013177
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,4096,32,0.006115555349323485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,65536,6144,0.12468800279829238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,65536,16384,0.34649067454867893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,6144,128,0.016509332590632968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,7168,3584,0.024485333098305598
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,10240,12288,0.10036710898081462
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3072,6144,0.012116444607575735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,5120,768,0.005081777771313985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,2560,256,0.014882667197121514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,3072,2560,0.01627111103799608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,512,10240,0.007897777689827813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,7168,3072,0.023367111881573994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,64,64,0.006275555739800136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,6144,1536,0.006657777975002925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,64,2048,0.004605333424276776
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,1536,5120,0.02882400155067444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,256,65536,0.035727110173967146
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,2560,512,0.014177777700954013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,4096,65536,0.06745777527491252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,16384,16384,0.08317599693934123
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,7168,4096,0.18315200010935465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,5120,16384,0.060214221477508545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,12288,1536,0.00851111114025116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3072,512,0.01484444406297472
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,10240,16384,0.0388373335202535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,3584,6144,0.009786666267448002
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,64,1536,0.003973333372010125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3584,65536,0.06502044200897217
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,12288,2048,0.010759111079904767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1536,64,0.00923911150958803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,8192,2048,0.007712889048788283
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,16384,5120,0.025969778498013813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,7168,256,0.01685688893000285
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,5120,65536,0.07093244128757052
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2048,3072,0.0064328888224230866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,2560,128,0.014839111102951897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1024,64,0.0022533333135975734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,5120,16384,0.040538665321138166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,768,512,0.013345777988433838
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,512,768,0.015275556180212231
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,12288,768,0.018889778190188937
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,512,10240,0.022724444667498272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,10240,8192,0.036208000448015
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3072,10240,0.07875200112660725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,10240,16384,0.06462933619817098
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3584,6144,0.013407111167907715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,64,2048,0.004631999880075455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,65536,1536,0.1198728879292806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3072,5120,0.009515555368529426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,5120,4096,0.010067555639478896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,256,512,0.005084444251325396
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,1024,7168,0.022777777579095628
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,64,16384,0.005963555640644497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,32,65536,0.007063999772071838
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,12288,16384,0.0406257775094774
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,12288,12288,0.0314977765083313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,768,3072,0.00556355549229516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,1024,128,0.005548444473081165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,4096,16384,0.048361778259277344
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1536,16384,0.012133333418104382
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,16384,5120,0.026405332816971674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,32,5120,0.014550222290886773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,7168,4096,0.053893334335751004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,256,1536,0.014819555812411837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,2048,512,0.014108444253603617
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,7168,512,0.005565333283609814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,12288,16384,3.2540399763319225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,7168,2048,0.006607999818192587
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,2048,65536,0.7488702138264974
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,2560,12288,0.017662222186724346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3072,128,0.0021662221600612006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,7168,3584,0.00869422240389718
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,4096,12288,0.04671288861168755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,65536,8192,0.17806043889787462
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,3584,16384,0.06843111250135633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,10240,12288,0.29510045051574707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,512,256,0.00833333366447025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,8192,6144,0.32599377632141113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,32,8192,0.06613422102398343
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,256,1536,0.02488800055450863
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,2048,65536,0.5747475624084473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,12288,12288,0.28323645061916775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,768,8192,0.030099554194344416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3584,768,0.016890666551060148
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,12288,12288,0.048325333330366343
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,6144,2048,0.018440888987647164
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,256,2560,0.013831111292044321
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,8192,5120,0.02739022175470988
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,64,1536,0.006969778074158563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,8192,2048,0.025444444682863023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,4096,512,0.005582222094138463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,5120,64,0.005932444499598608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2048,1024,0.01424800025092231
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,8192,65536,0.1437822182973226
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,16384,1536,0.010826666322019366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,4096,12288,0.01937244501378801
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,64,2560,0.004546666724814309
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,10240,6144,0.022843556271659002
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3072,16384,0.029110223054885864
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,64,65536,0.08206578095753987
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2048,3072,0.0069422221018208405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2560,256,0.002505777817633417
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,7168,65536,0.14188710848490396
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2560,5120,0.008957333034939235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,512,16384,0.09078577491972183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,512,12288,0.00871733327706655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3072,3072,0.007642666498819987
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,6144,65536,0.11750755045149062
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,3584,768,0.020008888509538438
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,256,10240,0.011388444238238864
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,2048,1536,0.01683999929163191
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,7168,65536,0.07951555649439494
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,5120,5120,0.011287111375066968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,3584,10240,0.01587199999226464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,2048,768,0.01757777730623881
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,2560,512,0.015389333168665567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,4096,768,0.04569333460595873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,10240,3072,0.01312711089849472
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,512,256,0.005997333261701796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,8192,7168,0.014226666755146451
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,2048,1024,0.00600622221827507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,1536,128,0.005588444570700328
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3584,12288,0.017477333545684814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,512,1024,0.005621333503060871
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,3072,12288,0.03478488988346524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,512,4096,0.00797333319981893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,256,16384,0.006643555644485686
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,64,4096,0.03230133321550157
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,4096,256,0.010831111007266574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,128,65536,0.07172799772686429
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,12288,5120,0.02239199976126353
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,65536,3072,0.040960888067881264
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,16384,768,0.007704888780911763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2560,10240,0.023286221755875483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,12288,128,0.006019555446174409
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,2560,2560,0.0069351109365622205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,4096,32,0.00216977774269051
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,65536,768,0.031516446007622614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3584,8192,0.012900444368521372
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,65536,5120,0.46009688907199436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,1536,128,0.005636444522274866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,64,3584,0.007640000018808577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,12288,8192,0.03851822349760268
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,512,3584,0.015156444576051501
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,6144,64,0.003576000117593341
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,8192,256,0.005582222094138463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,7168,3072,0.030255112383100722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,256,32,0.005933333188295364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,512,6144,0.009038222332795462
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,768,6144,0.012441777520709567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,2560,768,0.01684711045689053
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,3072,16384,0.02130399975511763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3072,1536,0.004597333156400257
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,5120,2048,0.025102222959200542
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2048,3584,0.00794488853878445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,8192,3072,0.2394933435651991
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,3072,3072,0.00831288927131229
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,10240,32,0.05385955505900913
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,5120,5120,0.19860355059305826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,3072,12288,0.04620888829231262
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,64,128,0.0022693334354294673
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,2048,4096,0.018216000662909616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,256,65536,0.050033777952194214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,4096,512,0.0062773335311147906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,3584,512,0.02550044490231408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,16384,32,0.0025173332542181015
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,1536,2560,0.015188443991872998
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,1536,1536,0.006274666637182236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,2560,1024,0.014855111638704935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,4096,10240,0.02609866691960229
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,5120,512,0.0026568888376156488
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,5120,16384,0.041872888803482056
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2048,6144,0.12689244747161865
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,256,16384,0.015575110912322998
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,16384,65536,0.3124755488501655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,7168,4096,0.009382222261693742
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,65536,7168,0.2048462231953939
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,6144,4096,0.011081777513027191
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1024,2048,0.006059555543793573
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,10240,5120,0.015808888607554965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,3584,3072,0.017266665895779926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1536,256,0.014771555860837301
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,2560,8192,0.03807199994723002
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,8192,12288,0.03237777617242601
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2048,7168,0.0114702218108707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,65536,16384,0.22762844297620985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,3072,8192,0.011758222348160215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,2048,5120,0.018000889155599806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,16384,256,0.023568888505299885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,768,256,0.005564444594913059
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,768,32,0.0021911110315057966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,8192,12288,0.1725884411070082
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,65536,256,0.011801777614487542
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2560,6144,0.017072889539930556
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,7168,2560,0.009655111365848118
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,32,32,0.001984888894690408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,256,2560,0.007264888948864407
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,65536,512,0.012187555432319641
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,5120,5120,0.012210666305489011
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,7168,768,0.017854221993022494
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,5120,1536,0.010875555376211802
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,8192,2048,0.007727111379305522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,64,8192,0.005718222094906702
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,7168,12288,0.031234665049446955
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,16384,8192,0.029863999949561224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,3072,12288,0.016560888952679105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,16384,12288,0.1309102243847317
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,3072,256,0.007687999970383114
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,5120,256,0.016951110627916124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,4096,6144,0.022825777530670166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,10240,256,0.004015999949640698
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3072,5120,0.009434666898515489
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,1024,768,0.01573511130279965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,5120,16384,0.02237866653336419
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,2560,5120,0.02183466653029124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,65536,2048,0.09863111045625474
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3584,512,0.002895111011134254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,12288,512,0.005621333503060871
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,2048,5120,0.008536000218656328
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,16384,1536,0.03797244363360935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,256,6144,0.018287110659811232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,8192,65536,0.08956088622411092
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,2048,12288,0.019733332925372653
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,2560,256,0.009407111340098912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1024,8192,0.052545779281192355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3072,16384,0.04986933204862806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,65536,128,0.0052266667286554975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1536,8192,0.010495999621020423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,65536,2560,0.03222399950027466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,4096,128,0.003925333420435588
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,5120,16384,0.21869066026475695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,32,256,0.005635555419656966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,4096,2048,0.010585777461528778
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,12288,3584,0.05265066689915127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,3584,1536,0.01553600033124288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2048,8192,0.012676444318559436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,32,10240,0.012473777764373355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,32,12288,0.005389333185222413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,10240,1024,0.02548888822396596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,64,16384,0.03578133384386698
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,5120,128,0.017148445049921673
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,8192,7168,0.021070222059885662
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,128,3584,0.007617777420414819
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,8192,128,0.01034311122364468
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,16384,4096,0.015454222758611044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,10240,256,0.006220444622966979
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,128,7168,0.01890222231547038
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,8192,768,0.006251555350091722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,65536,3584,0.09539377689361572
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,4096,16384,0.14627288447486028
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2048,128,0.016349333855840895
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,3584,1536,0.01718577742576599
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1536,8192,0.02197066611713833
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,8192,8192,0.4195093313852946
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,12288,4096,0.031523555517196655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,768,8192,0.01035822182893753
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,128,10240,0.005568888866239124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,5120,3072,0.02699822187423706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,4096,3584,0.009681777821646797
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,1536,768,0.016148444679048326
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,16384,16384,0.045864890019098915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,6144,10240,0.04200177722507053
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,2048,65536,0.06623466809590657
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,64,1536,0.0035617777870761026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3072,10240,0.1235111157099406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3584,3072,0.007522666619883643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,2048,10240,0.24403733677334258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,16384,10240,0.034846223062939115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,16384,7168,0.17131022612253824
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,2560,6144,0.0931733316845364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,65536,12288,0.2994222111172146
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,7168,8192,0.0462408893638187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,256,2560,0.005607999861240387
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,3584,65536,0.9246400197347006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,2560,1024,0.01334755536582735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1024,65536,0.12927377223968506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,1536,5120,0.01718577742576599
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,512,2560,0.004920000003443824
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2560,10240,0.016599999533759225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,10240,128,0.024041778511471216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,16384,3584,0.0196106665664249
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,4096,12288,0.03710222244262695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,1024,5120,0.022192888789706763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,12288,4096,0.09381955862045288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,6144,16384,0.042818665504455566
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3584,3072,0.08317777845594618
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,128,128,0.007657777931955125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,12288,3584,0.014203555054134794
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,16384,1024,0.008644444247086843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3072,256,0.0052968888647026485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1024,65536,0.37745067808363175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1024,128,0.0022115556316243275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2560,512,0.002895111011134254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,64,1024,0.0032008888406885993
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,3584,256,0.019928889142142404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,64,256,0.006634666687912411
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,16384,1536,0.01824977828396691
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,768,2560,0.00702133360836241
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,128,8192,0.014480888843536377
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,256,4096,0.007746666669845581
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,6144,1024,0.007630222373538547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2560,6144,0.010690666735172272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,7168,256,0.002558222247494592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,7168,8192,0.021432888176706102
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,512,1536,0.006707555717892117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,6144,1024,0.008030222521887885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2048,6144,0.06717599762810601
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,1024,7168,0.019704000817404855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,2560,4096,0.048932443062464394
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,10240,16384,0.0551111102104187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,8192,768,0.0635440018441942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,64,4096,0.00831200016869439
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,12288,1024,0.007367110914654202
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,5120,64,0.002558222247494592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3584,3072,0.007103111180994246
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,8192,12288,0.06973777876959907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,12288,1024,0.013384888569513956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,7168,10240,0.04108000132772658
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3584,3584,0.008764444126023186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,5120,8192,0.017567111386193168
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,65536,2048,0.05784622165891859
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,16384,6144,0.033751110235850014
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,7168,10240,0.01668444441424476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,2048,2048,0.006628444625271692
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,2560,768,0.010448888772063784
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,4096,512,0.005580444302823808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,1024,1536,0.016229333149062265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,64,4096,0.004736888739797804
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2560,512,0.005659555395444234
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,1536,4096,0.026452443665928308
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2560,2560,0.007344889144102733
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,7168,10240,0.03205777870284186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,65536,1024,0.06501155429416232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,32,256,0.006116444451941385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2048,512,0.010126222338941362
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,768,512,0.012807110945383707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,16384,8192,0.030091553926467896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,16384,10240,0.46112622155083555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,16384,10240,0.09163022041320801
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,128,64,0.011615999870830111
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,10240,32,0.010025777750545079
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,32,4096,0.004040889028045866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,512,16384,0.027630221512582567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,512,10240,0.013911111487282647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1024,4096,0.007978666987684038
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,16384,256,0.006859555426571105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,6144,5120,0.01109777804878023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3072,16384,0.03196444445186191
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2048,2048,0.004606222112973531
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,1536,4096,0.017920888132519193
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,12288,256,0.009872000250551436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,768,128,0.014119999276267158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,6144,4096,0.18366311656104195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1536,4096,0.0635146697362264
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,3584,256,0.0155102229780621
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,256,1536,0.013125333521101209
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,128,32,0.006747555401590135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3584,128,0.0025537777692079544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2560,3584,0.00775999989774492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,256,1536,0.012730666332774691
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3584,1536,0.006212444355090459
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,12288,2560,0.012460444536474017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,8192,2560,0.026351999905374315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2048,512,0.002883555574549569
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,128,16384,0.13059377670288086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,10240,3584,0.013175999952687157
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,5120,512,0.016149333781666227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3584,512,0.0028320000403457214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,32,10240,0.004729777574539185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,10240,10240,1.372783131069607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2560,65536,0.41437599394056535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,2048,5120,0.009108444054921469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,65536,10240,0.3694924513498942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,3072,2560,0.08358222246170044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,6144,128,0.012638222012254926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,32,1024,0.005974222388532426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,512,12288,0.014106666048367819
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,1024,2048,0.01585244470172458
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,128,512,0.010802666346232096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,768,16384,0.011374221907721626
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,128,768,0.006194666855865055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,2048,12288,0.039120889372295804
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,128,12288,0.00554577757914861
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1536,4096,0.06539733542336358
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,256,12288,0.013453333742088742
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,16384,64,0.0031991110493739447
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2560,512,0.005962666538026597
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,7168,1536,0.005886222339338726
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,10240,10240,0.03891822364595201
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,7168,4096,0.012444444828563266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,12288,3072,0.09400710794660781
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,2048,6144,0.009688888986905416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,768,10240,0.02616888946957058
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,2560,1024,0.015249777171346875
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,12288,2560,0.07679288917117648
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,8192,5120,0.02684355609946781
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,7168,32,0.005236444373925527
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,16384,8192,0.06516266531414457
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,512,768,0.006299555715587404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,2560,2048,0.01699466672208574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,10240,1024,0.008064000142945183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,12288,5120,0.02439822256565094
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,128,2048,0.012422222230169507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,10240,1024,0.006735111276308696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,10240,7168,0.03304977880583869
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,7168,5120,0.01127022256453832
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,768,3072,0.005572444448868434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,5120,32,0.002528000002106031
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2048,16384,0.04642755455440945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,7168,7168,0.019352888067563374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,256,2048,0.0143857780430052
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,1024,6144,0.009024889104896123
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1024,3584,0.007975999679830339
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,12288,8192,0.023296000228987798
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,8192,6144,0.023931556277804907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,8192,12288,0.04265866676966349
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,5120,8192,0.013829333086808523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,512,5120,0.008830221990744272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,128,2560,0.004953777624501122
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,8192,3584,0.010288000106811523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,768,3072,0.005583999885453119
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2048,512,0.005397333453098933
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,7168,768,0.008691555923885768
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,8192,8192,0.049560887946022876
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,6144,1536,0.07315733035405476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,3584,16384,0.04534133275349935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,2560,1024,0.014444443914625378
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,7168,256,0.01683466633160909
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,512,6144,0.006280000011126201
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,128,6144,0.009727111293209923
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,65536,1536,0.10040444135665894
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1024,768,0.005600888695981767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,768,8192,0.006268444574541516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,3584,65536,0.09692621893352932
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,1536,8192,0.038408888710869685
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,10240,1024,0.00792711145348019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,2048,10240,0.02826933397187127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,512,128,0.00249777775671747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,3584,2560,0.01203555530971951
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1024,32,0.011679111255539788
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,32,256,0.006665777828958299
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,512,64,0.002149333349532551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,512,8192,0.020959110723601446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,10240,1536,0.024301333559883967
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,512,3072,0.007000889215204451
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,2560,3072,0.01736000014675988
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,768,8192,0.010631110933091907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,16384,2560,0.2557315561506483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3072,2560,0.007449777589903937
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,6144,2560,0.03455288873778449
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,6144,3072,0.023191110955344305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,8192,2048,0.009280000295903947
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,65536,3584,0.20638133419884575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,5120,128,0.0028595555987623003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,16384,12288,0.058670222759246826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,2048,3072,0.024853333830833435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,2048,512,0.0053591111467944244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,65536,8192,0.07358666923311022
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,1536,5120,0.01761866609255473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,7168,1024,0.006035555568006303
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,16384,6144,0.027837332752015855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,16384,5120,0.05472444494565328
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,4096,2560,0.017301332619455125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,5120,16384,0.07149244679345025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,16384,2048,0.014492443866199918
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,512,64,0.005696889013051987
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,1536,1536,0.014191110928853353
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,7168,6144,0.018784888916545443
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,128,64,0.001842666624320878
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,7168,256,0.005427555491526921
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,256,65536,0.049253334601720176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,3584,1536,0.016576000385814242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,256,6144,0.006663110935025745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,7168,3072,0.030285331938001845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,10240,12288,0.04980977707438999
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,7168,256,0.04309689005215963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,10240,8192,0.0251875552866194
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,16384,5120,0.03520977828237746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,65536,2560,0.035439110464519925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,768,2560,0.007275555696752336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,2048,1536,0.015112888481881885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,12288,16384,1.1649270587497287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,12288,6144,0.018411555223994784
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,6144,6144,0.026705778307384912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3072,2048,0.006574222197135289
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1024,32,0.005562666803598404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2560,64,0.005919999960396025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,8192,64,0.002489777695801523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,12288,8192,0.029711110724343195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1536,3072,0.027594667341974046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,7168,64,0.006569777925809224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,2048,8192,0.015463110473420886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,4096,4096,0.08582933081520928
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,128,12288,0.060862223307291664
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3072,12288,0.018432888719770644
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,65536,12288,0.09853600131140815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,768,10240,0.021008888880411785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,3584,256,0.017437333861986797
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,32,1024,0.005943999936183293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,65536,3584,0.10570577780405681
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,3584,4096,0.06092888779110379
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,12288,6144,0.021538666552967493
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,4096,7168,0.013181333740552267
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,1024,65536,0.07413155502743192
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,5120,3072,0.00907022257645925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2560,768,0.005783111270931032
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,7168,65536,0.08331200149324205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,768,512,0.0028453332682450614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,6144,7168,0.028373334142896865
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2048,256,0.01719999975628323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3584,768,0.0028826666788922418
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,2560,10240,0.016685333516862657
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,5120,3584,0.0188417782386144
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,5120,768,0.005997333261701796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,65536,2048,0.0801244444317288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,2048,3072,0.007722666694058313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,16384,1536,0.0347324444188012
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,10240,6144,0.024720888998773363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,6144,1536,0.005920889063013925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1024,16384,0.01830666595035129
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,2048,2560,0.0069368887278768755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,65536,5120,0.06224355432722303
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,768,12288,0.027093332674768236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,32,6144,0.00489688871635331
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,256,3584,0.007907555335097844
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,32,4096,0.003896000070704354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,12288,12288,0.07179555628034803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,65536,6144,0.14947199821472168
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,8192,3072,0.011937778029176923
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,1024,1536,0.006595555692911148
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,256,3584,0.0155395550860299
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,6144,8192,0.01575644479857551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1024,768,0.003560888684458203
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,7168,65536,0.120779554049174
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,6144,16384,0.03459555572933621
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,1536,256,0.006647111227114995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,3072,3584,0.018952889574898612
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,10240,512,0.010635555618339114
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,3584,512,0.005579555614127054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3072,256,0.002534222271707323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,12288,64,0.007635555333561367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,256,768,0.014486221803559197
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,2560,128,0.005561777700980504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2048,16384,0.015027556154463025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2048,128,0.005611555443869696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,3584,5120,0.009880888793203566
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,8192,1024,0.01633777717749278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1536,2560,0.007254222200976477
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,8192,256,0.02159999973244137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1536,32,0.002152888932161861
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,6144,2560,0.007344889144102733
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,65536,32,0.009993777506881291
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,5120,768,0.005569777968857024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,10240,10240,0.1529973347981771
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,3584,768,0.01650311052799225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,65536,32,0.010052444206343757
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,10240,6144,0.3772631221347385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,10240,4096,0.1966924402448866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,512,1536,0.013178666432698568
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,5120,1024,0.01884622209601932
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,8192,4096,0.0291244453854031
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,5120,7168,0.01516533394654592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,1024,8192,0.03134399983617995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,1536,12288,0.032467554012934365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,4096,3072,0.010431999961535135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3584,64,0.0024906665914588505
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3072,7168,0.013786666923099093
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,65536,2048,0.05648889144261678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,10240,2560,0.011423110961914062
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2048,256,0.009337777892748514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,3584,7168,0.012764444781674279
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,8192,3072,0.010075555907355415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,4096,128,0.00591822216908137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,7168,256,0.005928888916969299
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1024,512,0.002507555608948072
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,1536,65536,0.0859137773513794
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3072,512,0.011428444749779172
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,6144,512,0.005699555493063397
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,16384,6144,0.024344889654053584
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,8192,128,0.0042257776690853965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,1536,256,0.005606222069925732
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,5120,2560,0.0074026667409473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,4096,768,0.0032684444967243406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,7168,1024,0.021295999487241108
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,256,3584,0.016948444975747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3584,10240,0.1408382256825765
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,512,2560,0.013763555222087435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,1536,4096,0.011807111402352652
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,512,10240,0.011845333708657159
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1024,768,0.007338666253619724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,5120,64,0.01628977722591824
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,2048,2560,0.006928000185224745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,1536,128,0.014503111441930136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3584,2560,0.016175111134847004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,8192,768,0.006214222146405115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,6144,10240,0.0378115541405148
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,128,2560,0.013778666655222574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,6144,6144,0.012871999707486896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,8192,4096,0.0269822229941686
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,768,768,0.02270222206910451
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,8192,10240,0.03700977894994948
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,16384,7168,0.0277528895272149
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2560,2560,0.006976889239417181
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,32,512,0.007781333393520779
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,6144,512,0.0069724445541699724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,6144,10240,0.577430248260498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,2048,7168,0.021741333934995864
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,3072,16384,0.04989599850442675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,256,8192,0.018923555811246235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,7168,2560,0.03126222226354811
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2048,10240,0.010439111126793755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,128,512,0.003910222401221593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1024,12288,0.02485688858562046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2560,256,0.0035075553589397003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,64,5120,0.01071911139620675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1536,1536,0.006243555496136348
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3072,64,0.0018586667461527719
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,8192,256,0.01718133356836107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,32,1024,0.007283555964628856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,1024,5120,0.008632000121805403
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2560,3072,0.007311111523045435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,7168,3072,0.009550222092204625
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,6144,65536,0.4898097780015733
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,512,4096,0.026195555925369263
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,768,3072,0.005580444302823808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,1024,128,0.01555199921131134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,8192,5120,0.012541333006487953
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,256,1024,0.007320889168315464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,3584,1024,0.005977777971161737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,12288,12288,0.1923804415596856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,64,16384,0.0155395550860299
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,128,256,0.006732444382376141
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,768,12288,0.01349422252840466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,32,65536,0.006294222341643439
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,768,4096,0.017300445172521803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,2048,3072,0.0070471109615431885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,128,512,0.01146666705608368
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,4096,64,0.005324444423119227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,16384,2560,0.05771733654869927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,32,10240,0.0121760004096561
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,512,16384,0.01529600057337019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,6144,6144,0.01604977746804555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,256,128,0.005592888842026393
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2048,3072,0.0058764442801475525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,4096,256,0.005305777821275923
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,1536,12288,0.07763733466466267
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,7168,512,0.016581333345837064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,128,4096,0.00462400002612008
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,768,8192,0.08059110906389025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,32,1536,0.0087013335691558
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,8192,128,0.005693333430422678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2560,3584,0.007703111403518253
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,768,1024,0.003683555457327101
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,6144,12288,0.030969778696695965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,4096,3584,0.009149333669079674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,128,7168,0.0637520021862454
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,12288,128,0.015242666006088257
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2560,3072,0.007079111205206976
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,32,4096,0.01366666704416275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,2048,6144,0.08420533604092068
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,128,8192,0.0058666666348775225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,32,12288,0.004548444516128964
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,12288,1024,0.022653333014912073
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,2560,5120,0.008966222405433655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1536,2560,0.005186666631036335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1024,256,0.00828888929552502
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,1024,3072,0.018219555417696636
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,10240,7168,0.04764444298214383
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,4096,256,0.015944000747468736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,5120,6144,0.013497778111033969
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,256,3584,0.007336888876226213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,512,128,0.00629688882165485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,6144,1536,0.008430221842394935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,8192,3072,0.08726666371027629
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,16384,12288,0.04248799880345663
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,512,1024,0.012788444757461548
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,6144,512,0.031862222486072116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,12288,3072,0.013215111361609565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,16384,256,0.006673777682913675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,6144,1024,0.011134222149848938
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,768,65536,0.05927555428610908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,128,10240,0.03310399916436937
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,2560,3072,0.03853333327505324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,256,3584,0.01686844395266639
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,6144,3584,0.026689777771631878
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,5120,12288,0.025978666212823655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,4096,2048,0.007262222468852997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,64,4096,0.004299555387761858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3072,4096,0.008793777889675563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,12288,512,0.033915556139416166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,32,2560,0.004912888838185204
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,12288,1536,0.008052444292439355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,64,256,0.005937777873542573
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3072,7168,0.011400889191362592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,1536,6144,0.1339662207497491
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,768,256,0.005561777700980504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,8192,3072,0.010761777559916178
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,128,768,0.024989333417680528
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,2560,256,0.0052471111218134565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2560,1536,0.005609777652555042
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2560,128,0.0025235555238193935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,64,4096,0.0047786666287316214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,128,5120,0.005045333256324132
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,768,7168,0.01647200021478865
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,8192,7168,0.014461333552996317
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,16384,65536,0.32708443535698783
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,12288,32,0.018346667289733887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,8192,16384,0.06076355775197347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,512,32,0.005594666633341048
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,2560,6144,0.055141333076688975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,1024,1536,0.006613333192136552
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,1536,1536,0.01421333352724711
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,5120,4096,0.008023111356629265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3584,16384,0.021198223034540813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,8192,7168,0.05330400003327263
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1536,2560,0.007607111500369177
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,1536,10240,0.03874933388498094
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,3072,2048,0.01808355583084954
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,7168,2048,0.008091555701361762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1024,5120,0.008633777499198914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1536,4096,0.006618666566080517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,512,128,0.002159111201763153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,1536,65536,0.08186222447289361
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1536,2560,0.005417777846256892
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,768,12288,0.008722667064931657
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,8192,3072,0.1734293301900228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3584,768,0.002926222152180142
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,6144,10240,0.016985777351591323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,2560,12288,0.03277333246337043
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,5120,3072,0.00982844415638182
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,5120,2560,0.030258665482203167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,8192,2048,0.0617413322130839
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,65536,2560,0.030209776428010728
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,7168,2048,0.03690133492151896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,3584,128,0.014940443966123792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,7168,4096,0.03010400136311849
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,5120,4096,0.022994667291641235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,128,8192,0.005968889014588461
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,1536,8192,0.17372088962131074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,2048,2048,0.0069360000391801195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,2048,2048,0.01757333344883389
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,512,2560,0.004926222066084544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,16384,4096,0.018820444742838543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2048,3584,0.007681777907742395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,3584,8192,0.024460444847742718
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,16384,5120,0.11225067244635688
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,3072,2048,0.0069511110583941145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,6144,5120,0.014429334137174817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,7168,10240,0.4661244286431207
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,7168,7168,0.012623111406962076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,65536,4096,0.1636319955190023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,10240,128,0.02029244436158074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,256,128,0.002250666626625591
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,3584,1536,0.016187555260128446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,65536,8192,0.8927279578314887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2048,4096,0.17047554916805693
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,512,16384,0.03747999999258254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,3072,1024,0.021282666259341772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,4096,7168,0.04656444324387444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3072,7168,0.11050666703118218
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,128,7168,0.010392888552612728
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,10240,3584,0.09342044591903687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,12288,8192,0.302565336227417
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,2048,3072,0.025046222739749487
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1024,512,0.00591822216908137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,3072,768,0.022289777795473736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,1024,1536,0.018206222189797294
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,8192,128,0.005566222386227713
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,4096,16384,0.1584764454099867
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,512,3584,0.012446222205956778
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,2560,6144,0.009327111144860586
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,7168,3584,0.1811760001712375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,2048,65536,0.0867955552207099
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,64,768,0.0028239999794297745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2048,7168,0.010989333192507425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,2048,256,0.005292444593376583
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,3584,7168,0.010066666536860997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,4096,2048,0.007989333735571967
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1024,512,0.005612444546487596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,256,12288,0.03641511003176371
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,7168,128,0.005568000177542369
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,5120,256,0.016599110431141324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,32,768,0.005613333235184352
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,7168,10240,0.16803199715084502
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1024,768,0.004648000001907349
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,32,2560,0.004918222212129169
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,5120,6144,0.011153777440388998
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,10240,12288,0.029680000411139593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,4096,3584,0.008223999705579545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,10240,6144,0.0634595553080241
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,5120,16384,0.6256657706366645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,4096,8192,0.054392887486351855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,512,5120,0.031918221049838595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,768,768,0.01074577785200543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,3584,256,0.0069884442620807225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,512,1024,0.005912000106440649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,16384,8192,2.3752053578694663
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,2048,10240,0.012511110968059964
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,16384,32,0.005930666708283954
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,128,65536,0.010811555716726514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,2560,2048,0.015087111128701104
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,256,2560,0.007296889192528195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,4096,1024,0.00498488876554701
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,32,6144,0.0053173332578606075
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,128,10240,0.02163377735349867
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,16384,512,0.005009777843952179
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2560,2048,0.005932444499598608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,65536,1024,0.02500088844034407
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,10240,1536,0.0076453329788313965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,7168,16384,0.04660088817278544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,256,2048,0.0064702220261096954
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,65536,768,0.0380657778845893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,4096,2048,0.016823111308945548
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,512,1536,0.005542222410440445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,6144,3072,0.010355555348926121
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,768,768,0.005747555444637935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,6144,65536,0.07279111279381646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,128,256,0.011855111353927188
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,12288,3584,0.015236443943447538
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,7168,3584,0.011267555256684622
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,8192,3584,0.013823111024167804
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,256,768,0.015200000670221118
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,256,1024,0.012497777740160624
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,256,10240,0.012881778180599213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2048,5120,0.010489777558379702
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,10240,256,0.0026319999661710528
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,64,2048,0.004638222356637319
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1024,12288,0.018025777406162687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,768,12288,0.01795733306143019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,512,1024,0.006311111152172089
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3072,768,0.00300177786913183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,5120,2048,0.055366221401426524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,512,65536,0.019704000817404855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,10240,65536,0.23913688129848906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,1024,5120,0.00851022203763326
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,10240,768,0.006351111249791251
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,5120,3584,0.028228445185555354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2560,8192,0.019305777218606737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,2048,4096,0.06092800034417046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,5120,65536,0.09903822342554729
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,32,32,0.0018373332503769135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,1536,2048,0.0063271112740039825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,2048,2560,0.006931555353932911
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1024,65536,0.05752888653013441
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,5120,7168,0.012609778179062737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,256,6144,0.008932444784376357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3072,1536,0.006610666712125142
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,8192,3072,0.009282666775915358
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,32,64,0.0019022222194406723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,16384,2560,0.04110577702522278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,64,12288,0.012768000364303589
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,16384,12288,0.04210222098562452
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2560,2560,0.005756444401211209
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,256,8192,0.023010666171709698
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,16384,1024,0.007453333172533248
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,768,2048,0.004538666870858935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,64,512,0.00535111129283905
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,12288,1024,0.007753777835104201
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,5120,1536,0.017535999417304993
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,10240,32,0.002560888934466574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,16384,2560,0.01388711151149538
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,16384,128,0.01351111133893331
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,4096,3072,0.029148442877663508
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,512,1024,0.006256000035338932
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,2560,7168,0.04265422291225857
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,128,2560,0.006797333144479328
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,768,3584,0.01087288889620039
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,64,1536,0.003583999971548716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2048,65536,0.059302224053276904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,10240,128,0.0025759999536805684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,128,256,0.004960000101062986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,16384,3584,0.01536622146765391
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,65536,16384,0.3737893369462755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,256,2048,0.006953777538405524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2048,4096,0.0074480002125104266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,256,3072,0.00868622213602066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,8192,768,0.004980444494220945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1024,2048,0.004385777645640903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,256,128,0.012215110990736218
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,768,65536,0.09399466382132636
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,6144,16384,0.09268444114261204
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3584,256,0.0244159996509552
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,2048,256,0.006330666856633292
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,8192,512,0.0027431110954946945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,256,10240,0.005668444352017508
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,3072,10240,0.03256533212131924
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,128,12288,0.02864711152182685
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,12288,16384,0.2987697654300266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,64,5120,0.008943110704421997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,10240,12288,0.03551288776927524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,12288,10240,0.3190595573849148
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,128,7168,0.08700178066889445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,4096,1024,0.08598844210306804
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,3072,3584,0.04453600115246243
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,5120,65536,0.11469156212276882
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2560,10240,0.015818667080667283
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,12288,1536,0.01036622209681405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,16384,2560,0.017110221915774874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,12288,128,0.038196444511413574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,10240,8192,0.052942223019070096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,7168,3584,0.03260177705023024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,128,5120,0.008620444271299574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,7168,64,0.0034311111602518293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,6144,256,0.006152888966931238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,65536,7168,0.15619822343190512
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,12288,6144,0.03425333234998915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,512,512,0.006075555665625467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3072,7168,0.010994666980372535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,65536,256,0.007172444628344641
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,10240,1536,0.025634666283925373
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1536,256,0.0026782221264309357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,12288,8192,0.041026665104760066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,3072,2560,0.006986666884687211
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,4096,12288,0.05428533421622383
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,7168,2560,0.12032266457875569
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,7168,1536,0.0199635558658176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,128,3584,0.0069253332912921906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3584,1536,0.008286221987671321
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,64,1024,0.003195555466744635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,128,256,0.0021484444538752237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,10240,3584,0.01349244432316886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,256,64,0.007999999655617608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,8192,1536,0.006592000110281839
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,7168,12288,0.20365599791208902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,4096,3072,0.0935093363126119
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,256,6144,0.010207110808955299
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1024,512,0.0059057776298787855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3584,5120,0.010084444450007545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,32,2560,0.0069395556218094295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1536,32,0.005462222215202119
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,6144,65536,0.11650578180948894
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,128,2048,0.016872000363137987
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,16384,12288,0.293283568488227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,8192,128,0.0028568889117903183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,6144,4096,0.014527110589875115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,65536,3072,0.12460177474551731
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,1024,1024,0.016550223032633465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,65536,256,0.00830577810605367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,128,512,0.010400888820489248
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,5120,1024,0.016197333733240765
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,7168,2048,0.007278222176763747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,5120,12288,0.023896000451511804
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,4096,3584,0.033893333541022405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,16384,3072,0.021783111823929682
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,64,64,0.005688000056478713
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1024,4096,0.007983999947706858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,4096,1536,0.061843554178873696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,768,2560,0.009343999955389235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,6144,65536,0.11795555220709907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,7168,7168,0.019291554888089497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1536,5120,0.009279111193286048
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,256,4096,0.021493333909246657
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,65536,64,0.01127199994193183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,768,512,0.007057777709431118
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,4096,8192,0.01701866587003072
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,1024,5120,0.013467555244763693
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3584,3584,0.008385777473449707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,128,4096,0.022925333844290838
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,128,6144,0.005342222336265776
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,65536,12288,0.2189075549443563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,3584,7168,0.023932443724738225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,256,2560,0.007113777928882175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,64,2560,0.007271111011505127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,12288,768,0.0196435550848643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,256,2048,0.015167999598715039
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,2048,1024,0.01660977800687154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3072,12288,0.018145778112941317
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,10240,2560,0.023365333676338196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,3072,12288,0.020346666375796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2048,256,0.002891555635465516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3072,4096,0.014574221438831754
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,16384,10240,0.06354844570159912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,16384,3072,0.0374311109383901
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,7168,1536,0.0063493334584765965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,5120,2560,0.007054222126801808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,5120,512,0.005676444619894028
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,512,3072,0.0162426663769616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,10240,2048,0.026698667142126296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,128,64,0.0018462222069501877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,65536,10240,0.20392799377441406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,7168,10240,0.03358755509058634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,8192,128,0.016875555117925007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,2048,4096,0.013829333086808523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,65536,1536,0.02536711096763611
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,8192,3072,0.008085333638721043
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,7168,1536,0.01926577753490872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,8192,6144,0.479564454820421
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,10240,3072,0.1970764398574829
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1536,65536,0.10456444157494439
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1024,8192,0.011054221954610614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,10240,2048,0.011046222514576383
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,5120,6144,0.02781955401102702
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1024,768,0.00907911111911138
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,7168,2048,0.11259644561343723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,7168,65536,0.08639288610882229
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,1536,768,0.01720533271630605
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2048,32,0.0063582224150498705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,2560,8192,0.013861333330472311
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3072,3072,0.00923111124171151
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,12288,5120,0.026571555270089045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,10240,512,0.025456888808144465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,5120,2048,0.018253333038753934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,65536,2048,2.243164486355252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,64,65536,0.05672622389263577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,12288,65536,0.19106577502356636
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,65536,6144,0.06794755326377021
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,16384,10240,0.036093334356943764
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,12288,7168,0.02084355552991231
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1536,512,0.003157333367400699
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,65536,1024,0.07227733400132921
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,4096,1536,0.0064204442832205035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,10240,5120,0.018058665924602084
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,256,2048,0.0046302220887608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,1536,2048,0.014565333724021912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,128,1024,0.013008000122176277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2560,768,0.004975999808973736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3072,12288,0.04722311099370321
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,768,32,0.005669333454635408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,8192,2560,0.0277511113219791
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3072,65536,0.11659199661678737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,1024,2048,0.016177778442700703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,3072,7168,0.015607999430762397
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,8192,6144,0.026199999782774184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,12288,10240,0.060382220480177135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,768,12288,0.010470222267839642
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,8192,65536,0.10592799716525608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,8192,8192,0.04906577865282694
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,256,2560,0.004959110998445087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,768,32,0.002228444442152977
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,12288,2560,0.01296622223324246
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3584,12288,0.01854044364558326
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,16384,4096,0.022302221920755174
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2048,64,0.0021884443445338146
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,16384,3072,0.45231199264526367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,4096,8192,0.03681333197487725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,768,4096,0.02520888878239526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,768,6144,0.012517333030700684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,4096,768,0.006304889089531369
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,12288,16384,0.04072355561786228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,32,1024,0.0063075555695427795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,64,65536,0.05587289068433973
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,3584,2560,0.02204177776972453
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,4096,1536,0.015486222174432544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,128,2560,0.013105777402718862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,4096,3584,0.02718044486310747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,128,5120,0.008360000120268928
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,512,512,0.0053226666318045715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,128,12288,0.013444444371594323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,12288,512,0.0042453333735466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,8192,8192,0.016940444707870483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2048,128,0.002226666650838322
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,8192,64,0.017464000317785475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2560,64,0.00222044438123703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,3072,6144,0.009131555755933126
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3072,10240,0.013863999810483722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2560,5120,0.007999999655617608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,3072,5120,0.008399999803966945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,7168,3584,0.011767999993430244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,4096,4096,0.021116443806224402
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,2048,6144,0.012838222086429596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,6144,4096,0.01257511145538754
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1024,2560,0.007274666594134436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2560,1536,0.004920000003443824
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,256,512,0.0025351111673646504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,5120,768,0.025397333833906386
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,65536,4096,0.05088177654478285
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,10240,256,0.0061039999127388
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,10240,1024,0.0069955554273393415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3584,64,0.006029333505365584
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,12288,6144,0.03029688861634996
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,2560,256,0.015264888604482016
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,10240,3584,0.023218666513760883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,5120,512,0.005573333137565189
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,2560,6144,0.009679111341635386
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,2048,16384,0.020073778099483915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,2048,4096,0.04006577862633599
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,1024,10240,0.03355822298261855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,12288,32,0.012437333663304647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3584,32,0.002185777864522404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,6144,16384,0.04346399837070041
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1024,7168,0.011400000088744693
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1024,32,0.0021848889688650766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,5120,768,0.0028533333291610083
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,1536,2560,0.006933333145247565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,6144,16384,0.09056799941592747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,4096,65536,0.08025333616468641
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,65536,3584,0.3348088794284397
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,12288,2048,0.012455110748608908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,64,12288,0.028328001499176025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,128,16384,0.07391466697057088
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,65536,6144,0.12026755015055339
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,5120,64,0.001826666709449556
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,4096,128,0.014875556031862894
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,6144,2560,0.02068444424205356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,65536,8192,0.17883911397722033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,10240,2048,0.05650133556789822
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,768,768,0.011031111081441244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,3584,6144,0.02783999840418498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,8192,3072,0.00848711116446389
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3072,1024,0.006258666515350342
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,6144,16384,0.060814221700032554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,64,1536,0.0035297779573334586
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,128,768,0.005671111245950063
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,256,2048,0.007963555554548899
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,65536,12288,0.18386667304568818
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,16384,7168,0.03762400150299072
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,4096,16384,0.03780355718400743
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,3072,3584,0.007704888780911763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,768,3584,0.015969778100649517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1536,4096,0.020080000162124634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,12288,10240,0.026679111851586237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,768,7168,0.006325333482689328
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,10240,2048,0.0233271105421914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,7168,1024,0.02239199976126353
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,7168,1024,0.0042408886882993905
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,512,64,0.00628266649113761
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,2048,5120,0.0291244453854031
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,4096,3584,0.007669332954618666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,10240,5120,0.03054311209254795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,256,1536,0.005939555664857228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,512,3072,0.007629333270920648
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,5120,5120,0.00943377779589759
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,2048,256,0.015455111861228943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,256,256,0.011746666497654386
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,12288,2048,0.011738667057620155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,512,128,0.0021644443687465456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,6144,256,0.00793866647614373
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,3584,7168,0.010402667025725046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,16384,4096,0.20369244946373832
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,512,32,0.005558222118351195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,16384,128,0.006246222390068903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,1024,3584,0.010567111273606619
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,32,65536,0.15651199552747938
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,2560,16384,0.04764622118737963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,8192,512,0.02240000002914005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,2560,7168,0.011352000137170156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,16384,128,0.002505777817633417
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1024,64,0.00629066675901413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,6144,16384,0.033118221494886614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,10240,128,0.015418666932317944
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,7168,12288,0.14262044429779053
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,8192,1536,0.012040889097584618
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,3584,65536,2.244826634724935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,7168,6144,0.01719466679626041
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,4096,256,0.005519999812046687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,7168,10240,0.23011554612053764
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,512,6144,0.010092444717884064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,512,4096,0.00795377790927887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,10240,3584,0.015604444675975375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,3072,256,0.006588444527652528
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,128,1536,0.01240799989965227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,7168,2560,0.013532444834709167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,768,16384,0.020040889581044514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,6144,256,0.006616000086069107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,768,512,0.005646222167544895
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,65536,128,0.1046782202190823
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,10240,16384,0.042986667818493314
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,16384,512,0.012977778083748288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,2048,2560,0.015461333923869662
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,6144,5120,0.012938666674825879
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,512,1024,0.0031999999450312722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,64,512,0.002500444443689452
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,4096,3584,0.007611555357774098
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2048,8192,0.018723555737071566
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,768,768,0.009947555760542551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,8192,4096,0.016964443855815463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,1024,10240,0.035946667194366455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,2560,3072,0.02269777821169959
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,1024,512,0.013423110875818463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,65536,64,0.05451911025577121
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,12288,8192,0.04009333252906799
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,256,6144,0.01793511046303643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1024,7168,0.015840000576443143
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,256,10240,0.0199635558658176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,6144,2048,0.007280000381999546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,12288,7168,0.21267822053697374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,65536,32,0.01276266657643848
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,16384,6144,0.10404266913731892
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,768,512,0.005968889014588461
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,16384,5120,0.056208001242743603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,5120,32,0.007613333563009898
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,16384,8192,0.37885244687398273
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,65536,16384,0.29142488373650444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,12288,8192,0.04595999916394552
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,512,2560,0.007275555696752336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,3584,3584,0.01891911029815674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,12288,3072,0.04827022221353319
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,16384,8192,0.07384088966581556
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1536,65536,0.03715644280115763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,128,2560,0.00703111125363244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,32,6144,0.00508088908261723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3584,2560,0.006981333096822103
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,2560,768,0.01682844426896837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,768,1024,0.006026666611433029
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,128,3584,0.004920888692140579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,512,1024,0.0064293332397937775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,256,64,0.0018871112002266778
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,10240,6144,0.19263822502560088
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,2048,12288,0.014827556080288358
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,512,65536,0.06369955672158135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,10240,12288,0.044849776559405856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,4096,2560,0.007364444434642792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,768,768,0.015639111399650574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,8192,256,0.016594666573736403
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,2560,512,0.013539555172125498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,7168,2048,0.05951733059353299
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,2560,1024,0.02243822150760227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,4096,1024,0.04335288869010078
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,8192,8192,0.11063466469446818
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,8192,32,0.016169778174824186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,6144,64,0.0021404443929592767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,7168,512,0.005567111074924469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,128,1536,0.003580444388919406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,10240,512,0.006708444406588872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1536,12288,0.019311999281247456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,7168,256,0.002560888934466574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,7168,3584,0.011026666396194033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,12288,3584,0.017663111289342243
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,65536,5120,0.1202053361468845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,768,4096,0.00795111142926746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,32,768,0.0028400001012616684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,4096,128,0.0033084443873829315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,512,5120,0.006579555571079254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,3584,5120,0.008997333546479544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,512,8192,0.01037688884470198
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1024,16384,0.011426666544543372
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,768,1024,0.0035386664999855887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2048,3072,0.006800000038411882
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,6144,5120,0.035747554567125105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,1024,65536,0.16612799962361655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,64,64,0.005623999983072281
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,512,1024,0.006615110983451207
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,5120,65536,0.1022648877567715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,7168,5120,0.06787555747561984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,2560,3584,0.007656888829337225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,2560,12288,0.01556799974706438
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,6144,3584,0.022638221581776936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,4096,1024,0.0059164443777667145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,5120,10240,0.11127021577623154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,8192,128,0.005951111101441913
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,6144,6144,0.03756355577045017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,32,64,0.0018915554715527429
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,32,6144,0.009963555468453301
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,8192,2560,0.05602755811479357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3072,4096,0.013822221921549903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,64,2048,0.0069297779765393995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1024,1536,0.0052879999081293745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,768,3584,0.008006222546100616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,256,6144,0.01718222267097897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,2560,65536,0.06405155526267158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,256,16384,0.026148445076412622
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,16384,32,0.008447999755541483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,10240,65536,0.14917333920796713
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,2048,512,0.005656888915432825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1024,4096,0.006054222169849608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,10240,768,0.013057777451144325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,12288,256,0.0684444440735711
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,32,6144,0.010088889135254754
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3072,768,0.012520888613329994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,4096,7168,0.014899555179807874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,12288,16384,0.04250311189227634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3072,1024,0.003592888928121991
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2560,8192,0.010653333531485664
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,256,4096,0.015340444114473132
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,65536,64,0.04227733280923632
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,5120,2560,0.006904888898134232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,512,3072,0.015259555644459195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,16384,2560,0.08214222060309516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,10240,1536,0.11094488700230916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,7168,7168,0.024192000428835552
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,8192,1024,0.007790221936172909
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,10240,768,0.026549332671695288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,12288,768,0.006761777732107375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,3584,256,0.0052933332820733385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,10240,128,0.01571555601225959
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,7168,65536,0.1600373321109348
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1024,6144,0.007072889142566257
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1024,2048,0.004974222017659081
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,4096,2048,0.01702577703528934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2048,10240,0.01255644443962309
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,512,1024,0.005976888868543837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,7168,768,0.004549333204825719
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,10240,2048,0.08330044481489393
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,5120,4096,0.010054222411579555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1536,16384,0.020608888732062448
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1536,3584,0.007699555820888943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,1024,512,0.015528000063366361
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,4096,8192,0.04505066739188301
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,12288,3072,0.012770666844315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3584,12288,0.0422071119149526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,16384,7168,0.034649779399236046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,1536,65536,0.06328888734181722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,65536,64,0.29647111892700195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,5120,2048,0.007009777757856581
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,4096,3584,0.010613333019945355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,65536,7168,0.17088622517055937
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,256,128,0.005620444400442972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,2048,1536,0.016869333055284288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,128,256,0.012237333589129977
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2560,256,0.005576888720194499
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,65536,256,0.008636444807052612
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,12288,2560,0.014674666855070325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,128,3584,0.01075111081202825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,6144,4096,0.049534221490224205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,2048,128,0.005660444498062134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,6144,10240,0.01609777741962009
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,768,6144,0.010392888552612728
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,512,6144,0.029159112109078303
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,1536,10240,0.013711999687883588
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,10240,65536,0.16024800141652426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,10240,256,0.017273777061038546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,8192,768,0.01775644388463762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,3072,3584,0.017626666360431247
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,1024,1536,0.02004888819323646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3072,512,0.012484444512261285
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,3072,5120,0.009398221969604492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3584,1536,0.006331555545330048
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1024,3584,0.007721777591440413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3584,16384,0.060030221939086914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,10240,512,0.004751111070315043
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,4096,512,0.005941333456171884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1536,8192,0.010737777584128909
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3584,2560,0.007356444166766272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,512,5120,0.006639110959238476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,1536,512,0.013801777528391944
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,128,1024,0.005655999812814925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,128,128,0.010814222196737925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1024,12288,0.01148888882663515
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,1024,2560,0.05982222159703573
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1536,1024,0.005609777652555042
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2048,3584,0.007672000262472365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,7168,1024,0.038252443075180054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,6144,12288,0.030881779061423406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,10240,65536,8.801564534505209
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,128,12288,0.010747555229398938
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3072,32,0.010782221953074137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,10240,64,0.005230222311284807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,512,3072,0.012113778127564324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,2560,8192,0.011060444845093621
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,4096,4096,0.02142133315404256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1536,1024,0.0035822221802340615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,768,6144,0.009407999614874521
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2560,2560,0.00700711127784517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,10240,12288,0.030376000536812678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,1024,512,0.013447110851605734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,256,12288,0.02540622154871623
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,10240,2560,0.12009421984354655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,1024,1536,0.012092444631788464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,5120,1536,0.015045333239767285
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,768,1536,0.006288888967699475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,128,5120,0.0069573331210348345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2560,3584,0.04687111245261299
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,65536,10240,0.10708710882398818
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,6144,128,0.0052533331844541765
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,7168,8192,0.018686221705542672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,10240,65536,0.14117422368791369
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,7168,8192,0.022472888231277466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,768,1536,0.027969777584075928
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,12288,5120,0.048229333427217275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1536,6144,0.009820444716347588
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,1024,1536,0.014945778581831189
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,768,7168,0.009177777502271865
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1536,65536,0.1064293318324619
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,3584,128,0.0053022222386466126
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,512,5120,0.005447111195988125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2048,65536,0.05307555529806349
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,7168,65536,0.07988178067737155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,16384,3072,0.014850666125615438
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,64,3072,0.0046453335218959385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,2560,1536,0.019514666663275827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2560,5120,0.009696889254781935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3584,5120,0.010187555518415239
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,12288,16384,0.08540711137983535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2560,1536,0.0046471113132105935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,128,3072,0.004898666507667965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3584,2560,0.009751111268997192
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2048,3072,0.007835555407736037
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,768,8192,0.023016000787417095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,4096,256,0.01666933298110962
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,2560,512,0.01941066649225023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,10240,1024,0.012280888855457306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,3072,4096,0.018703111343913607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,10240,65536,0.20820444160037568
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,6144,10240,0.04920088913705614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,128,4096,0.04297777679231432
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,12288,1536,0.024853333830833435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,768,5120,0.007448888487286038
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,8192,32,0.0032035555276605818
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,10240,2048,0.02377333409256405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,65536,7168,0.1567902167638143
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,256,512,0.005967999911970562
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,7168,16384,0.06520089175966051
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,7168,512,0.005953777581453323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,16384,8192,0.033632887734307185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,768,4096,0.007719111111429002
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,65536,1536,0.07293688588672213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3584,32,0.007676444119877285
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,3072,1536,0.005958222266700532
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,768,512,0.005256888767083486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,4096,256,0.002526222210791376
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,1024,8192,0.01961599952644772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,7168,64,0.002584888910253843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,768,2560,0.017315554949972365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,1024,4096,0.016420443852742512
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,12288,64,0.003607999947335985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2560,256,0.005640888793600931
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,4096,3584,0.018928888771269057
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,7168,512,0.005706666658322017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,7168,2048,0.0114702218108707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,768,1536,0.006304889089531369
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2048,65536,0.13954399691687688
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3072,2048,0.019524445136388142
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,5120,2560,0.0203155560625924
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,10240,256,0.0025884444928831523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2560,16384,0.0440577765305837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1024,3072,0.007920888562997183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,6144,3584,0.007338666253619724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,65536,10240,2.9934285481770835
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,65536,32,0.007642666498819987
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,8192,16384,0.4081884490119086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,32,128,0.0021822222818930945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,5120,2048,0.008389333056079017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,8192,512,0.003930666794379552
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2048,768,0.004249777644872665
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,4096,64,0.00831200016869439
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1024,64,0.0034151110384199354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,2560,65536,0.06534844636917114
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,6144,1536,0.0070257774657673305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,5120,16384,0.39303021960788304
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,7168,12288,0.21152622169918486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,256,1024,0.005976000179847081
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,2048,6144,0.05241688754823473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,2560,16384,0.05868977970547146
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,512,8192,0.010732444624106089
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,16384,64,0.005961777849329843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,10240,6144,0.02015733387735155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,3072,1536,0.007000000112586551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,256,8192,0.018636445204416912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,10240,3072,0.013423110875818463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,256,1536,0.003910222401221593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,2048,4096,0.007622222105662028
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3584,2048,0.009526222116417354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2560,64,0.0024844445288181305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,5120,512,0.006272000157170826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,1024,2048,0.006615110983451207
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3072,128,0.0021955555097924336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2048,3584,0.008295111358165741
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,128,64,0.0021626665774318907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,2048,1024,0.005608888963858287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,5120,8192,0.028060442871517603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,4096,64,0.007703111403518253
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,5120,2048,0.016917333006858826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,32,16384,0.005993777679072485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,10240,3072,0.038754665189319186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,768,65536,0.056519110997517906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1024,7168,0.008622222476535374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,8192,10240,0.055496888028250806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,1024,10240,0.10720622539520264
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,10240,128,0.003540444291300244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,6144,256,0.016916443904240925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,256,1024,0.003216888962520493
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,512,4096,0.008377778033415476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,8192,3584,0.018547554810841877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,4096,1536,0.019703111714786954
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,256,6144,0.00527022240890397
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,3584,8192,0.01279022213485506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,1024,128,0.005326222214433882
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2048,1536,0.013636444177892474
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,4096,2048,0.021720000439220007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,7168,128,0.002540444541308615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,64,12288,0.005296000176005893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,12288,768,0.007325333025720384
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,32,2560,0.0068488890926043195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,8192,3072,0.029532445801628962
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,8192,256,0.0024986666523747975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,10240,3072,0.012160888976520963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,4096,3072,0.013293333351612091
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1536,7168,0.010392000277837118
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,256,16384,0.03632355398601956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,256,2048,0.006635555376609166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,10240,256,0.0031537777847713898
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,128,5120,0.016013332539134555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,7168,4096,0.03326133224699233
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,512,2048,0.004344888859324985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,3584,10240,0.015600000818570455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,65536,3072,0.08627822001775105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,4096,6144,0.011260444091426002
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,8192,7168,0.021848888860808477
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,128,128,0.005295111073387994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,4096,8192,0.02114222281508976
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1024,6144,0.009447111023796929
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,512,4096,0.006632888896597757
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,256,6144,0.022477333744366963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,12288,7168,0.020058666666348774
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,512,512,0.012390221986505719
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,65536,8192,0.194069332546658
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,4096,10240,0.014913777510325113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,10240,16384,0.26055288314819336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,10240,10240,0.052593777577082314
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,7168,1024,0.004424889054563311
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,128,64,0.0026604444202449587
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,2560,1536,0.006544888847404056
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,3072,6144,0.02077511118517982
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2560,3072,0.0074720001882976955
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,4096,64,0.006057777752478917
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,12288,6144,0.02377066678471035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,5120,5120,0.012206222448084088
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2560,12288,0.07156088617112902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,1024,1024,0.024157333705160353
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,1024,5120,0.03750577900144789
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,7168,8192,0.04116444455252753
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,4096,8192,0.043918222188949585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,12288,512,0.02225688927703434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,6144,16384,0.08558577961391872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,512,65536,0.18412178092532686
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,8192,3072,0.02697333362367418
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2048,512,0.005763555566469829
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,128,2560,0.0069208890199661255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,3072,4096,0.02104622291194068
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,65536,768,0.015936000479592215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,10240,8192,0.030255112383100722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,768,64,0.0021128888345426982
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,4096,64,0.005921777751710679
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,12288,7168,0.020432000358899433
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,12288,8192,0.5803199874030219
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,1024,16384,0.07718755801518758
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,32,128,0.00351555562681622
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,65536,16384,14.689914279513888
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3072,256,0.02885066800647312
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,7168,65536,0.310148451063368
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,65536,512,0.022600000103314716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,32,16384,0.005964444329341252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2560,2048,0.016484444340070088
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,6144,8192,0.16654488775465223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,512,2048,0.006335111127959357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,256,10240,0.007368000017272101
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,32,3584,0.007979555262459649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,5120,512,0.01719466679626041
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,7168,4096,0.01018666641579734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,6144,16384,0.0480328897635142
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,256,3584,0.007611555357774098
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,256,12288,0.013842666314707862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,4096,5120,0.00869955536392
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1536,7168,0.012004444168673621
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,4096,12288,0.17244356208377412
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,6144,16384,0.3462577660878499
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,10240,4096,0.01644533375898997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,2560,6144,0.026712889472643535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,2048,8192,0.01072533345884747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,256,3584,0.007679111427730984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2560,5120,0.008464888566070134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1024,8192,0.022114665971861944
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,65536,768,0.11803821722666423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,128,768,0.009669333696365356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,2560,12288,0.015936000479592215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2048,10240,0.022149332695537146
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,512,128,0.012831110921170978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,1536,512,0.005647111270162795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,1024,1024,0.005960000058015187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3584,3584,0.09619466463724773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2048,10240,0.025224889318148296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,1024,8192,0.010778666370444827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,4096,512,0.005716444303592046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3072,12288,0.017916444275114272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,1536,3072,0.007368000017272101
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,5120,512,0.011078221930397881
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,10240,1024,0.007083555890454187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2560,5120,0.010259555445777046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,3584,8192,0.07849599917729695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2048,16384,0.019539554913838703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,512,768,0.005639111002286275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,3584,512,0.01496088835928175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,512,2560,0.015184889237085978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3584,64,0.005976000179847081
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,256,64,0.005488889084921942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,1536,512,0.013167111410035027
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,10240,10240,0.026233777403831482
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,8192,8192,0.01974933346112569
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,8192,65536,0.20612266328599718
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,7168,12288,0.018630221486091614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,8192,6144,0.015406222807036506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,128,1024,0.013238222234778933
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,1536,12288,0.013359111216333179
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,4096,1024,0.02576622201336755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1536,64,0.013721778160995908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,3072,1024,0.018633777896563213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,7168,3072,0.0125937776433097
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,512,256,0.01510933372709486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,256,1024,0.006267555471923616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,65536,4096,0.14906577269236246
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,10240,12288,0.046316444873809814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,128,1536,0.006603555546866522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2560,10240,0.015544889701737298
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,4096,768,0.01789955629242791
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1536,2048,0.009746666583749983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,32,64,0.0019386667344305252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,4096,10240,0.01700799994998508
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,7168,1536,0.006728000111050076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,64,768,0.009728000395827824
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,10240,12288,0.5986853175693089
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,64,65536,0.09816000196668838
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,16384,2048,0.017157332764731515
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1024,7168,0.02221777704026964
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,16384,512,0.0399235553211636
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,65536,3072,0.18497422006395128
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,768,512,0.005702222386995952
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,128,3072,0.013920888304710388
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,3584,768,0.018046221799320646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,16384,3072,0.08452888992097642
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,7168,64,0.00907733374171787
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,64,1024,0.009351111120647853
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,10240,2560,0.030258665482203167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3584,6144,0.011759999725553723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,256,1536,0.013482666677898832
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,1536,65536,0.06303377946217854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,512,3584,0.015643555257055495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,3584,5120,0.016872889465755887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,32,5120,0.004608889006906086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,2048,3584,0.007530666887760162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,512,256,0.014496889379289416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,256,768,0.00590133335855272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,5120,256,0.016898666818936665
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1536,1536,0.00927466650803884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,2048,8192,0.010767999622556897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,5120,16384,0.050108446015252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,8192,3072,0.011006222003036074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,12288,12288,0.035585777627097234
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,768,1024,0.01475911173555586
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,768,32,0.0025439999169773525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,512,7168,0.01924533314175076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3584,3584,0.06215200159284803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,65536,8192,0.12550222873687744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1536,2560,0.006928000185224745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,12288,128,0.015542222393883599
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1536,64,0.0059057776298787855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1536,16384,0.02130488885773553
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,4096,2560,0.0075777777367168
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,65536,1024,0.0251857770813836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,8192,3584,0.011768889096048145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,4096,256,0.02694844371742672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,7168,7168,0.06199377775192261
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,7168,768,0.005869333528810077
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,65536,3072,0.05022844341066149
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,16384,6144,0.01976977785428365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1536,5120,0.017064000169436138
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,16384,2560,0.1319671074549357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,2048,8192,0.016917333006858826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,512,3072,0.015217777755525378
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,6144,7168,0.017447110679414537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,512,3072,0.016536888149049547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,10240,1536,0.007652444144090016
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,256,512,0.0053102220926019884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2048,10240,0.010874666273593903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,12288,512,0.019689778486887615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,1536,768,0.01793511046303643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2048,10240,0.015048000547620984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,768,64,0.0022053333620230355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,3072,7168,0.010191111101044549
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,6144,3584,0.02901866700914171
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3072,2048,0.018586667047606576
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,7168,512,0.0030168888883458245
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,8192,12288,0.15713332759009466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,6144,2560,0.012175111307038201
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2560,10240,0.02273688879277971
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,128,65536,0.04906488789452446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1024,12288,0.1250071128209432
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,512,12288,0.025507556067572698
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,6144,4096,0.024566221568319533
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,512,6144,0.006079999936951532
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,64,128,0.005606222069925732
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,2048,6144,0.008805333740181392
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3584,1024,0.005848888721730974
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,16384,2048,0.020043555233213637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,768,10240,0.011807999677128263
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,8192,3584,0.010097777677906884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,4096,7168,0.013518222504191928
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,65536,3072,0.037216888533698186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,10240,512,0.058352887630462646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,8192,256,0.01692800058258904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,256,128,0.0021911110315057966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,64,512,0.002566222308410539
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,12288,65536,0.15390577581193712
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,1536,65536,0.14913066228230795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,12288,7168,0.020066666934225295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2560,768,0.012359999948077731
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3584,2560,0.0069022224181228215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3072,32,0.002578666640652551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,512,5120,0.017370666066805523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,64,16384,0.006005333529578314
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,64,3584,0.004622222234805425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,64,256,0.0022622222701708474
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,16384,1024,0.007962666451931
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,128,1024,0.005603555589914322
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,1024,256,0.014831110835075378
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,12288,2560,0.033737778663635254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,512,5120,0.017298666967286002
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,5120,3584,0.008769777913888296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,4096,4096,0.03046400017208523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,65536,6144,0.13085688485039607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,3072,5120,0.020061333974202473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,64,256,0.0053075556125905775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,3584,3584,0.02313599983851115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,65536,8192,0.15913155343797472
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,2048,2560,0.021332444416152105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,64,512,0.002603555512097147
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,7168,16384,0.027041777968406677
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,7168,1536,0.027991112735536363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,512,10240,0.011770666473441653
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,64,3584,0.005589333259397083
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,1024,7168,0.026594666971100703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,6144,10240,0.04253866606288486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2560,2048,0.008443555898136562
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,768,2048,0.009319999979601966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1536,768,0.005657777604129579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,3072,8192,0.025416889124446448
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,512,5120,0.007307555940416124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,1024,8192,0.030861334668265447
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,65536,2048,0.03812888926929898
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,6144,3072,0.039184000757005476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,65536,1024,0.27085333400302464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,7168,3072,0.009288888838556077
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,64,3072,0.007294221884674496
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,512,5120,0.008303999900817871
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,128,32,0.005296000176005893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,16384,2560,0.03092355529467265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,64,16384,0.005601777798599667
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,128,32,0.005632888939645555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,256,16384,0.01588444411754608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,6144,16384,0.0248026665714052
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3072,3584,0.029256890217463177
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,7168,2560,0.01499466598033905
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1024,12288,0.014123555686738757
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,768,2560,0.006990222467316522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,768,12288,0.014284445179833306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1536,768,0.007311111523045435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,4096,10240,0.027227555712064106
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3584,64,0.002230222233467632
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,128,10240,0.01295644458797243
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,10240,7168,0.12589422861735025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,512,1536,0.013239999612172445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,3584,256,0.01682488951418135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2048,65536,1.8737742106119792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,32,4096,0.01036177741156684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,5120,128,0.01685688893000285
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,6144,1024,0.00869599978129069
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,256,3072,0.007660444411966536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,5120,4096,0.041645334826575384
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2048,7168,0.011927111281288994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,64,4096,0.008157333566082848
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3584,6144,0.01109155515829722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,8192,512,0.003318222239613533
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,16384,3072,0.031606223848130964
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1024,1024,0.003546666767862108
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,8192,65536,0.14545599619547525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2048,4096,0.02400711178779602
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,16384,256,0.006630222416586346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,3072,256,0.005283555636803309
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,1024,7168,0.009664889011118147
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,12288,256,0.003162666741344664
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,128,1536,0.00831200016869439
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,32,256,0.002513777878549364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,4096,1024,0.005961777849329843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,65536,4096,0.1362826691733466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,12288,7168,0.023053333163261414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1024,2560,0.007223111059930589
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,768,16384,0.012302222351233164
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1024,10240,0.0101724440852801
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,1536,7168,0.036350223753187395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1536,5120,0.04101422097947862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3072,64,0.006084444622198741
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,10240,64,0.0032764443506797156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,1024,1536,0.014896889527638754
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,4096,2560,0.007729777859316931
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,768,1024,0.003565333369705412
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3072,2560,0.007520889242490132
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1536,5120,0.014479999740918478
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,128,768,0.012699555191728802
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,4096,1024,0.005881777654091517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,8192,4096,0.013228444589508904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,6144,256,0.03999911083115472
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2048,65536,0.10059555371602376
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,4096,10240,0.016899555921554565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,16384,6144,0.5743946499294704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,65536,4096,0.08037066459655762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3584,7168,0.017672888106769986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,4096,4096,0.025046222739749487
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,768,1536,0.00601333338353369
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,4096,768,0.018276444739765592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,768,8192,0.010799110763602786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3072,64,0.002569777684079276
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,16384,256,0.022276444567574397
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,5120,2048,0.03718044360478719
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,512,1536,0.006340444501903322
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,256,5120,0.008654221892356873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,768,12288,0.02921600143114726
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3072,2560,0.07874933216306898
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,2560,1024,0.016524444023768108
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,3072,5120,0.024280000064108107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,768,16384,0.0159599996275372
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3584,6144,0.01202666676706738
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,8192,16384,0.02608088817861345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,16384,16384,0.37464266353183323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3584,768,0.006599111275540457
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1536,2560,0.005613333235184352
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3072,8192,0.01515111161602868
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,768,16384,0.03576888971858554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,10240,7168,0.03604622350798713
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,16384,6144,1.1367786195543077
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,4096,65536,0.13970222738054064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,256,1536,0.006455111006895701
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2048,64,0.00217333332531982
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3584,10240,0.029193778832753498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2560,128,0.0025333333760499954
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,32,1536,0.006261333409282897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,5120,2048,0.017190222938855488
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,128,16384,0.019515555765893724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1536,256,0.005903111149867375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,64,768,0.002569777684079276
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,7168,768,0.005919999960396025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1536,16384,0.014030221435758801
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,768,2560,0.00526577772365676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,12288,5120,0.024138665861553613
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,10240,3584,0.06364799870385064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1536,32,0.005982222242487802
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,8192,768,0.003942222230964237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,32,5120,0.026879999372694228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,256,10240,0.05416799916161431
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,128,2048,0.004975999808973736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,7168,12288,0.03846399982770284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3584,65536,0.12021866109636094
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,768,64,0.0021688888470331826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,10240,128,0.04347733325428433
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,1536,16384,0.33068444993760854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,128,768,0.006001777946949005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,768,768,0.005686222265164058
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,7168,1536,0.005932444499598608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,32,512,0.0025635556214385563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2048,2560,0.0069360000391801195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,16384,6144,0.08403733041551377
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,6144,5120,0.15458399719662136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1024,8192,0.010408889088365767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,512,7168,0.010072000324726105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,7168,12288,0.03504533237881131
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,256,8192,0.010400888820489248
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,65536,12288,0.3893848790062799
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,2048,1024,0.012673777838548025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,10240,65536,12.168238321940104
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,65536,1024,0.029198222690158423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,4096,16384,0.04418311185306973
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,8192,768,0.04767555660671658
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,768,512,0.002532444480392668
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,512,512,0.01200799975130293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,65536,2560,0.063701331615448
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,6144,7168,0.029143999020258587
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,65536,8192,0.14467022154066297
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,1024,256,0.006304889089531369
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,8192,8192,0.2201688951916165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,12288,5120,0.0602968864970737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,512,16384,0.03569599986076355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,5120,12288,0.06254844533072577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2048,6144,0.011086222198274402
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,128,2048,0.006589333216349284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,7168,8192,0.11402222845289443
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,7168,128,0.002512000087234709
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,1024,2560,0.009003555609120263
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,64,2048,0.004903111192915174
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,16384,10240,0.11345333523220485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2560,8192,0.014134221606784396
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,128,10240,0.011815110842386881
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,65536,128,0.01037599974208408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,2048,6144,0.03429422113630507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,8192,2048,0.007283555964628856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,512,2560,0.004870222260554631
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,128,10240,0.02685955497953627
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,512,256,0.0022035555707083806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,768,16384,0.012950222525331708
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,3584,128,0.007635555333561367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,4096,1536,0.04125244418780009
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,32,4096,0.01051111105415556
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,3072,8192,0.042048000627093844
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,5120,1536,0.005561777700980504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,2560,10240,0.015454222758611044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,12288,1536,0.12952889336480036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,64,65536,0.007744000189834171
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3584,65536,0.22575110859341094
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,6144,3584,0.009369778136412302
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,7168,256,0.0052888890107472735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,256,768,0.002826666666401757
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,65536,64,0.004237333519591225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,1024,2560,0.007976888782448238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,64,7168,0.011765333513418833
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,8192,512,0.027669333749347266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,3072,8192,0.015857777661747403
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,128,12288,0.013080000049538083
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,6144,10240,0.024805333879258897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,1024,128,0.005509333478079901
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,5120,768,0.006265777680608962
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,128,4096,0.008048888709810045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,128,64,0.00565422202150027
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,65536,10240,0.08256622155507405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,10240,2048,0.016813332835833233
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,1024,256,0.005575110928879843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,5120,64,0.00886933339966668
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,1536,7168,0.010346666806273991
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,256,5120,0.008832889298597971
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,3584,8192,0.013206221991115145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2048,64,0.005970666805903117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,64,2048,0.006633777585294511
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,12288,128,0.007387555307812161
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,256,3584,0.007997333175606197
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,7168,16384,0.3836142222086589
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,10240,32,0.007267555428875818
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,5120,256,0.008367111285527548
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,768,768,0.0034435554924938413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,512,512,0.0024968888610601425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,512,128,0.005236444373925527
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,10240,512,0.00796622203456031
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,6144,1536,0.006302222195598815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2560,32,0.0021582223061058256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,512,10240,0.006618666566080517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,6144,2048,0.01963644391960568
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,10240,4096,0.015878222054905362
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,5120,16384,0.03602755400869582
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,512,2560,0.014516444669829475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2560,2048,0.05789955457051595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,128,3584,0.007445333732499017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,4096,16384,0.025567111041810777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2048,4096,0.007656888829337225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,768,1024,0.006231110956933763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,64,2048,0.008975110948085785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,32,6144,0.009308444129096137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,5120,16384,0.02973777718014187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,7168,2048,0.023168888356950548
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,10240,1536,0.05922222137451172
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,3072,256,0.01514311134815216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,16384,5120,0.026113778352737427
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,64,10240,0.012059555285506778
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,4096,768,0.0059368887709246735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,3072,2560,0.022967999180157978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1536,768,0.00554666668176651
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,128,768,0.01346666696998808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,32,10240,0.009002666506502364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,2560,4096,0.007627555893527136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,65536,3584,0.09053244193394978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,4096,5120,0.008577778107590145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1536,512,0.003192000091075897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,768,64,0.004927111168702443
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,5120,3584,0.008545777863926357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1024,12288,0.00982844415638182
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,16384,65536,0.3006497753991021
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,12288,2048,0.011729777687125735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,6144,12288,0.04288799895180596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,32,2048,0.004540444248252445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,256,5120,0.01756266587310367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,2048,256,0.005568000177542369
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,16384,1536,0.010215111076831818
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3072,5120,0.009874666730562845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,32,256,0.005626666463083691
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,32,64,0.005569777968857024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,2560,1536,0.016178665889634024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,7168,12288,0.02351555559370253
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,512,3584,0.04180622100830078
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,16384,768,0.10817333062489827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,256,1024,0.0032026666320032547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,65536,12288,0.11635643906063503
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,8192,12288,0.034635557068718806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,256,4096,0.014516444669829475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,512,65536,0.09709155559539795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,512,3072,0.00797155582242542
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3584,3072,0.051585776938332446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,256,65536,0.08797778023613824
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,768,768,0.0058959999846087555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,12288,4096,0.018025777406162687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,2048,16384,0.01995822290579478
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,1536,1024,0.013805333111021252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,256,256,0.015171556009186639
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,7168,6144,0.025535111625989277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3584,64,0.002143111079931259
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,768,256,0.005919999960396025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,8192,1024,0.020598222812016804
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,7168,512,0.005912000106440649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,6144,3584,0.008316444853941599
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,12288,2048,0.02804622385236952
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,65536,16384,0.1643893321355184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1536,256,0.002528888897763358
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,10240,65536,0.34175909890068906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,12288,7168,0.04879200127389696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,768,12288,0.04956266615125868
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,16384,512,0.005433777968088786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,2560,768,0.005678222411208683
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,256,3584,0.008190222084522247
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,10240,256,0.01108444482088089
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,256,2560,0.007355555891990662
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,3584,1536,0.016903999778959487
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,5120,65536,0.11228356096479629
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1536,10240,0.016558221644825406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,65536,512,0.0127688886390792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,7168,6144,0.2913751072353787
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,4096,768,0.0075493330756823225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2048,512,0.006002666635645761
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,32,2048,0.0063635553750726916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,5120,6144,0.018319111731317308
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2048,512,0.009141333401203156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,64,7168,0.010465777582592435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3584,6144,0.01108088923825158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,512,16384,0.04077333211898804
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,12288,3072,0.019916445016860962
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,10240,1536,0.023755555351575215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3072,2560,0.0069413334131240845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,128,65536,0.010046222143703038
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,65536,512,0.013515555196338229
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,12288,65536,0.3123697704739041
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,4096,8192,0.026465778549512226
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2048,6144,0.02015466656949785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,16384,5120,0.026848889059490625
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,65536,5120,0.2127608855565389
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,65536,64,0.024710221423043147
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,128,768,0.007681777907742395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,5120,2048,0.016695110334290397
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2560,5120,0.011333333121405708
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,1024,2048,0.006963555183675554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,256,4096,0.004905777672926585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,5120,256,0.005658666706747479
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,8192,16384,0.02591111097070906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,10240,768,0.023955555425749883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,16384,16384,0.07092710998323229
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,12288,768,0.0069679998689227635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,2560,10240,0.02991644541422526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,16384,1024,0.008667555948098501
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,65536,5120,0.04899466700024075
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,6144,768,0.017816889617178176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,128,65536,0.05083910955323113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,32,1024,0.0029288888391521242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,12288,4096,0.018583110637134977
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,128,256,0.006966222491529252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,7168,10240,0.02648888859483931
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,768,10240,0.021570665968788996
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,6144,3072,0.0939742194281684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,12288,3072,0.22759644190470377
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,128,65536,0.08009777466456096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,6144,10240,0.015736889508035447
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,768,8192,0.0217875556813346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1536,2560,0.0069448889957533936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,1536,2048,0.006656888872385025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,512,6144,0.01959466603067186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3584,12288,0.020478222105238173
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2560,10240,0.02943999899758233
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,16384,768,0.011910222470760345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,768,65536,0.056623107857174344
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,65536,64,0.004831999954250124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,16384,4096,0.015360888507631091
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,64,2560,0.0052560000783867305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,3584,12288,0.016724444097942777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,32,1536,0.0035208890007601846
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,2048,256,0.01478844384352366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,4096,7168,0.013407111167907715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,32,5120,0.008960888617568547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,7168,3584,0.008967999782827165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,512,65536,0.07355378071467082
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,1536,5120,0.008658666577604082
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,512,7168,0.040727999475267194
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,10240,1536,0.007141333487298753
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3072,12288,0.02861066659291585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,256,768,0.0028533333291610083
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,768,65536,0.13267110453711614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1024,1536,0.00794488853878445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,16384,512,0.023373333944214716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,8192,8192,0.04935822221967909
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,7168,65536,0.1483502255545722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,256,2560,0.016924444172117446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,768,2560,0.030007110701666936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,10240,768,0.011252444651391772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1536,64,0.0021786666992637846
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,12288,10240,0.15672355228000218
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,64,512,0.005614222337802251
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,10240,65536,0.14416621790991888
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,8192,6144,0.015839111473825242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,1536,3584,0.01719999975628323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,256,4096,0.00795199970404307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,3072,65536,0.08553777800665961
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,256,512,0.005507555686765247
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,4096,2560,0.0064204442832205035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,32,2560,0.0053093334039052325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,8192,16384,0.026991999811596338
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,512,256,0.0069653333889113525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,128,1536,0.012545777691735161
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3072,256,0.005941333456171884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,16384,3072,0.018389332625601027
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,4096,1024,0.026405332816971674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,8192,5120,0.04746577805942959
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3584,4096,0.008822222550710043
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,65536,1536,0.024001777172088623
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,1536,256,0.015836444165971544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,4096,128,0.005595555735958948
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,768,7168,0.029675556553734675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3584,5120,0.010056888891590966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,2560,4096,0.02405866649415758
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3072,64,0.00221422231859631
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,7168,65536,0.08358310990863377
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,32,65536,0.045055111249287925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,32,6144,0.005355555564165115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,3584,6144,0.02276266614596049
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,512,3584,0.007293333609898885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,5120,7168,0.2341946760813395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,256,3584,0.006991110742092133
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,3072,512,0.018287110659811232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,65536,6144,0.05554666784074572
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,32,12288,0.004945777770545748
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,512,16384,0.038217776351504855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,3072,7168,0.013857777747843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,64,4096,0.00526577772365676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,128,6144,0.00908622228437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,1024,3584,0.013216888739003075
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,7168,768,0.01755288905567593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1024,512,0.0026293332791990707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3072,7168,0.03087110983000861
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,1536,10240,0.012425777812798819
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,32,7168,0.010244444840484196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1024,7168,0.010584000084135266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2048,8192,0.013493333425786761
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,10240,8192,0.02946044339074029
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2048,3584,0.007652444144090016
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,4096,4096,0.01163200040658315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,1536,512,0.014122666584120857
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,5120,2048,0.041533334387673274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,128,65536,0.3135582341088189
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2560,10240,0.017126222451527912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,2048,256,0.0052560000783867305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,65536,4096,0.11098666985829671
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,7168,6144,0.01350844403107961
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,5120,16384,0.14188178380330405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,65536,7168,0.07047644588682386
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,5120,64,0.006501333167155583
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,6144,256,0.005586666779385672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,64,128,0.0022328889204396144
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,8192,64,0.02401955591307746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,768,65536,0.06224711073769463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,8192,768,0.005029333134492238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,32,768,0.002895111011134254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2560,10240,0.035339554150899254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,2048,128,0.014492443866199918
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,6144,12288,0.042507555749681264
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,12288,6144,0.018242667118708294
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,5120,2048,0.013889777991506787
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,768,8192,0.01568800045384301
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,7168,3072,0.007789333661397298
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,768,6144,0.019260444574885897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,16384,256,0.004222222086456087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,16384,128,0.0064311110311084324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,8192,256,0.009322666459613377
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,768,128,0.0021386666016446217
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,256,1536,0.0062871111763848206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,4096,64,0.005277333160241445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3072,7168,0.011887999872366587
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2048,8192,0.014557333456145393
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,1024,65536,0.06500177913241915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,8192,10240,0.14440088801913792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1536,7168,0.010045333041085137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3584,12288,0.01829955478509267
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,1024,768,0.015808888607554965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,1024,12288,0.036115556955337524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,64,6144,0.004620444443490771
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,32,10240,0.04952889018588596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,7168,4096,0.009329777624871995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,7168,128,0.021311110920376245
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,128,768,0.015132443772421943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,4096,3584,0.008291555775536431
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,3072,8192,0.039263112677468195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,12288,8192,0.022912888063324824
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2048,768,0.0031884445084465873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,6144,2048,0.11816355917188857
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,256,3584,0.02068800065252516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,16384,64,0.012411555482281579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,256,768,0.015466666883892484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,2560,3584,0.04405600163671705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,1536,3072,0.016509332590632968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,12288,1536,0.01804266704453362
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1024,4096,0.00795555528667238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,64,2048,0.004551110996140374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,1024,1536,0.014447111222479077
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,1536,2560,0.006909333169460297
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,7168,16384,0.044292443328433566
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,8192,8192,0.01959377858373854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,16384,128,0.016169778174824186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,32,32,0.005625777774386936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,32,2560,0.0042035554846127825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,12288,2048,0.025026667449209426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1024,8192,0.010062221851613786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,65536,16384,0.2279608779483371
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,6144,6144,0.017772444420390658
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,16384,12288,0.07371022303899129
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,10240,1024,0.021333333518770006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,128,16384,0.015799111790127225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,2048,3072,0.018079999420377944
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,6144,65536,7.05462646484375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,5120,1024,0.02850311001141866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,10240,768,0.03295644455485874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,4096,768,0.007697777615653143
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,3584,10240,0.015084443820847405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2560,64,0.0068408888247278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,8192,2048,0.0658586687511868
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2560,512,0.0136835558546914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,3584,65536,0.18099021911621094
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,65536,2560,0.03308444552951389
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,64,16384,0.015574221809705099
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2560,65536,0.061182220776875816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,65536,3072,0.08089600006739299
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,2048,2048,0.008957333034939235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,4096,2048,0.0063279999627007385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,4096,1536,0.004476444588767158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,5120,4096,0.0773128867149353
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,256,512,0.002526222210791376
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,65536,3072,0.030142221185896132
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1024,5120,0.0144213338692983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2048,768,0.003601777884695265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,256,3584,0.012180444267061023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,12288,128,0.015376889043384127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,64,2560,0.010048000348938836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,8192,1536,0.007663110891977946
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,512,32,0.002185777864522404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3072,8192,0.011571555501884885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,64,7168,0.010065777434243096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,4096,128,0.0022728888110982045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,6144,12288,0.042295998997158475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1024,65536,0.09937422143088447
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,10240,4096,0.013980444934633044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,1024,10240,0.19647733370463052
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,12288,2048,0.0110880004035102
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,5120,65536,0.946939574347602
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,768,3072,0.012969777815871768
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,12288,3072,0.023836443821589153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,65536,4096,0.050422221422195435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,10240,2048,0.013282666603724161
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,65536,512,0.06053600046369764
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,65536,128,0.004877333425813251
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,8192,4096,0.11303822199503581
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,16384,1024,0.008385777473449707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,3584,8192,0.014132445057233175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3072,32,0.02000266644689772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,3584,5120,0.02405333353413476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,256,4096,0.0045422220395671
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,5120,6144,0.013744889034165276
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,8192,32,0.008980444735950893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,10240,2560,0.019208888212839764
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,5120,2560,0.019112888309690688
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,8192,768,0.018176888426144917
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,4096,65536,0.07864711019727919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3072,65536,0.8497751024034289
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1536,2560,0.009314667019579146
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,256,2048,0.016827555166350473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,7168,8192,0.019673777951134574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,64,7168,0.013415999710559845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2560,768,0.0028195555011431375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,6144,3584,0.010984888507260216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,4096,2560,0.00869066682126787
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,12288,10240,0.03182755576239692
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,10240,12288,0.10159111022949219
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,65536,12288,0.09517866373062134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,32,10240,0.00664444433318244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,5120,5120,0.05039733317163256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,256,6144,0.0059573331640826324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,65536,6144,0.1209075583351983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,6144,1536,0.005379555539952383
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,65536,3584,0.10516089200973511
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,16384,3584,0.018808888064490426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,256,128,0.008279999925030602
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,64,5120,0.010213333699438307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2048,64,0.0021484444538752237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,768,10240,0.020139555136362713
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,6144,16384,0.04209066761864556
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,3072,256,0.015081778168678284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,256,65536,0.06226933002471924
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3584,12288,0.019189332922299702
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3072,128,0.005552000055710475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,128,8192,0.020061333974202473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,3072,16384,0.04310933417744107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,16384,8192,0.025585777229732935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,12288,16384,0.05330666568544176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3584,10240,0.01628266606065962
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2560,64,0.005552000055710475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,16384,3584,0.01956977778010898
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,6144,64,0.002477333363559511
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,8192,768,0.0041146667467223275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,256,32,0.0021235555824306277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,256,5120,0.013367999758985309
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,32,1536,0.005903111149867375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,512,256,0.0024862223201327855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,2560,12288,0.01683911184469859
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,4096,3072,0.007720888488822513
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,256,128,0.005323555734422471
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,1536,8192,0.0214373336897956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1024,768,0.0033119999700122406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,16384,65536,0.16878044605255127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,768,4096,0.007987555530336168
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,16384,10240,0.08586133188671535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,10240,2560,0.013436444103717804
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,6144,7168,0.16163555781046549
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,10240,512,0.006590222318967183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,12288,2048,0.013274666335847644
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,128,3584,0.011809777882364062
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,32,5120,0.0042133331298828125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,768,6144,0.006549333532651265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2048,256,0.002858666703104973
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,4096,65536,0.07030755281448364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,5120,768,0.0059262220230367445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,65536,1536,0.025015999873479206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,1536,16384,0.01882133384545644
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,768,2048,0.014434667097197639
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,5120,6144,0.013580444786283704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,7168,16384,0.10686400201585557
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,256,8192,0.01000355515215132
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2048,5120,0.008911111288600499
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,6144,128,0.007606222397751278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,2560,2560,0.01957066688272688
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,2048,16384,0.020299555526839364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,8192,2048,0.0184906671444575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,64,768,0.0028124445428450904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,4096,8192,0.1674106650882297
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,3584,65536,0.0872382190492418
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2560,4096,0.009729777773221334
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,32,3072,0.004577777865860197
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,3072,8192,0.02895911203490363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,2048,128,0.00869066682126787
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1536,7168,0.03129422333505418
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,12288,10240,0.030124445756276447
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,7168,4096,0.06599111027187772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,512,32,0.005274666680230035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3072,4096,0.02039377722475264
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,3584,1536,0.018178666631380718
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,8192,2560,0.18648266792297363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,2048,4096,0.008021333151393467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,768,4096,0.01756266587310367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,6144,7168,0.019866666860050626
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,4096,1024,0.02357599967055851
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,128,12288,0.013099555340078143
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,2560,65536,0.06839733653598361
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,64,2560,0.0052871112194326185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,8192,5120,0.031903998719321355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,768,3072,0.018907555275493197
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,10240,12288,0.041377776198916964
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,4096,3072,0.06278844674428304
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,16384,2560,0.04641866683959961
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,512,3072,0.006288888967699475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,4096,12288,0.01973777843846215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,65536,10240,0.10757688681284587
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,8192,6144,0.04131199916203817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1536,10240,0.15176355838775635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,1024,256,0.005140444470776452
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,6144,3072,0.010402667025725046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,12288,768,0.029149333635965984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3072,7168,0.02994133366478814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,128,8192,0.027553778555658128
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,8192,512,0.00833155545923445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,3072,8192,0.011047999891969891
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,8192,32,0.006948444578382704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1536,6144,0.009122666385438707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,1024,1024,0.006260444306664997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,8192,3072,0.029106666644414265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1024,3072,0.009748444788985783
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,16384,6144,0.040355556541019015
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,128,6144,0.010010666317409938
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,128,32,0.0018560000591807896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,6144,12288,0.11924533049265544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,65536,16384,0.30074310302734375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,10240,1536,0.01314222233162986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,512,8192,0.01108000013563368
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,65536,10240,0.20787378152211508
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,768,12288,0.02500444485081567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3584,65536,0.0733226670159234
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,65536,8192,0.07083200083838569
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,128,65536,0.05750311083263821
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2048,768,0.005967111223273807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,5120,65536,0.07170399692323473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,10240,128,0.0021439999755885866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,256,65536,0.1256213320626153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,4096,2048,0.006538666784763336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,16384,2560,0.03493155704604255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3584,2560,0.08465066883299087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,65536,3584,0.07599021991093953
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,1536,2048,0.01553244392077128
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1536,16384,0.013311111264758639
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2560,16384,0.018043556147151522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,7168,7168,0.023224888576401606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,16384,512,0.006976000136799282
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2560,1024,0.006225777582989798
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3072,128,0.0024702221983008915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,10240,16384,0.08282844225565593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,7168,6144,0.02088533341884613
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3584,128,0.0021582223061058256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,8192,768,0.004005333201752769
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,512,6144,0.022520888182852004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,3584,3584,0.017279111676745944
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2560,8192,0.013803555733627744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,512,2560,0.006996444529957241
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,128,1536,0.012912888493802814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,512,1536,0.003620444486538569
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2048,5120,0.09466044770346747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,5120,10240,0.043325334787368774
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,4096,768,0.005893333504597346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,2560,7168,0.009683555199040307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,7168,7168,0.015793777174419828
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,8192,7168,0.016549333930015564
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,768,1536,0.0069511110583941145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,512,32,0.006909333169460297
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,65536,2560,0.08868977758619521
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1536,6144,0.00961866643693712
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,32,12288,0.006271111054552927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,8192,65536,9.07584974500868
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,3584,65536,0.06950133376651339
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,6144,4096,0.024671999944580927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2048,12288,0.011459555890825061
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,2048,65536,0.06559377908706665
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,768,768,0.007275555696752336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1024,5120,0.00965244405799442
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,768,10240,0.011752888560295105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3072,4096,0.007961778177155389
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1536,7168,0.016135111451148987
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,3072,10240,0.046635554896460645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,12288,256,0.003638222399685118
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,7168,256,0.0028213332924577924
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,5120,64,0.005599110904667113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1536,5120,0.009425777528021071
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,8192,128,0.028637333048714533
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,4096,3584,0.00795199970404307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,256,2560,0.0069360000391801195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,65536,3584,0.06948977708816528
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,7168,3584,0.02502577834659153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,12288,65536,0.20553155740102133
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,768,128,0.002346666736735238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,12288,1024,0.010564444793595208
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,6144,8192,0.01585155559910668
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,128,3584,0.007344000041484833
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,12288,3072,0.015463999576038785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2560,1024,0.005895111295912001
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,5120,3584,0.020572443803151447
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,128,12288,0.028620445066028174
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,3072,768,0.017515555024147034
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,65536,12288,0.10378844208187527
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,512,4096,0.007629333270920648
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1024,256,0.0022088889446523455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,16384,10240,0.07324800226423475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,512,768,0.015100444356600443
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1024,128,0.005520888914664586
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,128,768,0.013110222087966071
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,768,1024,0.01624000072479248
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,10240,16384,0.058896885977851025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2048,4096,0.00797333319981893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,4096,2048,0.02479733361138238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,8192,65536,0.13221332761976454
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1536,6144,0.01759199963675605
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,8192,3584,0.012206222448084088
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2048,16384,0.15791200266944036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,65536,5120,0.45360443327162003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,10240,64,0.0028497777465316984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,32,7168,0.009688888986905416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,7168,2560,0.008569777839713627
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,64,32,0.0052666668262746595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,6144,6144,0.013888888888888888
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,5120,12288,0.030379556947284277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2560,12288,0.13113689422607422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,8192,5120,0.01428533262676663
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,10240,3072,0.10902844534979926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,768,512,0.014287999934620328
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,65536,64,0.016089777151743572
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,4096,512,0.0062977779242727495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,10240,16384,0.052729778819613986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,512,1536,0.003871110992299186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,512,64,0.0021386666016446217
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1024,8192,0.00907022257645925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,6144,3072,0.02787289023399353
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3072,2560,0.007650666766696506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,5120,2560,0.019860444797409903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,256,8192,0.02097333305411869
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,2560,2560,0.016523554921150208
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,10240,1536,0.00795733349190818
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,1536,7168,0.009367999931176504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,8192,512,0.006007110906971826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,10240,8192,0.0292577776643965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3584,256,0.0028204443968004654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,4096,16384,0.025049777494536504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,1024,2560,0.0069671107663048645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,4096,65536,0.06672888994216919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2560,3584,0.007167110840479533
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1024,6144,0.01733955575360192
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,2048,2560,0.01552177800072564
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,8192,16384,0.06271111302905612
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,12288,12288,0.05471555391947428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,65536,128,0.026798221800062392
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,768,16384,0.04373866650793287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2560,128,0.005572444448868434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,7168,2560,0.02206666602028741
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,256,1024,0.005947555518812603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1536,7168,0.013487999637921652
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,8192,4096,0.021988444858127173
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,512,128,0.0052693333062860705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1024,5120,0.008819555242856344
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,12288,10240,0.026876444617907207
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,1024,12288,0.014647111296653748
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,2048,128,0.014575110541449653
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3584,2560,0.0151555554734336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,12288,8192,0.05115911033418444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,512,16384,0.04830044507980347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,7168,8192,0.024921778175565932
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3584,64,0.007930667036109501
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,10240,65536,0.20510133107503256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1536,32,0.0023786665664778813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,10240,10240,0.04420444369316101
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1536,5120,0.00830844458606508
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,16384,4096,0.022994667291641235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,8192,7168,0.02126755482620663
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,128,12288,0.023400889502631292
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3584,64,0.002838222309947014
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,1024,12288,0.02882222334543864
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,65536,2560,0.061667554908328585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,128,65536,0.08788000212775336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,5120,65536,0.11194845040639241
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,6144,128,0.014984000060293408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2560,256,0.003173333282272021
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1536,5120,0.01423911088042789
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3584,2560,0.00868800034125646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,4096,32,0.005439110928111606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,8192,7168,0.02299555473857456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,2048,16384,0.03174755639500088
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,4096,2048,0.016549333930015564
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,8192,7168,0.019759999381171334
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,64,6144,0.004944889081848992
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,256,2048,0.00490044429898262
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,8192,3072,0.008645333349704742
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,4096,2048,0.01425244410832723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,6144,2048,0.06477333439721002
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,2560,128,0.014500444134076437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,12288,128,0.01515111161602868
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,1536,5120,0.026014222039116755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,6144,2560,0.00795999997191959
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,64,2560,0.007159111400445302
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,4096,8192,0.015815110670195687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,8192,10240,0.021591999464564856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,2560,512,0.008533333738644918
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1024,768,0.003286222202910317
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,768,32,0.005688888745175467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,4096,5120,0.030990223089853924
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1024,7168,0.011514667007658215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,1536,512,0.00544888898730278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2560,256,0.005571555760171678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,65536,128,0.040991110934151545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,64,8192,0.01070577816830741
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,8192,8192,0.016910221841600206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,6144,7168,0.014818666709793938
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,128,64,0.008311111066076491
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,8192,64,0.0022168887986077205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,65536,2048,0.047725333107842334
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,1536,2048,0.020237333244747586
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,1024,12288,0.013545778062608508
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3072,7168,0.018389332625601027
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,1536,6144,0.009720889230569204
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1024,64,0.002149333349532551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,128,10240,0.012246222131782107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,128,2560,0.004952888935804367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,256,1024,0.008044444852405125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1536,2048,0.011976888610257043
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,7168,3584,0.008310221963458592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,64,4096,0.007710222568776872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,4096,2560,0.02588444451491038
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,7168,6144,0.01478488908873664
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,768,768,0.005583999885453119
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,6144,512,0.024528000089857314
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,256,2048,0.006648889018429651
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,3072,768,0.01720888912677765
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2048,2560,0.005963555640644497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,4096,6144,0.01313511116637124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,128,4096,0.006308444258239534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,7168,768,0.006457777900828256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,16384,12288,0.04998222324583265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,6144,10240,0.05939200189378527
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,128,1536,0.01277599980433782
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,16384,32,0.007315555380450354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,4096,7168,0.013791999883121915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,2048,4096,0.016545777519543965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,16384,8192,0.046979556481043495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,1536,2048,0.019039111004935372
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,256,3584,0.007576000359323289
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,4096,2048,0.03942755526966519
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2048,10240,0.03055644366476271
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,16384,2048,0.0322053333123525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,4096,5120,0.009029332962301042
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1536,10240,0.026362667481104534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,65536,6144,0.21082844999101427
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,6144,128,0.002247111043996281
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,16384,128,0.005937777873542573
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2560,4096,0.020216888851589628
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,5120,5120,0.020370667179425556
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3584,1024,0.00943288869327969
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2560,5120,0.023437332775857713
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,5120,5120,0.012157333393891653
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,16384,16384,0.1523546642727322
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,1024,512,0.0053137776752312975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,256,3072,0.019173334042231243
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,65536,768,0.023489778240521748
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,10240,64,0.006175111151403851
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,128,4096,0.014838222000333997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,128,2048,0.006670222100284364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,32,6144,0.00665155549844106
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,6144,256,0.00563733321097162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,7168,65536,0.20879467328389487
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,128,5120,0.02072355482313368
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,7168,2048,0.006603555546866522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,256,6144,0.009394666386975182
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,2048,6144,0.059402664502461754
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,6144,2048,0.007622222105662028
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,2560,5120,0.02242311173015171
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,8192,12288,0.6122462484571669
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,4096,1536,0.021673777037196692
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,12288,512,0.09310666720072429
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,4096,65536,0.09280444516075982
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,8192,65536,0.18613778220282662
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,2048,8192,0.19922488265567354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1024,65536,0.03540088733037313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3072,7168,0.011104889214038849
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,512,64,0.0059262220230367445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,256,64,0.0021973333011070886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,5120,5120,0.035655998521380954
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1536,10240,0.014514666464593677
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,2560,12288,0.017335111896197002
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1024,5120,0.010680889089902243
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,8192,12288,0.041280001401901245
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,5120,128,0.005591111050711737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,7168,3072,0.010392000277837118
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,4096,5120,0.021560000048743352
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,10240,10240,0.030798223283555772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,4096,12288,0.019838222199016146
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,65536,10240,0.10336177878909641
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,12288,32,0.019976889093716938
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,6144,3072,0.026206221845414903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,65536,4096,0.037385778294669256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,2560,3072,0.007649777664078607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,12288,768,0.011074666347768573
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,8192,256,0.00249422217408816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,6144,1024,0.005997333261701796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2560,128,0.00221422231859631
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,512,64,0.005585777676767773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,768,1536,0.0081386665503184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,7168,1024,0.005064888960785336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,65536,4096,0.052186667919158936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,256,256,0.0053022222386466126
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,256,2048,0.01699466672208574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,32,512,0.006609777609507243
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,2560,2560,0.0069368887278768755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,12288,3584,0.030893332428402368
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,128,7168,0.005647111270162795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,12288,12288,0.13350844383239746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,8192,10240,0.05702400207519531
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,65536,256,0.020399999287393358
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2048,128,0.006075555665625467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,12288,4096,0.2346097893185086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,3584,8192,0.0295688874191708
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,5120,768,0.01962844365172916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,256,7168,0.022321777211295232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,10240,3072,0.10118489132987128
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,128,6144,0.016898666818936665
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3584,4096,0.009223110973834991
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,6144,12288,0.02797244323624505
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,5120,1024,0.00602311102880372
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,2048,12288,0.05225511060820686
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2560,12288,0.022465777066018846
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,64,3072,0.010063110954231685
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,65536,3072,0.10316089126798843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,12288,768,0.0069777775141927935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,6144,3584,0.017667555146747164
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2048,5120,0.009852444132169088
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,65536,4096,0.1989039977391561
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,32,6144,0.022100443641344707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,12288,1536,0.009922666682137383
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,12288,3584,0.0162808895111084
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,5120,3072,0.012819555898507437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,6144,2560,0.011774222056070963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,2048,16384,0.020633776982625324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,6144,32,0.006287999865081575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3072,8192,0.015145777000321282
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,64,128,0.005966222120655908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,128,2560,0.005636444522274866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,2048,6144,0.022879999544885423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,8192,512,0.03140177660518222
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,2560,4096,0.026991999811596338
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,2048,2560,0.022367111510700647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,32,8192,0.019588443968031142
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,4096,64,0.005583111196756363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,12288,10240,0.046794666184319385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,4096,256,0.002439111057255003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,256,8192,0.011359111302428775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,10240,2048,0.010415111151006486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,16384,4096,0.01909955508179135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,7168,65536,0.3029786745707194
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,4096,3584,0.007671111159854465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,16384,12288,0.06047911114162869
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,512,12288,0.023708444502618577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,1024,512,0.005535111245181825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,64,1024,0.0031795555518733132
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,1024,1536,0.014500444134076437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,64,5120,0.005068444543414646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,10240,512,0.007048889166778988
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,8192,64,0.005517333332035277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,10240,512,0.003448888866437806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,12288,256,0.006265777680608962
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,128,256,0.00629688882165485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,65536,512,0.013139555851618448
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1024,3584,0.010664000279373594
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,7168,2048,0.007145778172545963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,6144,768,0.005526222288608551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,3072,1536,0.027172444595230952
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,10240,6144,0.01993777851263682
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,64,6144,0.0059262220230367445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,128,10240,0.011491555306646558
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1536,1536,0.006573333508438534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,1024,16384,0.016169778174824186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,8192,7168,0.05405422382884555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,8192,3072,0.011030221978823343
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,256,6144,0.018588445252842374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,7168,768,0.004316444612211651
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,32,10240,0.00489155575633049
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,768,2048,0.006592000110281839
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3072,7168,0.012454222473833295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,7168,2048,0.008632000121805403
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,768,3072,0.010738666686746808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,10240,16384,0.03957155678007338
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,4096,4096,0.0088755554623074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,256,6144,0.008973333570692275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,8192,2560,0.011754666765530905
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,256,3584,0.004852444347408083
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,2048,256,0.005234666582610872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,16384,4096,0.19228622648451063
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3584,12288,0.020587555236286588
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,12288,128,0.033575998412238225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,16384,768,0.041249778535630964
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,6144,32,0.018599111172888014
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,512,3584,0.0052728888889153796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,4096,512,0.005222222043408289
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,3072,8192,0.02515022291077508
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,10240,3072,0.02070488863521152
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,5120,1024,0.01683999929163191
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,16384,16384,0.10836622450086807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,768,1024,0.003963555726740095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,512,16384,0.007986666427718269
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,6144,1536,0.03975111246109009
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,64,128,0.002163555473089218
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1024,2048,0.0070257774657673305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,128,12288,0.1858044465382894
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,8192,128,0.02827111217710707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,768,16384,0.10132533311843872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2560,10240,0.062104887432522245
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,256,10240,0.012406222522258759
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,12288,128,0.0032195556494924757
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,512,10240,0.011863999896579318
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,64,3584,0.008045333127180735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,32,7168,0.010467555787828235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,16384,256,0.006272888845867581
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,5120,10240,0.020233778489960563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,5120,4096,0.010020444790522257
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,6144,65536,0.2652906576792399
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,32,65536,0.04547911220126682
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,512,1536,0.006703111032644908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,3072,128,0.005218666460778978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,32,8192,0.016502221425374348
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,256,1536,0.008646222452322641
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1024,65536,0.06372444497214423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,8192,8192,0.16624622874789768
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,16384,7168,0.05134222242567274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,12288,2048,0.01942933268017239
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,8192,8192,0.016055110428068373
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,512,2560,0.007065777977307637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,12288,6144,0.11639289061228435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,1536,12288,0.013464888764752282
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,512,1536,0.0063386667105886675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,3584,256,0.024522667129834492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,12288,1536,0.009633777870072259
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,32,7168,0.03468799922201369
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,512,10240,0.00871288859181934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,16384,65536,0.16645156012641057
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,64,2560,0.004913777940803104
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,65536,6144,0.05624177720811632
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,16384,16384,0.0904995534155104
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,5120,4096,0.02882400155067444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,1536,1024,0.03568711214595371
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,12288,4096,0.01940533353222741
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,3072,3072,0.025548444853888616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,128,1536,0.006278222219811545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2048,3584,0.011086222198274402
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,5120,5120,0.030051555898454454
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1024,7168,0.010794666906197866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,1536,2560,0.04155200057559543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,256,12288,0.015449777245521545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,32,3584,0.006031111296680238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,65536,2560,0.09343822134865655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,7168,1024,0.003944000022278892
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3584,2560,0.010411555568377176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,6144,4096,0.0090844440791342
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,10240,128,0.015645333462291293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,2048,12288,0.03126488791571723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,128,12288,0.013422222601042854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,64,768,0.005772444523043103
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,64,2560,0.009392889009581672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,128,6144,0.009103111094898647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,12288,12288,0.0483528905444675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2560,128,0.0022275555464956495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,4096,12288,0.03896888759401109
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2560,10240,0.016415110892719693
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,10240,65536,0.150144894917806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,3584,128,0.01863466699918111
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1024,3072,0.005254222287072076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,6144,1024,0.0165955556763543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,256,8192,0.005977777971161737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,2048,6144,0.009920000202125972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,65536,1536,0.04264800084961785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2048,1536,0.005952000204059813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,3072,1536,0.019300444258583915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2048,7168,0.010149333212110732
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,65536,256,0.2200995551215278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,7168,32,0.002855111120475663
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3584,3584,0.00776088900036282
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,256,1024,0.012472000386979846
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2560,1024,0.00646577775478363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,128,6144,0.009544889132181803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3584,3072,0.1046524445215861
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,128,2048,0.009696000152164036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2048,4096,0.010131555298964182
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,7168,6144,0.027679999669392902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,768,3584,0.0075031113293435835
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1536,3584,0.00794577764140235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,10240,2048,0.012940444052219391
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,2048,8192,0.02734666731622484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,6144,1536,0.024864888853496973
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,6144,5120,0.016883555385801528
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,65536,64,0.004519999855094486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1536,7168,0.010423999693658618
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,1536,10240,0.012053333222866058
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2560,3072,0.007802666889296637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,5120,65536,0.07485333416197035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,2560,128,0.007370666497283512
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2560,768,0.0029235554652081597
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,768,5120,0.006689777804745569
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3072,6144,0.012496000362767113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,64,512,0.006970666348934174
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,65536,256,0.013150222599506378
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3584,256,0.008022222254011366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,8192,32,0.0029271110478374693
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,1536,10240,0.02424088948302799
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,1536,1536,0.0158897770775689
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,1024,128,0.01459111107720269
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,12288,7168,0.027719999353090923
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,16384,2560,0.030058665408028498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,10240,10240,0.03916622201601664
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,512,2048,0.006619555668698416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,7168,3584,0.029119110769695703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,4096,7168,0.014136888914638095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,5120,1024,0.006274666637182236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,768,65536,0.07208444674809773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,12288,5120,0.016530666086408827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3072,5120,0.011376000112957425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,12288,6144,0.02642755541536543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,512,65536,0.15226756201850042
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3072,2048,0.010040889183680216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,7168,1536,0.01145066652033064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1536,1024,0.005917333480384614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2560,6144,0.011457777685589261
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,768,12288,0.09672088755501641
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,65536,16384,0.29965244399176705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,128,3584,0.008000888758235509
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,6144,2048,0.009159111314349705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,4096,10240,0.33339911037021214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,65536,3584,0.04231999980078804
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,10240,256,0.05849244197209676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,12288,12288,0.03144266539149814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,5120,7168,0.04070933328734504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,2560,768,0.005995555470387141
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3584,2560,0.006538666784763336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,2560,10240,0.08187022474077013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,10240,3072,0.20991644594404432
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,12288,8192,0.0277786652247111
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,128,3584,0.005602666487296422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,7168,5120,0.013882666826248169
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,6144,6144,0.01278133359220293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1024,8192,0.01223111069864697
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,8192,2048,0.010695999695195092
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,2048,5120,0.023708444502618577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1536,512,0.005565333283609814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,12288,5120,0.01649777756796943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,768,12288,0.013415999710559845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2048,1024,0.025607110725508794
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,65536,1536,0.022651554809676275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,2560,512,0.023036445180575054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,6144,512,0.015797333584891427
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,7168,128,0.00250133333934678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,1536,7168,0.033145778708987765
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,65536,3584,0.07657600111431546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,65536,128,0.01608622239695655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,65536,6144,0.06809777683681913
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,768,5120,0.008653333617581261
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3584,7168,0.014367111855083041
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2560,12288,0.046465777688556247
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,768,2560,0.0070568886068132185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,10240,64,0.05316266748640272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,32,1024,0.005980444451173146
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,768,8192,0.01146755533085929
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,3584,4096,0.01867111192809211
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,768,8192,0.028106666273540918
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,8192,512,0.005944889038801193
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1024,512,0.004032888760169347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,3584,6144,0.009696000152164036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1024,2560,0.009300444689061906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,768,8192,0.006979555719428592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,4096,5120,0.01310222182008955
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,6144,6144,0.0380968881977929
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,32,128,0.002208000048995018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3072,2048,0.006639110959238476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2560,6144,0.011529777612951068
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,65536,768,0.036899556716283165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3584,64,0.0052791109515561005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,7168,3584,0.013539555172125498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,5120,12288,0.04979377653863695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,65536,8192,0.17850311597188315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,512,128,0.014583110809326172
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,1024,4096,0.00775733341773351
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,1024,10240,0.021426666114065383
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,3072,16384,0.02105955613984002
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,3584,5120,0.02137511140770382
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,512,32,0.00221422231859631
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1024,3072,0.00629333323902554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,5120,32,0.0022675556441148124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,3072,2048,0.031488001346588135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,65536,256,0.19652533531188965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,12288,65536,0.15163644154866537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,512,768,0.010091555615266165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,10240,10240,0.027335110637876723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,4096,1536,0.01940622263484531
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,10240,10240,0.03401155604256524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3072,6144,0.011844444606039258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,6144,10240,0.0831022196345859
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,512,5120,0.005671999934646819
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,768,7168,0.01791288952032725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3584,16384,0.023139556248982746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,32,4096,0.004618666652176115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,768,1024,0.014167999227841696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,64,65536,0.008734222087595198
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,10240,32,0.007711999946170383
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,768,768,0.015607110328144498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,12288,512,0.008077333370844523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,8192,128,0.005959110955397288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2560,32,0.006591111007663939
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,12288,4096,0.021154666940371197
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,65536,5120,0.04566755559709337
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,8192,1536,0.030561778280470107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,5120,256,0.002546666603949335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2560,6144,0.01109955542617374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,16384,4096,0.04184000028504265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,1024,6144,0.018290667070282828
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,7168,12288,0.02276800076166789
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,65536,3584,1.452677302890354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,256,10240,0.0564933353000217
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,7168,16384,0.060257779227362744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,5120,4096,0.009000000026490953
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1536,2048,0.004622222234805425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,10240,2048,0.010853333605660332
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,65536,3584,0.15228977468278673
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,2560,65536,0.06429066922929552
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,16384,4096,0.025429333249727886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2560,512,0.0035911111368073356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,10240,1024,0.0074337778819931885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,10240,5120,0.04054488738377889
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,65536,7168,8.464063008626303
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,65536,3584,0.08707111411624485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,7168,65536,0.08850933445824517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1024,4096,0.048933333820766874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,32,65536,0.008053333395057255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,2048,512,0.01348444405529234
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2048,1536,0.006414222220579783
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2560,3584,0.027232888672086928
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,512,4096,0.0058808889653947614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,5120,7168,0.028096887800428603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,256,7168,0.04212355613708496
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,16384,2048,0.06100355254279243
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,65536,256,0.007323555648326874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,65536,3584,0.03540800015131632
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,6144,32,0.005933333188295364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,3072,7168,0.014389332797792224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,256,64,0.0052684446175893145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,768,768,0.006597333484225803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,32,65536,0.05428444345792135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,6144,64,0.0025155554629034465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,2560,256,0.021687111920780603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,6144,768,0.01796800063716041
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,6144,256,0.028689778513378564
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,6144,768,0.005937777873542573
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,6144,2048,0.018230222993426852
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,3072,4096,0.00830666638082928
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,3072,4096,0.010050666828950247
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,8192,256,0.02758755617671543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,65536,128,0.019882667395803664
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,5120,10240,0.09301422039667766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,65536,12288,0.11812622017330593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,256,768,0.0028400001012616684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,4096,5120,0.012458666331238218
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,16384,64,0.006607111129495833
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,6144,1536,0.017274666163656447
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,512,128,0.00249422217408816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,7168,128,0.0021235555824306277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,128,10240,0.016947555873129103
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,10240,65536,0.606552865770128
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,6144,1536,0.0181368887424469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,16384,7168,0.16574222511715359
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,1536,256,0.01479822231663598
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,256,64,0.005588444570700328
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,3584,2560,0.016490666402710807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,2048,512,0.005897777775923411
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,128,4096,0.02956177790959676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,3584,12288,0.01811555524667104
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,6144,8192,0.01921688848071628
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,2560,12288,0.018759111563364666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2048,2048,0.007723555796676212
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,12288,12288,0.07729511128531562
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2560,10240,0.19849688476986357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1024,5120,0.016685333516862657
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,64,1024,0.0028568889117903183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,2560,1024,0.005952888892756567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,8192,10240,0.2505706681145562
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,3584,5120,0.044553776582082115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1024,7168,0.01349688900841607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,10240,6144,0.02553244431813558
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,512,7168,0.00555377784702513
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,768,256,0.005239999956554837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,8192,1024,0.009716444545321995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3072,256,0.0025884444928831523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,8192,65536,2.3065812852647567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,64,16384,0.04485244552294413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,4096,128,0.007997333175606197
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,16384,64,0.006799110935793982
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,65536,1024,0.1291982200410631
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,512,128,0.009964444571071202
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2560,64,0.0019208888212839763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,7168,5120,0.016973333226309884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,4096,64,0.005582222094138463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,3072,65536,0.07371199793285794
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,512,7168,0.010330666270520952
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,3072,5120,0.02612000041537815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,3072,65536,0.12853777408599854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1536,32,0.0029271110478374693
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,64,12288,0.005576000031497743
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,64,2560,0.009296000003814697
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3584,2560,0.0705359975496928
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,128,3584,0.00756888919406467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,128,3584,0.028767999675538804
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,65536,10240,0.11595911449856228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,8192,2048,0.010460444622569613
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,2560,7168,0.01351288871632682
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,4096,2048,0.0058257778485616045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,128,2560,0.006953777538405524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1024,1024,0.008798222574922774
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,128,128,0.00526577772365676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,256,4096,0.00811644477976693
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,65536,3072,0.12411199675665961
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2048,3584,0.007275555696752336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2048,1536,0.004874666531880696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,5120,512,0.02994489007525974
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1536,1536,0.005959110955397288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,256,256,0.0052728888889153796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3584,1536,0.004228444563017951
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,512,768,0.0028328889360030494
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,65536,16384,0.13996177249484593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,7168,128,0.005675555517276128
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,3584,128,0.02772177755832672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,256,2048,0.00961866643693712
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,1536,768,0.015838222371207345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,512,65536,0.0557804438802931
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,16384,1024,0.07946755488713582
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2560,128,0.005894222193294101
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,2048,7168,0.03664533297220866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1536,3584,0.006664888726340399
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,128,256,0.0024746666765875286
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1024,128,0.007982222570313348
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,5120,1536,0.0372284452120463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,3584,10240,0.027882665395736694
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,512,7168,0.00998222248421775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2048,768,0.0058711109062035876
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,3584,256,0.015107555521859063
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,256,2560,0.004834666848182678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,128,12288,0.013463999662134381
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,1024,6144,0.009001777403884465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,32,12288,0.005220444252093633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,128,2048,0.006865777903132969
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,512,6144,0.06052977508968777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,5120,65536,0.6107475492689345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,16384,7168,0.057714667585161
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,8192,10240,0.0221368885702557
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,6144,256,0.017148445049921673
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,128,512,0.003206222214632564
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,256,2048,0.00831822223133511
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,128,128,0.012804444465372296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,10240,4096,0.013879111243618859
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,7168,256,0.016953777935769822
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,10240,768,0.006642666541867786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,128,1536,0.012450666891203986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,5120,512,0.0069297779765393995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,768,3584,0.005949333310127258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,12288,65536,0.25338488154941136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,768,64,0.002157333410448498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3584,512,0.003491555651028951
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,32,16384,0.015440000428093804
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,12288,2560,0.012275555895434486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1024,10240,0.012116444607575735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,1536,6144,0.00903911143541336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,8192,512,0.006014222072230444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,4096,128,0.014894222219785055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,32,1536,0.006618666566080517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,6144,512,0.015848888291252982
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,8192,1536,0.02163377735349867
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,512,768,0.008072888685597314
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,65536,768,0.01687466601530711
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3072,5120,0.009198221895429824
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,128,4096,0.005943999936183293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,1024,4096,0.014137778017255994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2560,768,0.005919111271699269
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,16384,2048,0.011333333121405708
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1024,768,0.00629066675901413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,256,12288,0.014182221558358936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,3584,1024,0.006164444403515921
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,64,2048,0.006659555352396435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,5120,128,0.002161777681774563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1536,65536,0.05845866600672404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,16384,5120,0.021572444174024794
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,12288,256,0.01756444407833947
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,7168,6144,0.013129777378506131
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1536,256,0.002566222308410539
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,3072,10240,0.09008889065848456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,6144,1536,0.005381333331267039
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,512,10240,0.010574222438865237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,65536,128,0.20432533158196342
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,768,65536,0.12942043940226236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,2048,512,0.013425778183672162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,32,4096,0.00795377790927887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,128,7168,0.011232000258233813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,5120,768,0.01835644410716163
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,2560,3072,0.09242133299509685
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,64,10240,0.005912000106440649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,65536,3584,0.11808089415232341
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,8192,7168,0.037666665183173284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,12288,3584,0.018211555149820115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3584,12288,0.01851022243499756
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,6144,16384,0.025422222084469263
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,1536,16384,0.018185777796639335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,7168,65536,0.13043999671936035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1024,10240,0.014194667339324951
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,4096,5120,0.010651555326249866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3584,10240,0.016329778565300834
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,256,64,0.005626666463083691
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2560,4096,0.010085333552625444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,3584,2560,0.036157333188586764
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,2048,6144,0.009011555876996782
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,5120,4096,0.013845333622561561
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2560,16384,0.017892445127169292
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,7168,8192,0.05515999926461113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2560,6144,0.013225778109497495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,5120,12288,0.03498844305674235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,12288,512,0.011824000212881299
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,4096,1536,0.014264000786675347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,8192,7168,0.03293155630429586
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,10240,4096,0.01385155568520228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,6144,2560,0.007722666694058313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,12288,8192,0.03602755400869582
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,512,768,0.006038222048017714
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,3072,7168,0.024079110887315538
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,5120,12288,0.031199110878838435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,5120,3072,0.01237155579858356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3584,128,0.005602666487296422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,1536,65536,0.17879111236996123
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2048,12288,0.06515822145673963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,5120,8192,0.017842666970358956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,64,6144,0.022113778524928625
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,256,512,0.005734222216738596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3072,512,0.0029057777590221832
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,65536,2560,0.06385333008236356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,256,2048,0.006705777926577463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,65536,6144,0.1470293336444431
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2048,65536,0.059897780418395996
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,64,768,0.0028408887899584244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,6144,128,0.015138667490747241
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,512,128,0.0021333334346612296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,2048,2048,0.01444533301724328
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,3584,65536,0.09823733568191528
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,65536,32,0.012305777933862476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,7168,65536,0.17416889137691924
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,768,2048,0.014450665977266101
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,256,7168,0.009983110758993361
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2560,6144,0.011576000187132092
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3072,3584,0.007718222008811102
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,65536,1024,0.015484443969196744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,2048,3584,0.007307555940416124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,3584,8192,0.02514222264289856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1536,6144,0.02773333258099026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1024,16384,0.021031111478805542
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,12288,768,0.006991110742092133
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,7168,3072,0.09889866908391316
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,256,512,0.007251555720965068
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1536,2048,0.019652444455358718
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3072,5120,0.009878222313192155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,12288,256,0.01723377737734053
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,65536,16384,0.7479759852091471
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,6144,12288,0.0306986669699351
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,4096,65536,0.08131644460890028
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,1024,3584,0.007711999946170383
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,65536,10240,1.4492603937784831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,2560,2560,0.006903111106819577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1024,6144,0.015044444137149386
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,10240,10240,0.16974577638838026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,16384,3584,0.02978666623433431
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,2048,128,0.01443822185198466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,128,2560,0.007054222126801808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,128,64,0.0018213333355055915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,768,3584,0.007613333563009898
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,1024,1536,0.013993778162532382
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1536,8192,0.00944266633854972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,1024,1024,0.006140444427728653
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,512,2560,0.007939555578761632
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,1536,4096,0.09322133329179551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,768,512,0.006909333169460297
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,1024,8192,0.010681777364677854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,128,3584,0.0079644446571668
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,3584,3072,0.022520888182852004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,256,10240,0.006278222219811545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2560,1024,0.003245333416594399
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,7168,128,0.015152000718646579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,128,7168,0.005564444594913059
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,7168,8192,0.022451554735501606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,64,512,0.002528888897763358
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,512,12288,0.007287110719415877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1024,3072,0.005598222215970357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,4096,16384,0.05648889144261678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,10240,256,0.017633777525689866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,128,256,0.005590222362014983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3072,64,0.0018862223045693503
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,768,6144,0.009278222090668148
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3072,768,0.009015999734401703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,2560,12288,0.04545955525504219
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,5120,128,0.0025235555238193935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,512,256,0.0025111111915773815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,7168,8192,0.01625333395269182
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2048,512,0.0028782222006056043
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,8192,2048,0.07100622521506415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,12288,16384,0.05315022336112129
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,3584,512,0.015224888920783997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,6144,4096,0.015269334117571512
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,1536,8192,0.02383288906680213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,5120,65536,0.12783732679155138
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,128,128,0.012160888976520963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,65536,1536,0.03167288833194309
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2560,1536,0.006339555399285422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2560,3584,0.09641866551505195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,6144,65536,0.21650311681959364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,3584,2048,0.036278221342298723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,65536,4096,0.05152444375885857
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,8192,3072,0.012800888882742988
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,2048,512,0.005379555539952383
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,7168,2560,0.016143111719025504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,1536,2560,0.006990222467316522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,10240,256,0.006440889090299606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,6144,16384,0.04495022363132901
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,128,16384,0.027648000253571406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,512,6144,0.0110871113008923
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,16384,2560,0.030265778303146362
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1536,8192,0.013091555900043912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,2560,1536,0.026153778036435444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,32,3072,0.008018666671382057
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,5120,128,0.006660444455014334
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,256,16384,0.03663288884692722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,5120,2560,0.013002666334311167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1024,2560,0.017158221867349412
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,1024,10240,0.011911999848153857
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,65536,10240,0.1067333353890313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,32,3584,0.007619555625650618
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,7168,768,0.006188444379303191
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,16384,3584,0.0270462218258116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,16384,10240,0.06406666835149129
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3584,2560,0.009447111023796929
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,1536,7168,0.019631110959582858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,768,768,0.005624889085690181
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,7168,1024,0.007735999921957652
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,8192,16384,0.1258231136533949
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,10240,768,0.0070159998204973005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,12288,512,0.020757334099875558
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1024,1536,0.0064195555945237475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,5120,7168,0.012124444047609964
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3072,16384,0.02128177715672387
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,768,4096,0.008048888709810045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,16384,3072,0.016940444707870483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,6144,1536,0.00637511122557852
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,10240,65536,0.14113067256079778
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,5120,6144,0.014409777190950183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,16384,65536,0.39881155225965714
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,256,12288,0.008421333299742805
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,5120,10240,0.05544266766972012
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,8192,1024,0.01902755598227183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,2560,768,0.005961777849329843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,16384,10240,0.06826666990915935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2048,64,0.0022008888837363985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3072,3072,0.007424888511498769
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,128,768,0.013224889006879596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,10240,1024,0.045287112394968666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,768,5120,0.008325333396593729
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,64,768,0.0029173334025674393
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3072,5120,0.009221333596441481
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,3584,2048,0.006495999793211619
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,768,8192,0.05502400133344862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,16384,12288,0.05549777878655327
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,65536,5120,0.12010222011142307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,256,512,0.005572444448868434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,5120,5120,0.04247377647293938
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,8192,2560,0.010600888894663917
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,4096,32,0.006611555400821898
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,7168,5120,0.013923555612564087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1024,768,0.011396444506115384
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,768,8192,0.010671999719407825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,2560,7168,0.038534220721986555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,12288,2048,0.031202667289310034
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,4096,3584,0.027790221903059218
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,16384,6144,0.02997066577275594
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,2048,8192,0.07585599687364367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,8192,12288,1.572712050543891
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,1024,10240,0.015127999915017022
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,7168,3584,0.011351111034552256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,5120,64,0.002526222210791376
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,5120,1024,0.015817777978049386
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3584,5120,0.012465777496496836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,768,1536,0.006315555423498154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,4096,256,0.005232000102599462
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,512,2560,0.007497777541478475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1536,3072,0.008664000365469191
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2048,768,0.006234666539563074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,512,32,0.0058702222175068325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,5120,8192,0.018453333112928603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2048,4096,0.00794666674402025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,256,256,0.0021502222451898786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,5120,768,0.004224888980388641
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,65536,2560,0.07201600074768066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,2048,2560,0.007266666326257918
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,3072,65536,0.06863644387986925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1024,64,0.0038648889296584656
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,3072,10240,0.030292444758945044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,6144,128,0.014855111638704935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3584,256,0.0024871110088295406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,8192,768,0.00591555568906996
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1024,16384,0.01478488908873664
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,65536,3072,0.08279911014768812
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,256,512,0.011756444142924415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,3072,6144,0.010414222048388587
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,32,768,0.0028648889727062653
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,12288,1024,0.012116444607575735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3072,4096,0.03321422139803568
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,512,3584,0.016189333465364244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,256,12288,0.029782222376929388
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1024,128,0.005610666755172942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,7168,12288,0.02370044423474206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,64,512,0.002479111154874166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,65536,1024,0.1578391128116184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,65536,10240,0.20345421632130942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,5120,7168,0.01553600033124288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,8192,1024,0.021047110358874004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,32,5120,0.005647999958859549
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,8192,5120,0.05589955382876926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,4096,10240,0.018726223044925265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,512,7168,0.01862044466866387
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,7168,3584,0.03091288937462701
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,128,16384,0.005964444329341252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,4096,3584,0.020819556381967332
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,10240,8192,0.12316444185045029
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,10240,2560,0.03053599927160475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,2048,256,0.0052693333062860705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,256,8192,0.04481066597832573
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,16384,768,0.03808533483081394
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3584,768,0.0032257777121331957
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,512,128,0.0021359999146726397
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,10240,4096,0.10919200049506293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,32,64,0.0052879999081293745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,65536,10240,0.16178754965464273
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,64,768,0.0029057777590221832
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,12288,12288,0.03505511085192362
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,65536,1024,0.26350310113694936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1024,7168,0.0073724447025193115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,12288,1024,0.04262488749292162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,5120,2560,0.009001777403884465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,16384,64,0.014843554960356818
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3584,2048,0.007921777665615082
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,4096,8192,0.026534222894244727
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,5120,7168,0.013479111095269522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,64,6144,0.049913777245415576
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,7168,2560,0.02366399930583106
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,2560,5120,0.00830844458606508
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,64,3072,0.007640889121426477
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,256,10240,0.011757333245542316
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,2560,512,0.006616888774765863
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3584,3584,0.03141511148876614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,3072,8192,0.012634666429625617
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,16384,5120,0.02166844407717387
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,1536,65536,0.08400977982415093
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,768,1024,0.0036275556517971885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,768,16384,0.011410666836632622
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,6144,8192,0.016607999801635742
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,6144,12288,0.23602755864461264
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,1536,12288,0.09684799777136908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,6144,7168,0.017143999536832173
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2560,2048,0.010075555907355415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,4096,10240,0.02684266699684991
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1024,12288,0.011577777564525604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,6144,256,0.007864889171388414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,65536,2048,0.08133066362804837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,5120,3584,0.12255733542972141
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,64,2048,0.019683554768562317
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,4096,12288,0.020125332805845473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,7168,512,0.009681777821646797
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,512,10240,0.01723377737734053
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,8192,4096,0.019858666592174105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,6144,7168,0.013965333501497904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1536,6144,0.009413333402739631
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,12288,7168,0.023994666006830003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,16384,1536,0.010767999622556897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,4096,10240,0.019946666227446664
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,16384,2048,0.01146933353609509
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1024,12288,0.013187555803192986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2560,65536,0.22318132718404135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2048,128,0.009407999614874521
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,256,512,0.005612444546487596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1536,7168,0.011067555182509951
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,3584,65536,0.16220000055101183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,3072,7168,0.02297244469324748
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,768,1024,0.006617777877383762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,16384,1024,0.008390222158696916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,8192,128,0.005620444400442972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,6144,4096,0.033542222446865506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,1024,768,0.01551555593808492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1536,3072,0.006169777777459886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,5120,6144,0.017832888497246634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,5120,2048,0.018570666511853535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,768,768,0.015619556109110514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,32,512,0.005952888892756567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,4096,3072,0.02528000043498145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,2048,3072,0.022288888692855835
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2560,128,0.00591555568906996
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,3072,768,0.00591466658645206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,6144,1536,0.006719111154476802
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,768,3072,0.00739911115831799
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,256,4096,0.007954667011896769
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,65536,768,0.03022222386466132
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,3584,12288,0.01734133395883772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,5120,32,0.002534222271707323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,768,12288,0.018295110927687753
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,64,32,0.0018800000349680583
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,3072,65536,0.0925111108356052
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,2048,7168,0.023287110858493384
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,128,768,0.0075333333677715724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,128,65536,0.018554665976100497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,256,3584,0.0053137776752312975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,4096,2560,0.10050044457117717
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,65536,2048,0.026857778429985046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,10240,1536,0.01089244418674045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,12288,16384,0.05669599771499634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,8192,1024,0.006284444282452266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,768,12288,0.024841777152485315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,12288,1024,0.0069413334131240845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1024,3072,0.005615111026499007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1536,128,0.006009777800904379
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,3072,128,0.014870221416155497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,8192,1024,0.004233777936961916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,2048,128,0.015580443872345818
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,6144,16384,0.49575646718343097
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,6144,1536,0.008240000241332585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3584,4096,0.019096000327004325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,512,12288,0.022890667120615642
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,3072,512,0.00556355549229516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,5120,512,0.023942222197850544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,512,4096,0.007992888490358988
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,7168,128,0.015379556351237826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,768,65536,0.08106044265958998
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,32,1024,0.006291555447710886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,128,65536,0.05484355489412943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2560,128,0.00590844452381134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,65536,16384,0.1455333365334405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,128,12288,0.013184000220563678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,64,6144,0.0052720002002186244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,1536,3072,0.007345777418878343
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,10240,768,0.018231110440360177
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,1536,6144,0.01790844400723775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,64,6144,0.005982222242487802
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,128,64,0.005311111195219888
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,8192,12288,0.042136000262366406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,256,16384,0.007377777662542131
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,2560,7168,0.009755555954244401
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3072,6144,0.011770666473441653
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,4096,512,0.0053431110249625305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,128,7168,0.005295111073387994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,12288,12288,0.22812888357374403
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,10240,7168,0.46199289957682294
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,10240,16384,0.03918755716747708
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,512,768,0.005952000204059813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,65536,128,0.10733599795235528
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,128,128,0.0022311111291249595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,1536,7168,0.009926222264766693
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,65536,512,0.02310577697224087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,12288,3072,0.03868444429503547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,64,256,0.0023351110931899813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,128,4096,0.014826666977670459
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,8192,10240,0.03191288974550035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,6144,65536,0.07900533411237928
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,8192,12288,0.04097955425580343
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,65536,1536,0.03600088755289713
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,4096,16384,0.22252000702752006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,64,2560,0.006892444358931647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,768,5120,0.05609244770473904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,12288,12288,0.0491466654671563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,2048,256,0.014834667245546976
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,16384,768,0.019720888800091214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,2560,1024,0.01643733349111345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3584,256,0.0028826666788922418
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,4096,16384,0.06042666567696465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,256,8192,0.0052871112194326185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,128,8192,0.012073777616024017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,6144,16384,0.022784888744354248
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,512,3584,0.007715555528799693
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,5120,4096,0.011363555159833698
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,64,10240,0.00778311077091429
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,6144,10240,0.01907999979125129
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,16384,3584,0.015980444020695157
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,7168,1024,0.0062773335311147906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,256,12288,0.007327111230956183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,3584,768,0.017983999517228868
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,5120,10240,0.0481697784529792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,2560,3072,0.020654223031467862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,6144,2560,0.124034669664171
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3072,128,0.005675555517276128
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,128,8192,0.032872888776991106
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,7168,1024,0.062466667758093945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3072,6144,0.01107911103301578
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2560,7168,0.045759999089770846
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,256,256,0.013228444589508904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,8192,65536,0.3129093382093641
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,16384,256,0.027505778604083594
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,512,2560,0.00867733359336853
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,4096,2560,0.0273333340883255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,5120,256,0.002558222247494592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,12288,128,0.0063884444534778595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,5120,512,0.015484443969196744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2048,1536,0.003945777813593547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1536,5120,0.009304888546466827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3072,1024,0.004958222309748332
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,2560,6144,0.020972443951500788
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,6144,2560,0.009045333498054082
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,3584,3584,0.01761777698993683
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,7168,32,0.002891555635465516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,6144,8192,0.01904177831278907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,4096,3072,0.00903733323017756
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,512,512,0.012481778032249875
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,4096,1536,0.004095111042261124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,5120,768,0.027179555760489568
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,10240,6144,0.043825778696272105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,7168,768,0.005943111247486538
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,16384,512,0.006269333263238271
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,65536,2560,0.02808977829085456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3072,65536,0.06289155615700616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,7168,3584,0.0888248880704244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1024,12288,0.011352000137170156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,7168,1024,0.004541333350870344
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,768,256,0.005612444546487596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,6144,7168,0.03147644466824002
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,5120,3584,0.009540444446934594
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,10240,5120,0.038006222910351224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,256,1024,0.006520000182920032
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2048,64,0.0021422221842739317
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,7168,768,0.02204177776972453
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,512,1024,0.003489777859714296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,32,1536,0.0035626664757728577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,768,2560,0.004950222041871813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,2048,1536,0.018813333577579923
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,1024,2560,0.0069751110341813825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,12288,4096,0.15594489044613308
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,10240,1536,0.00904977818330129
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,65536,2560,0.05563111437691582
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,16384,64,0.009777777724795872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,8192,1024,0.006533333410819371
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1024,10240,0.05444000164667765
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,128,2560,0.019555555449591745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,4096,2048,0.008829333715968663
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,10240,6144,0.04348533352216085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1024,1536,0.008840888738632202
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,4096,2048,0.007646222081449296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,1536,2560,0.00906666699382994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,6144,1024,0.01872177753183577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,4096,768,0.006014222072230444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,6144,7168,0.02921600143114726
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,3584,2048,0.006853333363930385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,2048,7168,0.010043555663691627
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2048,3584,0.007672000262472365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,1536,3072,0.019023999571800232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,7168,2048,0.007753777835104201
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,10240,512,0.009732444253232744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,8192,4096,0.015557333827018738
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,256,64,0.0019173332386546666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3072,5120,0.011770666473441653
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,256,3072,0.014476444986131458
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3584,8192,0.01982311076588101
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,2048,256,0.005311111195219888
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1536,2048,0.006236444330877728
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,8192,12288,0.15656267272101507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,1024,16384,0.019560888409614563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1024,7168,0.01034311122364468
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,7168,1024,0.006620444357395172
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2048,16384,0.015279110934999255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2048,128,0.0022479999396536085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,16384,16384,0.10462311241361831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,768,8192,0.011411555939250521
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,8192,128,0.01518311103185018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,3072,6144,0.010656889114114972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,128,3072,0.010391111175219217
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,7168,5120,0.013779555757840475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,256,768,0.006017777654859755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,16384,3584,0.040589332580566406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,65536,128,0.02956177790959676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3584,1024,0.00564533347884814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,65536,256,0.022988445229000513
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,64,5120,0.009325332939624786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2048,3584,0.0365626679526435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2560,512,0.0059279998143513995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,64,8192,0.017571555243598092
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1024,64,0.0021706666383478376
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,1024,4096,0.019848888119061787
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,256,128,0.0021422221842739317
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,512,65536,0.01847288840346866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3072,64,0.001826666709449556
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,10240,2560,0.036417778995301985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,16384,12288,0.039671109782324895
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,7168,6144,0.027775110469924078
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,2560,768,0.016900445024172466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,256,12288,0.005557333429654439
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,3584,128,0.014781332678265043
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,32,32,0.0018382221460342407
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,512,10240,0.011680888632933298
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,512,2048,0.014375110467274984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3584,10240,0.016301333904266357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3584,4096,0.009325332939624786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,128,1536,0.003531555748648114
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,256,256,0.012464889221721225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,512,5120,0.0162435554795795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,512,1536,0.013487111363146039
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3072,256,0.0058959999846087555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1536,12288,0.013996443814701505
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,4096,7168,0.011891555454995898
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,1536,6144,0.009687999884287516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,16384,256,0.00944800012641483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3072,1536,0.006280888699822956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3072,65536,0.15736622280544704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,768,10240,0.012479999827014076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,64,65536,0.01296622223324246
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,7168,128,0.005537777725193236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,64,3072,0.0070186663005087115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,2560,2560,0.006979555719428592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3072,6144,0.011425777441925473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,768,8192,0.02033244404527876
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,64,512,0.0076248885856734375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,2048,2048,0.006642666541867786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,5120,12288,0.030022223790486652
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1536,512,0.009723555710580613
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,16384,128,0.00906844437122345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,65536,2048,0.07600800196329753
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,5120,16384,0.05379377802213033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,256,12288,0.013959111438857185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,16384,12288,0.06405688656700982
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,12288,6144,0.03511555658446418
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,5120,8192,0.01443911095460256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,4096,64,0.0020853332761261198
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,5120,64,0.0022320000247822865
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,7168,12288,0.023574221465322707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,2048,4096,0.008000888758235509
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,4096,16384,0.021347555849287245
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,8192,256,0.0032435556252797446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,64,5120,0.005030222237110138
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2048,512,0.005643555687533484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,4096,3584,0.12706400288475886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,512,2560,0.016928889685206942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,6144,768,0.00997511131895913
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,128,5120,0.01422133379512363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2048,64,0.006958222223652734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,8192,2560,0.01699199941423204
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,10240,4096,0.02251911163330078
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,256,64,0.002514666567246119
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,512,7168,0.009045333498054082
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,10240,4096,0.014173333843549093
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,8192,10240,0.03173333406448364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,6144,1024,0.006131555471155379
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3584,65536,2.1113626692030167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,512,8192,0.010409778190983666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,768,128,0.005936000082227919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,10240,1536,0.009335999687512716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,12288,768,0.03331733412212796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,768,256,0.0158897770775689
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,768,7168,0.009697777529557547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1536,65536,0.26618578698900014
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1536,10240,0.015791111522250705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,32,5120,0.008985777695973715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1024,64,0.006083555519580841
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,768,3584,0.00590133335855272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,12288,128,0.006426666759782367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1024,32,0.0026559999419583213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,16384,2560,0.015239111251301236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3584,2560,0.006173333360089197
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,10240,4096,0.016693333784739178
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3072,5120,0.014873777826627096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,256,4096,0.005929777605666055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,7168,2048,0.007194666398896112
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,512,2560,0.007320889168315464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,6144,65536,0.12678310606214735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,128,6144,0.00903555585278405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,128,16384,0.02605244517326355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2560,256,0.019283556275897555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,12288,3584,0.014906666345066495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2048,2560,0.006535111202134027
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,12288,4096,0.29202132754855686
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,16384,256,0.017631111873520743
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,1536,16384,0.03802488909827338
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,16384,3072,0.03596266773011949
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3072,65536,0.5652026600307888
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,128,512,0.007687110867765214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,64,8192,0.01866755551762051
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,512,3584,0.009140444298585257
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3072,1536,0.04305866691801283
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,10240,3584,0.037949333588282265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,128,1536,0.012455110748608908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,3584,7168,0.03959555427233378
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,256,6144,0.01759733259677887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,65536,3072,0.09761955340703328
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,256,2560,0.0052133335007561584
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3072,2048,0.005569777968857024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,1024,65536,0.190247999297248
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,1536,512,0.005740444279379315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,768,3072,0.015962666935390897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,256,3072,0.0047395556337303585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,64,12288,0.01350933313369751
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,1024,4096,0.008019555773999956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,32,512,0.005951111101441913
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2560,1536,0.004025777594910728
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,3072,768,0.016996444927321542
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,512,8192,0.005613333235184352
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,12288,64,0.006322666588756774
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,4096,32,0.0038808890514903595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,1024,65536,0.13200978438059488
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,512,4096,0.00699733363257514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,5120,65536,0.14059200551774767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,128,5120,0.02697333362367418
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,256,65536,0.4971742100185818
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,512,65536,0.0544488893614875
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,3072,1536,0.01276622215906779
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,6144,12288,0.12213067213694255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,16384,32,0.0032186667538351486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1536,65536,0.05824977821773953
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,128,768,0.005619555711746216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3584,768,0.00628355559375551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,6144,16384,0.03819377885924445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,10240,3584,0.020576889316240948
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,65536,1536,0.05283555719587538
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,12288,768,0.005969777703285217
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,5120,2560,0.035024000538720026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,128,512,0.011767999993430244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,1536,7168,0.08791110912958781
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,2560,5120,0.009027555584907532
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,65536,256,0.012701333396964602
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2048,1536,0.003951111187537511
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,1024,1024,0.014859555496109856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,768,16384,0.026536888546413843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,1024,8192,0.033399111694759794
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,4096,16384,0.023736000061035156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,32,2560,0.004523555437723796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,256,4096,0.013078221844302284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,5120,65536,0.19089421961042616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,6144,2048,0.01017599966790941
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,65536,512,0.013854222165213691
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,65536,7168,0.16709421740637886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3072,32,0.005568888866239124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,5120,1536,0.016492444607946608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,16384,7168,0.02648977769745721
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,12288,65536,0.5844133165147569
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,3584,7168,0.0140337778462304
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,5120,2048,0.020883555213610332
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,128,3072,0.007728888756699032
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,64,12288,0.005219555563396878
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,4096,16384,0.03636711173587375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,2560,12288,0.02979733215437995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,65536,12288,0.23745955361260307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,768,128,0.0063866666621632045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,512,8192,0.008211555580298105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,16384,512,0.03511555658446418
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,5120,1024,0.006321777900060018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,8192,3072,0.011063999599880643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,768,12288,0.018651554981867473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,128,8192,0.018623110320832994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,128,12288,0.06642133659786649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,4096,3072,0.00796977761718962
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,768,4096,0.016217778126398723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,6144,1024,0.01921955578856998
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,128,12288,0.03648622168434991
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,256,5120,0.009365333451165093
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,512,1536,0.006254222244024277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3072,65536,0.13383200433519152
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,512,768,0.004709333181381226
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,4096,1536,0.019648000597953796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2560,768,0.006160000132189856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,128,32,0.0020453333854675293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,16384,128,0.0032222221295038858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,6144,5120,0.013556444810496436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3584,2048,0.01346666696998808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,2048,6144,0.03187555405828688
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,256,512,0.0028560000161329904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,1536,6144,0.008685333033402761
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,512,256,0.015505777464972602
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1024,8192,0.025656888882319134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2560,64,0.005939555664857228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1024,32,0.005584888988071018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,16384,1024,0.009399111072222391
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,10240,64,0.003968888686762916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,512,16384,0.0209840006298489
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,1536,256,0.015277332729763456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1536,3072,0.00739644467830658
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,3584,8192,0.01349422252840466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2560,5120,0.008120888637171851
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,512,5120,0.008391111261314815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,7168,16384,0.02792444494035509
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,1024,512,0.013179555535316467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1024,1024,0.0041351111398802865
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,12288,65536,0.15724533134036595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,512,768,0.028748446040683325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1536,1536,0.006423111177153057
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,512,5120,0.009720889230569204
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,3072,65536,0.06486933098898993
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,3584,2048,0.018999111321237352
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,768,64,0.005281777845488654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,64,10240,0.018577777677112155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,12288,768,0.019460444649060566
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,32,3072,0.0042133331298828125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,8192,1536,0.0199946661790212
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,8192,12288,0.5315617985195583
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1536,4096,0.00853777759604984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,6144,64,0.007979555262459649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,2560,3072,0.01588000026014116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,64,7168,0.010039110978444418
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,10240,128,0.015463999576038785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,256,2560,0.006682666639486949
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,1024,2048,0.023642667465739783
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,2048,12288,0.030718220604790583
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,8192,7168,0.3461368878682454
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,256,12288,0.022188444932301838
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,256,256,0.005254222287072076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,32,2048,0.0064791109826829694
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,65536,6144,0.06525866852866279
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,128,512,0.01146399974822998
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,64,10240,0.012093333734406365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,16384,768,0.024678222007221643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,8192,768,0.004751111070315043
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,3072,8192,0.03722133239110311
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,512,3072,0.01864177816443973
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,768,5120,0.005619555711746216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3584,2560,0.0070168889231152
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,512,1024,0.005705777969625261
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,32,2048,0.004378666480382283
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,512,256,0.004896000027656555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1536,128,0.0069368887278768755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,5120,1024,0.004866666677925321
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,256,2048,0.004606222112973531
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3072,4096,0.008016889293988546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,2560,5120,0.009110222260157267
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,768,64,0.002502222235004107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,7168,512,0.005631999837027655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,16384,7168,0.0314319994714525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,3072,8192,0.011744889120260874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,128,7168,0.009707555174827576
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,512,2048,0.013951111170980664
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,512,3072,0.00758755538198683
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,256,32,0.006245333287451003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3584,16384,0.03482222225930955
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,128,16384,0.020962667134073045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,64,32,0.005199999858935674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,1024,2048,0.006711999989218182
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,3584,16384,0.021260443660948012
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,12288,3584,0.028772443532943726
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,2048,65536,0.07677955759896173
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,10240,3584,0.03428355521625943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,128,7168,0.01018311083316803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,6144,12288,0.021381333470344543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,7168,512,0.04727377825313144
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,4096,2560,0.018398221996095445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,4096,12288,0.11993955241309272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,65536,2048,0.03195022212134467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,128,4096,0.007672888537247975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,512,3072,0.018925334016482036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,16384,10240,0.12860800160302058
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,4096,7168,0.031112889448801678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,6144,2048,0.022620444496472675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,768,1024,0.0064995553758409284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1536,10240,0.015499555402331881
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,8192,6144,0.02254666719171736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,2048,3584,0.0192586663696501
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3584,10240,0.05378399954901802
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,5120,7168,0.014376888672510782
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,5120,7168,0.028202666176689997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,16384,7168,0.08511555857128567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,3584,3584,0.020415110720528495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,512,3584,0.0162462227874332
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1536,1024,0.007852444218264686
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,3072,768,0.016826666063732572
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3072,1024,0.006320888797442119
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,6144,2048,0.024452444579866197
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,4096,512,0.0029333333174387612
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,128,1024,0.005599110904667113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,65536,4096,0.37433332867092556
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2560,32,0.008344888687133789
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,4096,3072,0.05184177888764275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,1024,2048,0.014615999327765571
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,16384,8192,0.03978133201599121
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,1024,2560,0.019537778364287484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3072,2048,0.005944889038801193
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,128,2560,0.006686222222116258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,5120,8192,0.039128889640172325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3072,12288,0.01570399933391147
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,8192,768,0.006372444331645966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,512,2048,0.004573333180612988
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,12288,64,0.002509333400262727
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,16384,2048,0.011966221862369113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,2560,12288,0.0560435586505466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,6144,512,0.00517866677708096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,10240,3072,0.013941332697868347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,5120,65536,0.09973955816692776
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,8192,32,0.002524444419476721
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3072,256,0.006588444527652528
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,4096,12288,0.029133333100212946
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,7168,8192,0.04914222160975138
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,10240,5120,0.02964444292916192
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1024,65536,0.056164443492889404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1024,8192,0.018976888722843595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,12288,3584,0.013600889179441663
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2048,3584,0.007894222107198503
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,128,768,0.005619555711746216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,65536,16384,0.1446337832344903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1536,1024,0.014305777019924588
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,768,128,0.013806222213639153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,32,4096,0.007753777835104201
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,7168,4096,0.03296711047490438
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,256,4096,0.015832889411184523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2048,12288,0.014327110515700446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1024,4096,0.005947555518812603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,8192,1536,0.012702222499582501
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,768,768,0.008240888516108194
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,16384,6144,0.5596284336513943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,16384,1536,0.012168000141779581
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,128,10240,0.020945777495702107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,12288,2048,0.010205333431561789
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,5120,768,0.005633777628342311
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,12288,768,0.018204443984561496
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,16384,5120,0.12521689467959934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1024,8192,0.010640000303586325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,3072,3584,0.025441777375009324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,7168,1024,0.01754488878779941
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,1536,256,0.005610666755172942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,8192,5120,0.012072000238630505
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,6144,1536,0.007272889216740926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,3072,5120,0.008292444050312042
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,8192,10240,0.036702222294277616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,5120,8192,0.135096894370185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,5120,4096,0.008437333007653555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,6144,128,0.0069253332912921906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,6144,768,0.005610666755172942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,128,10240,0.011975111232863532
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,16384,10240,0.3769724369049072
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,768,2560,0.005233777893914117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,512,6144,0.014382221632533602
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,3584,3584,0.02409422232045068
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,1536,65536,0.09819910923639934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,16384,4096,0.01538044379817115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,4096,2560,0.02032711108525594
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,4096,2560,0.008612444831265343
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1536,128,0.014163555370436775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,6144,3584,0.012816000315878125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1536,3072,0.005648889061477449
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3584,64,0.006619555668698416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,128,1536,0.012759110993809171
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,1536,128,0.014923555983437432
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,256,16384,0.026326222552193537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,6144,8192,0.04885422190030416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,128,65536,0.04669333497683207
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,512,3584,0.009673777553770278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,768,6144,0.009082666701740688
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,7168,8192,0.042040887806150645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,8192,1024,0.009276444713274637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,16384,65536,0.37270487679375547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,768,6144,0.0306915541489919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2560,768,0.003917333152559069
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,5120,3072,0.008446222378147973
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,10240,7168,0.025240000751283433
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,512,256,0.002552888873550627
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,10240,7168,0.9566862318250867
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,64,7168,0.009977777798970541
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,8192,4096,0.013390222357379066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,4096,1536,0.007309333317809635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,5120,65536,0.10172177685631646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,768,7168,0.009746666583749983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,7168,512,0.005692444327804778
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,3072,65536,0.23615201314290366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,10240,1536,0.014333332578341166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,128,8192,0.008164444731341468
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,768,12288,0.014860444598727755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,5120,768,0.005696889013051987
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,12288,64,0.002550222186578645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,5120,65536,0.0728257762061225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,6144,4096,0.009840889109505547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,2048,512,0.005562666803598404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,8192,4096,0.026522666215896606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,4096,1536,0.05496177739567227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,12288,256,0.006351999938488007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,8192,5120,0.014498665928840637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,1024,1024,0.013493333425786761
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,3072,768,0.018290667070282828
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,64,512,0.0028817777832349143
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,4096,10240,0.015619556109110514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,65536,64,0.01314222233162986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,1536,512,0.013806222213639153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2560,3072,0.007382222347789341
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1536,2048,0.007039111521508958
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,12288,4096,0.08279911014768812
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,512,768,0.0029146667155954572
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,6144,6144,0.09701866573757595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,6144,2048,0.00793511089351442
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1536,3072,0.008082666330867344
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,16384,1024,0.008420444197124904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,65536,128,0.022337777747048274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,8192,16384,0.046874665551715426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,16384,3072,0.03172889020707872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,8192,8192,0.04643733302752177
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,5120,6144,0.01974933346112569
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,12288,6144,0.02019377715057797
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1536,12288,0.016155555844306946
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,65536,16384,0.3030631012386746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,16384,2048,0.011002666420406766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,5120,1024,0.008734222087595198
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,10240,512,0.03089155422316657
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,12288,1536,0.00870666652917862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,2560,768,0.00795199970404307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,1536,7168,0.022078222698635522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,32,3584,0.007981333467695449
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,64,768,0.006222222414281633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,512,1536,0.00739555557568868
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2560,32,0.002206222257680363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3072,12288,0.07201777564154731
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,768,256,0.002542222125662698
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,256,2560,0.005144888742102517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,65536,5120,0.10393599669138591
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,4096,4096,0.007705777883529663
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,512,2560,0.014450665977266101
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,10240,2048,0.010767999622556897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,256,768,0.005660444498062134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2048,12288,0.01425066590309143
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,1536,512,0.00564177789621883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3584,3584,0.040376888381110296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,65536,768,0.017489777670966256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,768,3584,0.005889777921968036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,10240,10240,0.03466399841838413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,10240,5120,0.01570666664176517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,256,2048,0.004957333207130432
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,256,32,0.0053582220441765254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,6144,6144,0.04080355498525832
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3584,6144,0.01293777757220798
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,2560,2048,0.01973600023322635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,1536,4096,0.01644533375898997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,6144,768,0.016833777228991192
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,7168,256,0.013567999833159976
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,2560,10240,0.04085777865515815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,6144,3584,0.011583111352390714
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,7168,768,0.0029324444217814338
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,512,1024,0.007668444679843054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1024,4096,0.02942755487230089
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,12288,1024,0.062021334966023765
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,65536,1024,0.017483555608325534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,128,4096,0.007958222594526079
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,4096,8192,0.2696942223442925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,10240,64,0.002507555608948072
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2048,12288,0.03186577889654372
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1024,7168,0.007908444437715743
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1536,512,0.006698666761318843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,6144,16384,0.16000799338022867
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,512,65536,0.02791911032464769
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,768,128,0.008101333346631791
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,7168,2048,0.008652444514963362
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,3584,5120,0.031239999665154353
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,12288,768,0.013974222871992322
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,7168,256,0.007000889215204451
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,7168,65536,0.0843377775616116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,768,8192,0.15053866969214544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,128,1536,0.006295111030340195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,2048,256,0.021761778328153823
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,3584,5120,0.02143200072977278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3072,10240,0.017246221502621967
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,128,256,0.0022337778160969415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,5120,8192,0.020766221814685397
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,2560,5120,0.05823555257585314
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,512,1536,0.013448889056841532
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,16384,3584,0.0792906681696574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,16384,1024,0.008038221961922115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,6144,4096,0.024119999673631456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,6144,7168,0.013600000076823764
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,512,768,0.0032355555643637977
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,6144,6144,0.017280888226297166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,16384,5120,0.020463110672103036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,6144,10240,0.023631110787391663
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,7168,512,0.011415999796655444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,65536,768,0.01919644408755832
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,256,2048,0.004577777865860197
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3584,5120,0.015909334023793537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,1024,16384,0.028891556792789038
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,10240,4096,0.032292445500691734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,32,5120,0.0047146665553251905
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,6144,4096,0.04822044571240743
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3072,2560,0.008254222571849823
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,4096,16384,0.022601778308550518
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,768,2048,0.006682666639486949
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,512,10240,0.011808888779746162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,768,2560,0.006882666713661618
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,16384,1536,0.02836622132195367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,16384,16384,0.0867919988102383
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,65536,2560,0.19330667124854195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,8192,65536,0.09752888811959161
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,3584,7168,0.028632889191309612
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,12288,8192,0.059431109163496226
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1536,64,0.005981333139869902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,65536,2048,0.03906666570239597
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,512,768,0.0028648889727062653
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2560,128,0.005696889013051987
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,256,1024,0.012144000165992312
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,3072,1024,0.016474665866957772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,128,7168,0.017990221579869587
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,10240,1024,0.0076453329788313965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,2560,768,0.01789422167672051
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,3584,768,0.006311999840868845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,2560,3584,0.0181697772608863
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3072,1024,0.003567111161020067
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,5120,7168,0.030040889978408813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,2560,1024,0.01575999955336253
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,128,256,0.0052373334765434265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,2048,128,0.014592889282438489
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,4096,12288,0.01905333333545261
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,256,8192,0.0059279998143513995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,5120,1536,0.0783440007103814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1024,65536,0.09838577773835923
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,2560,3584,0.007979555262459649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,8192,6144,0.01575999955336253
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,64,2048,0.004587555511130227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,768,12288,0.013087111214796701
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,512,65536,0.058782219886779785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,8192,3072,0.010586666564146677
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,128,8192,0.01854933301607768
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,3584,768,0.01677600045998891
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,7168,1536,0.006623111251327727
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,12288,65536,0.5149075720045302
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,5120,128,0.0021600000974204806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,16384,8192,0.04544533292452494
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,2560,512,0.013798221945762634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,256,5120,0.006911110960774952
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,6144,768,0.018975110517607797
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,8192,1024,0.006269333263238271
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,12288,128,0.015453333655993143
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,2560,3584,0.05979466438293457
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,12288,64,0.008685333033402761
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,7168,64,0.003911111089918348
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1024,32,0.005952000204059813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,5120,1024,0.01755644381046295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,5120,7168,0.01219555570019616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,2048,16384,0.03227555420663621
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,16384,1024,0.027456889549891155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2048,65536,0.09763377904891968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,128,65536,0.020421332783169217
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,12288,3584,0.014739554789331226
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,16384,6144,0.023810666468408372
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,512,256,0.005572444448868434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,16384,768,0.00601244428091579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,768,256,0.0021742222209771476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,768,6144,0.006679111056857639
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,7168,2560,0.008135110967689091
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,12288,256,0.0076248885856734375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1024,64,0.005910222315125995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,16384,512,0.011567999919255575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,64,32,0.007975999679830339
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,7168,256,0.005264889034960005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,512,3584,0.0075635554061995606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,32,65536,0.009640889035330879
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,3072,128,0.02147644427087572
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,4096,768,0.005387555393907759
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,65536,6144,0.06561422348022461
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,10240,2048,0.02367288867632548
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,128,2048,0.0069368887278768755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,12288,1536,0.01367911116944419
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,4096,256,0.003286222202910317
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,12288,1024,0.012443555725945367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1536,7168,0.010344888601038191
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,768,65536,0.05701511436038547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,64,7168,0.005619555711746216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,7168,32,0.0028524444335036804
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,1536,4096,0.01629599928855896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,256,1536,0.013241777817408243
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,5120,4096,0.010288889209429422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,512,2560,0.0069324444565508105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,10240,65536,0.14350844754113093
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,65536,512,0.05747999747594198
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,8192,5120,0.021964444054497614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,65536,16384,0.16636710696750218
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,1024,2560,0.01520088811715444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2048,12288,0.02114222281508976
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,32,768,0.006266666783226862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1536,1536,0.0038906666967603895
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1024,768,0.0032266666077905228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,65536,16384,1.9463369581434462
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,3584,1536,0.006721777634488211
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,128,256,0.013869333598348828
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,2048,4096,0.008449777960777283
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,7168,12288,0.03372088736957974
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,4096,4096,0.1420079999499851
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,4096,1536,0.018212444252438016
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,2048,2560,0.0162471118900511
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,65536,256,0.011380444798204633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,3072,256,0.005623999983072281
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2048,1024,0.003522666792074839
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,5120,10240,0.017817777064111497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3584,3584,0.007616889145639207
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,8192,12288,0.02427911096149021
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,10240,128,0.015477332803938123
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,64,1536,0.006360888895061281
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,4096,12288,0.48808887269761825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,4096,512,0.005656888915432825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,128,6144,0.014146667387750415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,1024,6144,0.009171555439631144
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,16384,2560,0.035213334692849055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,6144,256,0.021224000387721594
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,16384,256,0.017613333132531907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,16384,3072,0.015212444795502557
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,3584,2560,0.012587555580668978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2560,5120,0.009390222529570261
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,12288,768,0.01791733337773217
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3072,64,0.0021955555097924336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,16384,3584,0.041922665304607816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,7168,3072,0.02306844459639655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,3072,256,0.005314666777849197
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,65536,16384,0.38658489121331113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,10240,128,0.0073724447025193115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,4096,768,0.006789333290523953
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,256,1024,0.008035555481910706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,6144,6144,0.015804444750150044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,768,4096,0.00904088881280687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,32,768,0.01274400038851632
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,768,128,0.007712889048788283
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,32,2048,0.0063457778758472866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,512,12288,0.009924444059530893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,32,65536,0.0070017774899800616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,64,3584,0.007768889268239339
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,16384,4096,0.020184000333150227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3072,12288,0.020041777027977836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,12288,128,0.006616888774765863
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,2048,2560,0.03155022197299533
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,128,3072,0.025638222694396973
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2560,2560,0.05587110916773478
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,16384,128,0.015581332974963717
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2560,6144,0.015009777413474189
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,3072,16384,0.021029333273569744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,3584,10240,0.30464087592230904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,256,12288,0.013358222113715278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,768,512,0.012776888906955719
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,1024,768,0.005988444305128521
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,8192,3072,0.010460444622569613
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3584,16384,0.021521778570281133
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,5120,128,0.007685333490371704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,65536,5120,0.13203022215101454
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,12288,256,0.039467556609047785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,7168,1024,0.005962666538026597
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,512,10240,0.012147555748621622
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,8192,16384,0.1611377795537313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1024,32,0.002198222196764416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,3072,6144,0.009745777481132084
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,2048,3584,0.008014221986134848
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1024,4096,0.008349333372380998
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,65536,32,0.005320888840489917
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,65536,768,0.01587377819750044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,6144,256,0.005991111199061076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3072,4096,0.008719111482302347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,10240,512,0.019280888968043856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,512,3072,0.0073315559162033936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,1536,16384,0.031818668047587075
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,8192,256,0.02478844424088796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,32,5120,0.004939555707905027
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,64,512,0.006033777776691649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3072,64,0.00536088893810908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,1536,8192,0.020670221911536325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,128,1024,0.011731555892361535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,4096,3584,0.011906666888131036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,16384,5120,0.03918044434653388
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,5120,7168,0.04123644365204705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,3584,7168,0.023686221904224817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,768,2048,0.01588355501492818
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,64,512,0.005647111270162795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,4096,16384,0.03710222244262695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,2048,3584,0.012534221841229333
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,1024,768,0.0158933334880405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,6144,64,0.0019902222686343724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,16384,6144,0.0399982233842214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,128,512,0.002537777854336633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,4096,256,0.005970666805903117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,8192,4096,0.015840889679061044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,256,10240,0.012792888614866467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,4096,512,0.010115555591053432
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,512,6144,0.00964355551534229
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2048,32,0.01635199950801002
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,5120,7168,0.12407733334435357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,8192,6144,0.01593866613176134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,768,32,0.0022257777551809945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,3584,7168,0.0314319994714525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,12288,10240,0.043271111117468945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,5120,8192,0.0310515562693278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,3072,3072,0.024728889266649883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,65536,256,0.024735111329290602
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,10240,32,0.006639110959238476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,2560,8192,0.011229332950380115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,16384,65536,0.1759493350982666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,512,2048,0.013914666242069669
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,32,8192,0.004233777936961916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,16384,1536,0.028014222780863445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1536,64,0.0052897776994440295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3072,512,0.005946666830115848
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,10240,4096,0.01384622189733717
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,768,3584,0.0074035558435651995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,128,12288,0.01346666696998808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,1024,512,0.012801777985360889
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,2560,10240,0.025435555312368605
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,12288,2048,0.0410791105694241
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1536,64,0.006317333214812809
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,128,7168,0.017650667164060805
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,5120,1536,0.006335999816656113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,512,768,0.006295111030340195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,512,2560,0.016535111599498324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,2560,65536,1.8129751417371962
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,256,512,0.013870221873124441
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,8192,16384,0.030095110336939495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,5120,32,0.01651733285850949
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,16384,1536,0.05267733335494995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,1536,3072,0.07553688685099284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,16384,768,0.019336000084877014
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2048,8192,0.17458400461408827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,512,256,0.00535733335547977
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,768,10240,0.03292977809906006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,32,8192,0.01075644459989336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,7168,5120,0.013201778133710226
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,5120,10240,0.020580444071027968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,1024,2560,0.00867733359336853
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,128,128,0.0021333334346612296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,64,7168,0.004976888911591636
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,65536,4096,0.329187552134196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1024,7168,0.028723554478751287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3584,3072,0.012444444828563266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,128,3072,0.0073520003093613525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,3072,16384,0.06023199690712822
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,64,4096,0.008023999631404877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,12288,8192,0.054377777708901294
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1536,12288,0.012198222180207571
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,7168,128,0.01514755520555708
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,10240,32,0.0032186667538351486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,1024,128,0.005562666803598404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,10240,768,0.004325333154863781
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,10240,8192,0.03757777810096741
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,128,4096,0.015190222197108798
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,2048,7168,0.009737778041097853
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1024,12288,0.025795555777019922
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,8192,3584,0.011568889021873474
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,1024,65536,0.09438044495052761
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,768,4096,0.012794666820102267
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3072,6144,0.04068266683154636
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,7168,2048,0.007308444215191736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,12288,16384,0.08748444583680894
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,12288,16384,0.11056177483664618
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,128,768,0.013478221992651621
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,16384,2560,0.022132444712850783
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3584,65536,0.0769875579410129
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,6144,2560,0.07010755274030897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,3584,10240,0.05354666709899902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,16384,1024,0.04387022058169047
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,8192,16384,0.07957777712080213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,8192,1024,0.006654222392373615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,4096,1536,0.004330666528807746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,16384,768,0.007892444729804993
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,128,768,0.0032559999575217566
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,4096,256,0.012063999970753988
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1024,256,0.004602666530344221
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,256,10240,0.02141422198878394
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,1536,4096,0.02663999961482154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,3584,768,0.005619555711746216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1024,12288,0.017295999659432303
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,64,64,0.0018453333112928604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,12288,4096,0.03622488843070136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,10240,12288,0.17699821790059408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,256,3584,0.014569777581426831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,5120,128,0.005241777747869492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,512,10240,0.012490666574902005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,16384,128,0.006627555522653792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,7168,7168,0.04351822204060025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,3072,2560,0.01755644381046295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,128,8192,0.00565244464410676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1024,256,0.005976888868543837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,6144,2048,0.007009777757856581
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,256,5120,0.016026667422718473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,1536,3584,0.01109688894616233
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,64,10240,0.006628444625271692
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,6144,256,0.0029200000895394217
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,6144,128,0.005930666708283954
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2048,2560,0.008762666748629676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,2048,2048,0.020033778415785897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,1024,1024,0.005936000082227919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,7168,64,0.005940444353553984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,5120,8192,0.02494311167134179
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,512,8192,0.0122097780307134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,256,1536,0.01347733371787601
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,768,6144,0.009408888717492422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,12288,7168,0.029511110650168523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,10240,8192,0.02940088841650221
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,1536,1024,0.011439999772442712
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,512,1536,0.018605333235528734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,5120,3072,0.007821333077218797
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,128,12288,0.025055999557177227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,2048,3072,0.023911111884646948
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,5120,4096,0.027480888697836135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,64,6144,0.009336888790130615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,256,65536,0.01616533266173469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,768,6144,0.02132266594303979
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2048,128,0.005920889063013925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3072,256,0.005959110955397288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3072,1536,0.016873778568373788
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,7168,8192,0.09930488798353408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,4096,768,0.007297777467303806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,5120,512,0.007115555306275685
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,6144,32,0.006729777902364731
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,5120,2048,0.006896889044178857
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3072,1536,0.006581333362393909
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,5120,512,0.00600355573826366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,1536,768,0.016201777590645682
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,768,3072,0.007707555260923173
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1024,12288,0.029463112354278564
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2560,6144,0.01038755559258991
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,768,128,0.005655111124118169
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,65536,12288,0.17398222287495932
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,16384,256,0.017567111386193168
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,768,256,0.006989333364698622
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,16384,3584,0.033005333609051175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,3072,10240,0.03341155582004123
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,6144,65536,0.12149155139923096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,1024,3584,0.008103999826643202
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,6144,4096,0.03184533450338576
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,5120,1024,0.020272889071040683
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,3584,128,0.005304000029961268
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,256,6144,0.004969777746333016
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,65536,3584,0.5398880110846626
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,3072,4096,0.05800355805291069
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,768,3584,0.01923822197649214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,65536,16384,0.22997511757744682
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,1536,8192,0.010804444551467896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,4096,128,0.007850666840871176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,512,1536,0.013507555756303998
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,7168,768,0.0565333366394043
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,2560,3072,0.007348444726732042
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,2560,256,0.015909334023793537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,4096,65536,0.4854782422383626
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1536,2048,0.005997333261701796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,6144,256,0.021673777037196692
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,64,8192,0.008662222160233391
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2560,12288,0.019302222463819716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,7168,12288,0.040844443771574236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,7168,512,0.007731555236710443
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,128,32,0.0018622221218215094
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,12288,3072,0.017110221915774874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,8192,65536,0.13365511099497476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1536,1536,0.004912888838185204
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,32,7168,0.02478222217824724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,256,8192,0.040130668216281466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2048,65536,1.413777775234646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,5120,5120,0.018412444326612685
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1024,1024,0.010402667025725046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,4096,65536,0.06799288590749104
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,16384,32,0.0086986662613021
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,65536,1024,0.019816888703240287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,65536,1024,0.0188719994492001
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,10240,12288,0.0317555566628774
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,3072,768,0.023802666200531855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,32,12288,0.0196115556690428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,8192,7168,0.37608000967237687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,5120,256,0.0025635556214385563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,6144,3072,0.007522666619883643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,5120,16384,0.022094221578703985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,5120,10240,0.17040267255571154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,256,768,0.013450666434235044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,10240,1536,0.015150222513410779
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,4096,768,0.024833778540293377
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2048,10240,0.016587555408477783
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,5120,5120,0.011745777395036487
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,10240,65536,0.20307555463578966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,12288,2048,0.011385777758227454
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1536,3584,0.05598400036493937
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,8192,8192,0.12293867270151775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,65536,3584,0.137210660510593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,256,3584,0.014866666661368476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,3072,7168,0.011154666543006897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,65536,2048,0.12092888355255127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,128,4096,0.015935111376974318
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,5120,768,0.050442665815353394
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,7168,512,0.025981333520677354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,5120,3072,0.007296889192528195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,1024,16384,0.03462666604253981
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,10240,65536,0.14849155479007298
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,2560,256,0.014133334159851074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1536,5120,0.02760533326201969
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,2048,8192,0.04293333490689596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,7168,64,0.007651555869314406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,2560,5120,0.023402666052182514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,5120,768,0.017235555582576327
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2560,12288,0.017302221722073026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,2560,8192,0.0103493332862854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2048,768,0.006846222198671765
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,8192,4096,0.026400888959566753
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,4096,2048,0.022566222482257422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2560,512,0.022052443689770166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,768,8192,0.006986666884687211
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3072,512,0.00258577780591117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,768,256,0.0052639999323421055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,65536,10240,0.21110844612121582
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2048,32,0.005280888742870755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,6144,12288,0.021730666359265644
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,16384,16384,0.09660444657007854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,512,32,0.008017777568764156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,4096,2560,0.018633777896563213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,10240,12288,0.044601778189341225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,4096,768,0.018211555149820115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,12288,1536,0.011412444214026133
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,512,3584,0.015398222539159985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3072,10240,0.013807111316257052
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,128,3072,0.018625777628686693
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,5120,10240,0.5388844278123643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,2560,6144,0.03838488790724013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,4096,1536,0.0048071112897661
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,16384,16384,0.048039999273088246
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,5120,768,0.013584888643688627
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,5120,1536,0.006244444598754247
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,7168,768,0.006276444428496891
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,2560,128,0.014860444598727755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,256,32,0.005646222167544895
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,768,5120,0.017599110802014668
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,12288,3072,0.013450666434235044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,2560,6144,0.020958221620983545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,4096,256,0.005924444645643234
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,2560,16384,0.03191733360290527
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3072,2048,0.008961777720186446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,10240,128,0.005991111199061076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,6144,8192,0.017557332913080852
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,3584,7168,0.01979111135005951
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,5120,128,0.00526577772365676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,6144,512,0.010085333552625444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,64,256,0.006276444428496891
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,16384,2560,0.013445333474212222
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,3584,2560,0.006907555378145642
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2048,512,0.005963555640644497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,768,6144,0.010122666756312052
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,2048,7168,0.010363555616802638
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,512,12288,0.027756444282001917
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,7168,6144,0.2680799961090088
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,1536,4096,0.008259555531872643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,12288,3072,0.03363822235001458
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,256,2048,0.006656888872385025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,128,128,0.005290666802061929
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,64,2048,0.005610666755172942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,4096,768,0.018580444984965853
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,16384,32,0.007699555820888943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,7168,128,0.005690666536490123
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,65536,512,0.05719644493526883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,2560,8192,0.01575199928548601
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,10240,5120,0.030786666605207656
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3072,16384,0.02106222179200914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3584,2560,0.009679111341635386
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,10240,128,0.01551555593808492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,512,1024,0.005990222096443176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,4096,3072,0.029271112547980413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,512,2560,0.014180445008807711
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,128,7168,0.005261333452330695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,2048,256,0.015832889411184523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,256,7168,0.009822222093741098
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,64,256,0.002491555487116178
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3072,3072,0.007687999970383114
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2560,3072,0.04200266798337301
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,7168,128,0.0022613333745135204
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,512,32,0.0022613333745135204
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,10240,5120,0.020768000019921195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,128,768,0.007409777906205919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,32,64,0.002489777695801523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,7168,256,0.017252445220947266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,65536,1536,0.06948888964123197
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2560,256,0.005943111247486538
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,5120,32,0.002943111169669363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,1536,512,0.016390222642156813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,16384,1024,0.08169333140055339
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,6144,128,0.009135111338562435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,6144,4096,0.01498044364982181
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,768,8192,0.01036622209681405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,8192,65536,0.1675013303756714
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,12288,10240,0.0329733325375451
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3072,32,0.006324444380071428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,32,768,0.002936888900068071
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,4096,256,0.005618666609128316
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,128,7168,0.009753777749008602
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,10240,5120,0.06082489093144735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,6144,768,0.023207111491097346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,6144,5120,0.0332622230052948
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,768,1536,0.006559999866618051
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3584,32,0.002518222149875429
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,5120,8192,0.2667226791381836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,128,2560,0.013439111411571503
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,8192,4096,0.015154666370815702
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,768,65536,0.06356000237994723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,65536,32,0.013761777844693927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,256,7168,0.010382222632567087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3584,65536,0.45060178968641496
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,128,1536,0.003825777934657203
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,65536,2048,0.04204533497492472
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,512,256,0.01994311147265964
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2048,6144,0.009649777577983009
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,10240,4096,0.03573955429924859
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,7168,5120,0.02611555490228865
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,5120,5120,0.08708888954586452
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,7168,1536,0.009373333719041612
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,12288,65536,0.17319733566708037
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,1536,2560,0.007962666451931
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,2560,1024,0.042261332273483276
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,8192,65536,4.78053961859809
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,16384,2560,0.013535999589496188
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,32,4096,0.00452622233165635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,16384,3072,0.01482399966981676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,256,8192,0.0052693333062860705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3072,128,0.0022373333987262514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,8192,10240,0.05746399694018894
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,12288,3072,0.01556622154182858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,32,256,0.005584888988071018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,2048,512,0.023297778434223596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,1536,2048,0.027455111344655354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,65536,768,0.016019556257459853
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,12288,1536,0.026718222432666357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,12288,32,0.008685333033402761
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,4096,128,0.011121778024567498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3072,128,0.005576888720194499
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,5120,2048,0.009399111072222391
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,2560,65536,0.6883982022603353
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,16384,16384,2.305831061469184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3584,256,0.009708444277445475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,768,256,0.00979200005531311
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3584,12288,0.02718755602836609
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,7168,1024,0.0070293330483966405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,32,3584,0.005040888984998067
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,10240,12288,0.058858666155073375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,256,6144,0.035537779331207275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,64,16384,0.015182221929232279
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,6144,128,0.002487999904486868
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,3584,768,0.021700445148679946
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,512,4096,0.006630222416586346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,512,8192,0.013809777796268463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,16384,1536,0.0273617770936754
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,512,12288,0.03715288970205519
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2048,1536,0.003921777837806278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3072,64,0.005581333405441708
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3584,6144,0.013410666750537025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,5120,1024,0.016311999824311998
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2560,1536,0.011381333072980246
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,7168,16384,0.023992889457278784
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3584,512,0.006224888894293044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,512,2560,0.00489066665371259
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,5120,768,0.0038684445122877755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,128,1536,0.006274666637182236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,16384,1024,0.026188444760110643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3072,2560,0.0064426664676931165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1536,64,0.006239111224810283
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,16384,32,0.007265778051482306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,768,12288,0.0466177761554718
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,768,8192,0.008355555435021719
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,1024,3584,0.016174222032229107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,6144,65536,0.07819822099473742
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,10240,1024,0.00903200027015474
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,256,4096,0.011078221930397881
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,128,256,0.008050666915045844
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1024,10240,0.009908444351620143
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,768,6144,0.0648933317926195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,256,8192,0.013632000320487551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,65536,2560,0.07966133621003893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,3584,1024,0.015099555253982544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,128,3584,0.007400000260935889
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,1536,1536,0.014431110686726041
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,128,8192,0.010462221999963125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1536,3072,0.005574222240183089
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,128,7168,0.006623111251327727
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,16384,7168,0.021225778592957392
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3072,7168,0.014031999640994601
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,12288,65536,0.24897421730889213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,12288,10240,0.04227377639876472
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,768,128,0.00815644446346495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,3584,512,0.005673777725961473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,4096,5120,0.016904888881577384
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,12288,64,0.002234666711754269
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,10240,64,0.007335999773608313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,512,12288,0.013227555486891003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,16384,6144,0.031103110975689355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,512,2048,0.007315555380450354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,16384,3072,0.013203555511103736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,7168,65536,2.7648870680067272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,2560,3584,0.03578755590650771
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,768,8192,0.021615111165576514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,1536,6144,0.01817599932352702
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,65536,16384,0.15888444582621256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,128,8192,0.021008888880411785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,5120,128,0.005612444546487596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,3072,4096,0.007616000043021307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,12288,256,0.006593777901596493
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3584,2560,0.007280000381999546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,2048,1536,0.006247111078765657
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,6144,8192,0.014905777242448596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,3072,8192,0.03200533323817783
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,1536,16384,0.01887022289964888
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,32,6144,0.017508443858888414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,16384,2048,0.032482667101754084
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3584,1536,0.006565333240562015
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,768,512,0.008014221986134848
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,768,12288,0.008381333616044786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,256,5120,0.009039999710188972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,12288,3072,0.031586666901906334
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,65536,2560,0.03235822253757053
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,6144,8192,0.014486221803559197
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3584,7168,0.1784728897942437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,65536,8192,0.7313386599222819
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,1536,256,0.005613333235184352
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,8192,7168,0.044496887260013156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,2560,65536,0.08943111366695827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,64,65536,0.05650133556789822
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,32,128,0.0021840000732077495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,8192,128,0.005345777918895085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,3584,12288,0.017457778255144756
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,7168,512,0.005582222094138463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,512,2048,0.014241778188281588
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,256,128,0.0069546666410234236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,6144,4096,0.01055111073785358
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,12288,1024,0.007654222349325816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,7168,10240,0.06813600328233507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2048,32,0.002228444442152977
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,65536,128,0.01218311074707243
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,10240,768,0.006973333656787872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,3072,8192,0.22992889086405435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,5120,128,0.007336888876226213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,7168,1024,0.004893333133724001
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,6144,8192,0.022080888350804646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,16384,12288,0.04225600096914503
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,512,3584,0.022272000710169475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,10240,512,0.004017777740955353
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,6144,16384,0.18173689312405059
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3072,5120,0.010025777750545079
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,128,768,0.006966222491529252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,16384,12288,0.05395288930998909
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,65536,7168,0.06473333305782743
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,128,3584,0.014198222094111972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2560,2560,0.005974222388532426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,1536,8192,0.020655110478401184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,12288,64,0.0069288888739215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,16384,128,0.006015111174848344
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1024,10240,0.011150221857759686
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,6144,32,0.005611555443869696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,3584,65536,0.0924160016907586
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,7168,10240,0.03421866562631395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,256,3072,0.007324444750944774
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,12288,3584,0.018087110585636564
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3072,256,0.005586666779385672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,16384,512,0.09282133314344619
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,10240,64,0.011408000356621213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,2560,8192,0.031503111124038696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,10240,256,0.005980444451173146
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,256,2048,0.00490399988161193
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,8192,2048,0.022096888886557683
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,3072,6144,0.03321866525544061
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,128,512,0.005614222337802251
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,65536,7168,0.17093155119154188
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,3584,512,0.020900444851981267
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,10240,2560,0.0170817772547404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,32,4096,0.007646222081449296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,16384,10240,0.03468622101677789
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,1536,3584,0.026371555195914373
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,2048,768,0.005972444597217772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,6144,2048,0.013244444297419654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,256,10240,0.05381955703099569
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3584,256,0.00327822214199437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,32,10240,0.011775111158688864
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,16384,7168,0.02219911085234748
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,3584,1536,0.006272000157170826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,7168,768,0.016886222693655226
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,8192,16384,0.08404799964692856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,3072,4096,0.008063111040327284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,768,2048,0.005222222043408289
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2560,3584,0.007857778006129796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,6144,1024,0.007476444045702617
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,2560,256,0.018153778380817838
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,768,5120,0.009337777892748514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,3584,1024,0.006291555447710886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,512,128,0.016757332616382174
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,5120,128,0.012025777664449481
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,256,3584,0.00981422182586458
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,128,7168,0.01015733347998725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,12288,128,0.0092613332801395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2048,32,0.0058133333093590206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,256,12288,0.02293155590693156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,3072,16384,0.020988444487253826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,65536,2048,0.06628533204396565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,1024,7168,0.009674666656388177
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,512,256,0.006609777609507243
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,64,2048,0.006959111326270633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2560,256,0.0029013332807355454
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,7168,128,0.019384000036451552
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,5120,2560,0.022999111149046157
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,4096,2048,0.008303999900817871
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2560,8192,0.010298666854699453
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,256,4096,0.007676444119877285
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1024,32,0.0053022222386466126
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2560,256,0.00795022232664956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,768,1536,0.006262222097979651
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,12288,1024,0.023050665855407715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,128,16384,0.010138666464222802
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2048,5120,0.011447110937701331
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,16384,5120,0.02647466626432207
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,2048,128,0.006286222073766921
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,3072,5120,0.018431999617152743
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,6144,3072,0.01297688898113039
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,1536,6144,0.00903911143541336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,6144,64,0.0022044444663657085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,10240,65536,0.1428577767478095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,16384,2048,0.011167999770906238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,12288,4096,0.019350222415394254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,8192,1024,0.018277333842383493
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,256,8192,0.01962844365172916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,7168,3584,0.011163555913501315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2048,768,0.005991111199061076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,65536,2048,0.023864888482623633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,64,5120,0.004927111168702443
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,1536,1024,0.014192000031471252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,10240,16384,0.03859733210669623
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,8192,1536,0.006872889068391588
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,7168,64,0.005757333503829108
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,256,16384,0.01647555496957567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,10240,8192,0.03184355629814996
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,65536,3584,0.10976177453994751
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,2048,1024,0.005970666805903117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2048,256,0.005995555470387141
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,768,8192,0.03219022353490194
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,16384,65536,0.16693600018819174
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,7168,512,0.028284443749321833
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,256,3584,0.00794222205877304
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,1024,7168,0.14280800024668375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,64,128,0.0021315556433465746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3584,16384,0.41866223017374676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,12288,8192,0.057358225186665855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2048,12288,0.018626666731304593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,7168,1536,0.022515555222829182
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,1024,10240,0.02409155501259698
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1024,12288,0.0384497774971856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,16384,512,0.005407111098368962
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,32,16384,0.005865777946180767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,4096,512,0.020929776959949072
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2560,8192,0.013594666288958656
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,65536,12288,0.28771644168429905
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,512,32,0.0022248888595236675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1536,6144,0.009190222455395592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,10240,1024,0.013424889081054263
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,5120,512,0.0053431110249625305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,512,10240,0.011882666912343768
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,768,256,0.014474666780895658
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,65536,1536,0.12920978334214953
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,7168,128,0.017265778448846605
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,768,768,0.005584888988071018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,3584,6144,0.08494755294587876
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2048,128,0.00544177782204416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,256,16384,0.005587555468082428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,128,10240,0.007670222057236566
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,3584,2560,0.006991110742092133
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,5120,3072,0.014727999766667684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,65536,16384,0.3931786749098036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,512,65536,0.03323555654949612
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,2048,16384,0.07996710803773668
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,2048,256,0.01332622187005149
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,256,10240,0.015287111202875773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,256,65536,0.07028622097439237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,10240,32,0.0069884442620807225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,1536,7168,0.016558221644825406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,3072,512,0.005623999983072281
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,1024,5120,0.008361777497662438
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3584,2560,0.008350222474998897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,32,3584,0.005748444547255834
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,128,6144,0.017675555414623685
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,512,1536,0.006284444282452266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,256,128,0.0022195554855797025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2048,2048,0.006931555353932911
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,3072,6144,0.05948355462816027
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,3072,768,0.00565422202150027
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,12288,12288,0.03177155719863044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,64,4096,0.0046533333758513136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,8192,256,0.002601777720782492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,16384,3584,0.03851022322972616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,768,768,0.005899555567238066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,7168,4096,0.010420444111029306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,8192,2048,0.007673777639865875
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1536,2560,0.006167999986145232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,12288,2048,0.011388444238238864
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,128,1536,0.0036524443162812125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,6144,6144,0.027456000447273254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3072,65536,0.06372000111473931
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,8192,3072,0.013029333617952136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,6144,65536,1.7334701750013564
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,2560,6144,0.027535999814669292
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,7168,32,0.00416711096962293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,8192,8192,0.0337066650390625
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,256,1536,0.0038577777643998465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,1024,5120,0.022452443838119507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,2048,5120,0.013003555436929068
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,6144,2560,0.04080355498525832
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,16384,16384,0.049977779388427734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,7168,64,0.0020871110674407748
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,256,8192,0.016959110895792644
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,8192,3584,0.010984888507260216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,10240,128,0.019517333971129525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3072,128,0.005590222362014983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3584,1024,0.006039999839332368
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,512,3584,0.005374222166008419
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,3072,768,0.005616888817813661
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,7168,256,0.005628444254398346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,1536,768,0.005281777845488654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,768,5120,0.008665777742862701
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,8192,2048,0.02647022240691715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,768,1536,0.006278222219811545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,3072,12288,0.018239999810854595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,512,768,0.013459555804729462
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,7168,3072,0.008003555238246918
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,1536,2048,0.006603555546866522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,512,128,0.01312355531586541
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,7168,3072,0.05494399865468343
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,3584,768,0.005976888868543837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,16384,10240,0.7755040062798394
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,512,5120,0.020589333441522386
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3584,6144,0.023176888624827068
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,12288,3072,0.12987021605173746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,3584,12288,0.018292443619834054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,16384,4096,0.02144266664981842
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,256,2560,0.015098666151364645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,1024,10240,0.011775111158688864
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,1024,5120,0.019973332683245342
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,8192,256,0.009110222260157267
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,65536,1024,0.01739911072784
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,12288,8192,0.027523555689387854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3072,128,0.005707555760939916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,7168,64,0.005952888892756567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,12288,5120,0.20002755853864881
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,1024,3584,0.008005333443482717
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,16384,3072,0.026456000076399908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,2560,12288,0.05065955718358358
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,768,3072,0.01721244388156467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,512,4096,0.017509332961506315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3584,5120,0.010042666561073726
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,4096,12288,0.01720266706413693
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,3584,65536,0.9810738033718533
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,512,32,0.007310222420427535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,65536,2048,0.7327849070231119
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,64,64,0.00868444475862715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,4096,1536,0.008025777836640676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3072,8192,0.02128000060717265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,512,768,0.005606222069925732
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2560,768,0.0035235554807715942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,3072,1536,0.015196444259749519
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,6144,16384,0.03715111149681939
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,4096,16384,0.03541422221395705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,512,512,0.008378666308191087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,2560,3584,0.022426666484938726
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,6144,5120,0.026421333352724712
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,1536,8192,0.09830933147006565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,64,8192,0.010810666614108615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1536,1536,0.006628444625271692
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,4096,10240,0.020534222324689228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,16384,2048,0.015126221709781222
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,768,128,0.0021911110315057966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,12288,512,0.020644444558355544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3584,768,0.0038604442444112566
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,512,2560,0.0053484443989064955
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,3072,65536,0.9261030620998807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,6144,2560,0.15104356076982287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,256,8192,0.021021333005693223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,2048,512,0.005558222118351195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,256,64,0.006664000037643645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,256,4096,0.014866666661368476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,6144,256,0.016864889197879367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,512,2560,0.006911110960774952
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,3584,6144,0.01841600073708428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,6144,2048,0.019496000475353666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,32,3072,0.007245333658324347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,128,32,0.0018951110541820526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,4096,512,0.014846222268210517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,3072,2560,0.019746666153271992
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,128,2560,0.004872000051869286
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1024,16384,0.1006773312886556
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,12288,128,0.0031724443866146933
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2560,16384,0.020478222105238173
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,7168,5120,0.011487111449241638
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,512,5120,0.017817777064111497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,10240,3072,0.03123377760251363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,256,2560,0.006980444822046492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,32,32,0.00536088893810908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,768,8192,0.01075644459989336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,7168,12288,0.2615191141764323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,65536,3584,0.05693511168162028
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2048,512,0.0063680000603199005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,512,3584,0.06794222195943196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,16384,2560,0.03435466686884562
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2560,10240,0.018998222218619455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,8192,512,0.005946666830115848
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,12288,12288,1.1461653179592557
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,1536,768,0.005604444278611078
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3584,1024,0.0032106666929192017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,16384,16384,0.5078142484029134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,256,5120,0.015543111496501498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,256,128,0.003264888914095031
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,2560,12288,0.02160177793767717
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,8192,4096,0.03335288829273648
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,16384,2560,0.04019644525316026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,16384,16384,0.10723644495010376
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,768,4096,0.01720977822939555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,64,12288,0.013491555220550962
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,7168,32,0.03837066557672288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,1024,5120,0.010271111296282874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,3584,2560,0.007316444483068254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1536,32,0.006004444426960415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,16384,2560,0.01237066669596566
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,65536,10240,0.4355448881785075
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,65536,4096,0.05400177836418152
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1536,768,0.0032417778339650896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,5120,2560,0.007104888558387756
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,768,3584,0.017317333155208163
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1024,12288,0.027614222632514104
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,6144,3584,0.0229102224111557
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,512,512,0.005347555710209741
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,3584,1536,0.02215822206603156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3072,3072,0.007643555601437886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,16384,5120,0.027790221903059218
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,8192,7168,0.07509511046939425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,768,512,0.01442399952146742
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,8192,32,0.0021991110924217436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,16384,1024,0.007288888924651676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,4096,12288,0.020446222689416673
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,5120,7168,0.012844444149070315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3072,2560,0.007263999846246507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2048,7168,0.008380444513426887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,5120,256,0.006667555620272954
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3072,1024,0.0062773335311147906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,768,64,0.003003555453485913
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1024,32,0.002532444480392668
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,8192,3584,0.09651022487216526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,768,65536,0.020486222373114694
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,768,5120,0.008340444829728868
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,1536,10240,0.0811377763748169
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,7168,2560,0.049031111929151744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,2048,10240,0.0767422252231174
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,256,2048,0.008651555412345463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,12288,256,0.0107342220014996
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3072,3072,0.010326222413116032
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,6144,2048,0.007092444433106317
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,128,128,0.005745777653323279
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,32,12288,0.009860444400045607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,1024,256,0.0052426668504873914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,768,128,0.003181333343187968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1536,65536,0.06713600291146173
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,128,1024,0.012314666476514606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2048,32,0.0031662223239739737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,6144,4096,0.014477334088749357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,512,2560,0.005250666704442766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,5120,3072,0.008044444852405125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,64,512,0.00832444429397583
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,128,2560,0.007300444775157505
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,3584,6144,0.022023111581802368
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,128,3584,0.0052622221410274506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,32,5120,0.004592888885074192
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2048,64,0.005693333430422678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,4096,1024,0.004912888838185204
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,4096,2560,0.012134222520722283
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3072,4096,0.014709333578745524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1536,256,0.0025759999536805684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,512,1024,0.0052764444715446895
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,65536,2560,0.03471022182040744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,12288,8192,0.025891555680169
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,256,10240,0.009768000079525841
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,3584,65536,0.0695706672138638
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,1536,6144,0.01372977760103014
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,512,5120,0.008234666453467475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,7168,10240,0.13439555962880453
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1024,7168,0.018967999352349173
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,1536,3584,0.018111111389266122
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,256,7168,0.009663110805882348
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,16384,3584,0.0539244446489546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,2048,6144,0.032636443773905434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,256,3072,0.005587555468082428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,12288,512,0.004595555365085602
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,65536,8192,2.4688470628526473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,2560,1536,0.0063164445261160536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,512,4096,0.01272266689274046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3584,16384,0.29725244310167104
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,1536,6144,0.027456889549891155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3072,64,0.002472888885272874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,128,1024,0.008608889248636033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2560,64,0.012390221986505719
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,65536,1024,0.021292444732454088
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,2048,16384,0.040729777680502996
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,5120,2048,0.033230221933788724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,16384,3072,0.15196088949839273
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,10240,6144,0.02260977857642704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,12288,512,0.01961866683430142
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,256,768,0.013059555656380124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,2560,2048,0.006239999913507038
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,7168,1536,0.0185253338681327
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,512,32,0.002135111019015312
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,2048,1536,0.006218666831652324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,10240,65536,0.44565333260430234
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,768,3584,0.017504889104101393
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,512,7168,0.005225777626037598
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3584,8192,0.016141333513789706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,8192,10240,0.021368889345063105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,12288,10240,0.3499146567450629
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,12288,8192,0.023790222075250413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,1024,3072,0.007367110914654202
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,7168,256,0.007998222278224098
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2048,7168,0.021946666969193354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1536,2048,0.006629333313968446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,3072,3072,0.02403555479314592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,65536,1024,0.018106665876176622
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,2048,8192,0.02232088810867733
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,6144,8192,0.10043999883863662
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,64,768,0.007338666253619724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,4096,16384,0.022791999909612868
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1536,7168,0.01272266689274046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,7168,3584,0.01963733302222358
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,3072,12288,0.030054221550623577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,128,10240,0.020067556036843192
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3584,512,0.0036151111125946045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,128,512,0.005487111293607288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,7168,7168,0.02609688871436649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,16384,12288,0.06667377551396687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,4096,7168,0.015188443991872998
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,1536,2560,0.015188443991872998
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,64,3584,0.020657777786254883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,256,8192,0.006033777776691649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,4096,1024,0.018696889281272888
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,256,7168,0.017893332574102614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,512,64,0.005593777944644292
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,128,1536,0.015287111202875773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,16384,128,0.054584887292650014
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,5120,6144,0.03587466809484694
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,12288,3072,0.015644444359673392
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,10240,4096,0.025591999292373657
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,512,128,0.0052720002002186244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,4096,3584,0.007631111476156447
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,512,64,0.005636444522274866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,8192,512,0.016903110676341586
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1536,512,0.00254844439526399
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1536,768,0.007937777373525832
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,8192,7168,0.017111111018392775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,64,6144,0.009328888522254096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,10240,5120,0.029222223493787978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3072,12288,0.018960000740157235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,5120,128,0.005611555443869696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,768,16384,0.0162435554795795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,2560,2048,0.006296000132958095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,4096,512,0.003022222262289789
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,1024,6144,0.018956444329685636
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2560,65536,0.11147911018795437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,12288,12288,0.047388444344202675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,8192,3072,0.009384888741705153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,16384,1536,0.009382222261693742
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2048,65536,0.27702222930060494
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,128,1536,0.017532444662517972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,256,1536,0.006320888797442119
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,8192,12288,0.041125333971447416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3072,3072,0.008904000123341879
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,8192,3584,0.017670222454600863
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,512,7168,0.011128889189826118
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,4096,16384,0.07050755288865831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3584,3072,0.0076142218377855085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,512,2560,0.034176889393064715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,5120,65536,0.11985777484046088
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,1536,3072,0.010079111489984725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,10240,2048,0.008551999926567078
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2048,7168,0.011078221930397881
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,6144,512,0.016893333858913846
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,64,32,0.0019422223170598347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1536,12288,0.09555911355548435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,8192,1536,0.01958311100800832
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1024,12288,0.01201333353916804
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,4096,512,0.0038159998754660287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,10240,256,0.00332266671790017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,65536,32,0.09433333079020183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,512,1024,0.009312888814343346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,7168,128,0.01514755520555708
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,128,4096,0.008028444316652086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,32,4096,0.0046862223082118565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,32,2048,0.00904799997806549
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,128,7168,0.017287110288937885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2560,2560,0.006562666760550604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,2048,512,0.005550222264395819
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,4096,256,0.008032889001899296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,6144,1024,0.006907555378145642
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,3072,16384,0.02215555475817786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2048,8192,0.0199626667631997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,16384,1024,0.008727110922336578
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,6144,1536,0.006301333506902059
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,256,3584,0.006046222315894232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,7168,256,0.0053591111467944244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,12288,32,0.0028782222006056043
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,128,1024,0.004966222163703707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,2560,16384,0.04285688863860237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2560,768,0.0038737778862317405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,5120,5120,0.017794667018784415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,1536,256,0.005280888742870755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,768,64,0.005684444473849402
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,65536,7168,0.13026132848527697
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3584,3072,0.02883022361331516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,256,1024,0.028175112273957994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,8192,1024,0.00572533326016532
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,2560,256,0.0052986666560173035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,12288,6144,0.033702221181657575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,128,8192,0.01661777827474806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1024,2048,0.0069439998931354946
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,10240,16384,0.12008889516194661
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,1024,7168,0.023194667365815904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,3072,5120,0.013898667362001208
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,10240,512,0.006641777853171031
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,10240,768,0.01790844400723775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,10240,1536,0.009144888983832465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,6144,512,0.012216000093354119
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,10240,2048,0.05226044522391426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,6144,8192,0.08428444465001424
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,12288,12288,0.0337759984864129
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3584,64,0.007992000215583378
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,3072,4096,0.018037334084510803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,6144,128,0.014862222803963555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,32,768,0.002974222310715251
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,7168,128,0.007327111230956183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,65536,3072,1.039907561408149
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,64,5120,0.018587556150224473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,3072,2048,0.02779200010829502
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,7168,128,0.047037333250045776
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1536,3072,0.016442666451136272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,7168,768,0.005956444475385878
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,6144,12288,0.035989334185918175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,64,128,0.0021866667601797315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,6144,256,0.002248888835310936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,12288,16384,0.06343111064698961
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,1536,16384,0.03529511226548089
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,5120,256,0.03701333204905192
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,5120,256,0.016914667354689706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,65536,6144,0.12161244286431207
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,256,128,0.013091555900043912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,128,65536,0.12527999613020155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,512,6144,0.03978666663169861
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,3584,2560,0.05410755674044291
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,8192,4096,0.032695998748143516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,256,5120,0.01647733317481147
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,10240,2560,0.01258044441541036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,16384,7168,0.0221368885702557
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,6144,6144,0.03974577784538269
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2048,256,0.0052622221410274506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,4096,8192,0.013383111192120446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,2048,256,0.01520088811715444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1536,5120,0.009008888569143083
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,5120,256,0.002587555597225825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,2560,10240,0.01277599980433782
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,4096,3584,0.044159111049440175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,5120,768,0.005995555470387141
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,64,3584,0.004936888813972473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,768,5120,0.00908622228437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,128,5120,0.01276711126168569
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,4096,2560,0.007048000064161088
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,64,65536,0.09785599841011895
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,3584,3072,0.0255084451701906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,6144,768,0.04983199967278374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,8192,3584,0.010024000373151567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,64,2048,0.004247111164861255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,8192,1024,0.005283555636803309
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,1024,256,0.014848000473446317
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,6144,16384,0.03669333457946777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,10240,2560,0.011185777684052786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,7168,256,0.03031022349993388
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,64,2560,0.007660444411966536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,16384,3584,0.03372888763745626
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,256,1024,0.0075093333919843035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,12288,128,0.002893333426780171
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,1024,256,0.014916444818178812
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,3072,4096,0.05436266793145073
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,64,7168,0.01953066719902886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1024,1536,0.019183110859658983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,1536,8192,0.010597333312034607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,7168,2560,0.007338666253619724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,32,10240,0.012208888928095499
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,64,256,0.005658666706747479
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,64,3584,0.00775733341773351
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,1024,512,0.018906666172875296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,64,2048,0.006953777538405524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,768,10240,0.18382133377922905
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,12288,10240,0.028065777487225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,128,5120,0.01585422290696038
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,65536,4096,0.11396800147162543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,32,2048,0.00462400002612008
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,7168,16384,0.026943999860021803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,12288,3584,0.02810844447877672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,128,7168,0.019479999939600628
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,16384,768,0.006940444310506185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,256,4096,0.016167999969588388
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,6144,768,0.021012443635198805
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,5120,10240,0.02511111067401038
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,2048,7168,0.033668445216284856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,65536,4096,0.0401760008600023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,7168,5120,0.01372977760103014
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,128,512,0.002536000063021978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,4096,6144,0.013223999904261695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3072,512,0.0035253332720862497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,64,1536,0.007908444437715743
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2048,64,0.005600888695981767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,3584,1536,0.006562666760550604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,256,10240,0.05158755514356825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,10240,8192,0.037703110112084284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,64,16384,0.00600622221827507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,768,2560,0.01690933273898231
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,7168,16384,0.025288888149791296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,1024,6144,0.024241778585645888
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,64,1536,0.003645333151022593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,512,2048,0.008642666869693333
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,5120,16384,0.037984887758890785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,8192,12288,0.0207306659883923
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2048,1024,0.0047733332547876565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,65536,10240,0.09679200251897176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,256,3072,0.01477599971824222
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,5120,8192,0.016162667009565566
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,256,3072,0.015488000379668342
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,16384,4096,0.3793191115061442
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,3584,1536,0.013120888835854001
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,128,16384,0.015467555986510383
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,4096,768,0.01869333287080129
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,256,256,0.0069217777086628815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3072,64,0.0053484443989064955
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,128,1536,0.012840000291665396
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,64,2048,0.0048942222363419
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,768,256,0.005603555589914322
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3584,256,0.005892444401979446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,1024,16384,0.05121689041455587
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,16384,65536,0.1735831101735433
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,16384,2560,0.38439732127719456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,768,4096,0.006974221931563483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,768,12288,0.01721600029203627
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,7168,3584,0.009835555321640438
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,6144,768,0.005899555567238066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3584,10240,0.02422844370206197
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,32,128,0.005987555616431766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2048,7168,0.01017066670788659
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,768,3072,0.025599110457632277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,12288,12288,0.0678115553326077
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3072,768,0.005591111050711737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,16384,2560,0.014808000789748298
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,65536,4096,2.058880911933051
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,10240,128,0.021947556071811255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,5120,2048,0.00832266691658232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,12288,1536,0.015819556183285184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,3072,512,0.02033955521053738
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,128,16384,0.02643377747800615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,10240,10240,0.03682488865322537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,16384,65536,2.2592061360677085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3584,128,0.013105777402718862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1024,6144,0.07551466756396823
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,768,128,0.0021964444054497615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,6144,12288,0.044807109567854136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1536,16384,0.019949333535300363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,65536,128,0.00755555596616533
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,10240,2560,0.06851822137832642
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,4096,256,0.015881778465376962
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,5120,512,0.002599999929467837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,768,3584,0.007713778151406183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,512,1536,0.006646222124497096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,10240,2560,0.012808000048001608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,128,32,0.002209777840309673
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,8192,3584,0.07482400205400255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,16384,3584,0.01461955573823717
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,8192,4096,0.010957332948843638
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,1536,12288,0.023429332507981196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,65536,256,0.010533333652549319
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,5120,1536,0.006343110981914733
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,512,10240,0.17445511288113066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,128,256,0.005423999908897612
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2560,7168,0.011810666984981961
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,4096,3584,0.05841422080993652
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2048,128,0.0069155556460221606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3072,6144,0.011785777906576792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,65536,3072,0.09936800267961289
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,7168,512,0.005314666777849197
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,65536,6144,0.05262488789028592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,8192,1536,0.05897777610354953
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,6144,10240,0.023009777069091797
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2560,2048,0.006606222026877933
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1536,2048,0.00872444444232517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,7168,2048,0.06685955656899346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3072,64,0.005570666657553778
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,3072,5120,0.08299110995398627
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,64,7168,0.004927111168702443
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,10240,65536,0.1983333296245999
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,16384,2560,0.01239822225438224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,10240,2560,0.012072888513406118
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,1536,3584,0.0076657773719893555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,1024,128,0.005520888914664586
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,10240,65536,0.18799022833506265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,512,6144,0.02092711130777995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3072,2560,0.007272889216740926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2560,16384,0.21895021862453887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2560,6144,0.012170666621790992
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,10240,512,0.0069573331210348345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,768,3072,0.0070568886068132185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,8192,32,0.006464889066086874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,768,8192,0.01074399964676963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,8192,5120,0.014013333453072442
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,2560,128,0.014846222268210517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,7168,4096,0.012638222012254926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,4096,128,0.014767110347747803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,1024,65536,0.06420533524619208
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,1536,7168,0.05113510953055488
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,5120,1024,0.006676444576846228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,65536,1024,0.07766222291522555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,65536,128,0.019307555423842538
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1536,16384,0.019676445258988273
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3584,128,0.01147911118136512
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,6144,256,0.016957332690556843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,7168,2560,0.022022222479184467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,768,3584,0.007719111111429002
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,4096,32,0.005912000106440649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,10240,4096,0.0292559994591607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,8192,5120,0.011500444677140979
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,16384,1536,0.009330666727489894
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,64,768,0.002826666666401757
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,10240,6144,0.10038399696350098
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1536,32,0.013806222213639153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,64,7168,0.007373332977294922
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,12288,1536,0.026683555708991155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,512,16384,0.023141332798533972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,256,12288,0.00599911105301645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,65536,512,0.05765689081615872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,10240,3072,0.011751111182901593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,12288,1024,0.023393778337372675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2048,4096,0.008285333712895712
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,8192,65536,0.1489528947406345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,12288,2048,0.05796977546479967
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,512,256,0.014480888843536377
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,768,10240,0.012122666670216454
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,65536,3072,0.037294222248925105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,5120,256,0.018363555272420246
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,12288,5120,0.0413333343134986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,2048,1024,0.016435555285877652
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,1024,16384,0.01621955633163452
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,8192,5120,0.25932712025112575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,4096,5120,0.015640000502268474
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,8192,5120,0.027564444475703772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,512,2560,0.007096889118353526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,256,2048,0.00941066692272822
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,128,7168,0.02986488739649455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1536,65536,0.03967466619279649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,32,1024,0.007638221813572778
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,10240,256,0.005958222266700532
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,768,7168,0.018928888771269057
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,7168,1024,0.0059262220230367445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,16384,10240,0.030631999174753826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,64,10240,0.023722666833135817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,6144,256,0.005647111270162795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,4096,128,0.002505777817633417
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,512,2048,0.006622222148709827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,65536,7168,0.1741422282324897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,4096,16384,0.026541334059503343
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,8192,2048,0.011800000237094032
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,65536,6144,0.5288515620761448
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,3584,4096,0.011511999699804517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3072,768,0.007330666813585494
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,256,7168,0.02142133315404256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,1536,7168,0.028421332438786823
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,5120,12288,0.021115556359291077
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,3072,12288,0.03011200163099501
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,512,128,0.013068444199032254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,8192,32,0.003177777760558658
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,3584,768,0.04332888788647122
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,256,1536,0.003822222352027893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,32,512,0.0028479999552170434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,128,128,0.010741333166758219
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,32,3584,0.004918222212129169
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,256,8192,0.010763555765151978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,65536,2048,0.023377777801619634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,16384,10240,0.037250667810440063
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,512,3072,0.005601777798599667
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,1024,16384,0.04876177840762668
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,128,6144,0.011066666907734342
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,128,5120,0.004944889081848992
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,2048,65536,0.08061066601011488
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,256,6144,0.005937777873542573
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,7168,2048,0.007873777714040544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,5120,2048,0.010302222437328762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1024,2048,0.0069155556460221606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,16384,12288,0.03550666570663452
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,32,2560,0.007305777735180325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3584,16384,0.023017777336968318
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,32,64,0.005626666463083691
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,256,8192,0.0052844443255000645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,128,10240,0.0059279998143513995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,768,768,0.016387555334303115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,7168,1536,0.006258666515350342
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,512,64,0.002256888896226883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,4096,8192,0.013521778086821238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,8192,12288,0.025443555580245122
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,10240,5120,0.02088977727625105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,3072,12288,0.03583822316593594
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2560,4096,0.00903200027015474
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,16384,2560,0.18675555123223198
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,10240,3584,0.016338666280110676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,7168,256,0.005361777626805835
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2560,8192,0.013771555489963956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,4096,6144,0.013864888913101621
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,12288,6144,0.22973956002129448
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,16384,5120,0.017729777428838942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,512,1536,0.0063164445261160536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,6144,4096,0.026311111119058397
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,4096,2560,0.01569955547650655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,16384,8192,0.04806222187148201
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,10240,1536,0.008395555946562026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,2560,1024,0.015814221567577787
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,512,4096,0.005759999983840519
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,256,3072,0.005648889061477449
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,512,32,0.00600355573826366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3584,2560,0.008729777402347988
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,4096,5120,0.01183377785815133
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,10240,10240,0.0255022214518653
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,10240,10240,0.05188622077306112
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,10240,32,0.0029057777590221832
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,64,128,0.008062221937709385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1024,2048,0.014660444524553088
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,65536,512,0.011770666473441653
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,3584,5120,0.026758222116364375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,512,16384,0.010326222413116032
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,1024,12288,0.04701155424118042
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,128,1536,0.003555555724435382
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,65536,6144,0.051348444488313466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,5120,16384,0.02383199996418423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,5120,16384,0.06068888637754652
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,16384,5120,0.017608889275126986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,2560,3072,0.01826755536927117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,6144,512,0.016551999582184684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,2560,256,0.007339555356237624
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,256,7168,0.00962400022480223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,12288,1024,0.008061333662933773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,3072,65536,0.06578133503595988
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2560,2560,0.011542222566074796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,5120,5120,0.012864000267452665
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,7168,1024,0.015170666906568738
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,16384,6144,0.03077510992685954
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,7168,256,0.029666665527555678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,12288,4096,0.014521777629852295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2048,768,0.007338666253619724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,768,12288,0.02581866747803158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2048,7168,0.012451555165979596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,1024,6144,0.017603556315104168
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,7168,7168,0.020218667056825426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,1536,256,0.015609777635998197
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2560,3072,0.036863111787372164
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,5120,12288,0.018262222409248352
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,65536,8192,0.09568978018230861
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2560,32,0.010053333308961656
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,10240,7168,0.02604977786540985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,128,2048,0.013502221968438892
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,256,3072,0.010441777606805166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,10240,2560,0.011762667033407422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2048,2048,0.0064151109092765385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2560,8192,0.15882489416334364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,32,16384,0.008058666355080074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,4096,256,0.0059279998143513995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,1536,512,0.009648888475365108
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,3584,6144,0.03994311226738824
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,512,256,0.014467555615637036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,1024,1536,0.006188444379303191
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,1536,256,0.014838222000333997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,256,512,0.011701333026091257
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,7168,8192,0.2005831135643853
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,6144,1024,0.009372444616423713
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,3072,128,0.0053004444473319584
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,256,512,0.0026604444202449587
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,5120,65536,0.10248800118764241
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,8192,2560,0.01181422256761127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,16384,10240,0.2302791012658013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,1536,12288,0.05208622084723579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,4096,1536,0.02556177808178796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,128,3072,0.018240888913472492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,65536,1024,0.040157334672080144
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,8192,2048,0.02409066590997908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,8192,512,0.005944889038801193
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,2048,3584,0.008012444608741337
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,8192,1024,0.005941333456171884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,65536,2048,0.04749155706829495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,768,512,0.005948444621430502
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,128,12288,0.013119999733236102
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,768,12288,0.01346577786737018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,6144,65536,0.1449751059214274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,2048,8192,0.01184088902340995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,64,5120,0.009029332962301042
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,8192,3584,0.28399909867180717
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,65536,768,0.02313599983851115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,2048,16384,0.02128622266981337
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1024,1024,0.004535111288229625
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,6144,3584,0.02949777907795376
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,65536,128,0.022269333402315777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,32,10240,0.005642666584915585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,16384,1536,0.1216497818628947
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,256,3584,0.005882666756709416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1536,64,0.0018586667461527719
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2560,128,0.0022355556074115965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,7168,512,0.017619555195172627
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,64,3072,0.004545777622196409
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,768,1536,0.006242666807439592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2048,65536,0.5670151180691189
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,512,7168,0.035663998789257474
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,6144,768,0.0034977777136696708
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1024,4096,0.006950221955776215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,2048,16384,0.08050577508078681
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,1536,2048,0.054126222928365074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,16384,32,0.006233777850866318
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,2560,768,0.0059262220230367445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,5120,3584,0.007688889073001013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,768,1536,0.0038275557259718576
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,32,12288,0.01314133322901196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,64,7168,0.0052622221410274506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,128,8192,0.02141155633661482
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,65536,2560,0.03202844328350491
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,768,1024,0.005635555419656966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,128,5120,0.015467555986510383
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,256,1024,0.006243555496136348
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,128,64,0.0019137778629859288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,256,8192,0.005684444473849402
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,4096,768,0.003192000091075897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,256,32,0.005945777727497949
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2048,768,0.003212444484233856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,65536,8192,0.08985333310233222
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,65536,3584,0.056222220261891685
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2048,3072,0.0073724447025193115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,512,12288,0.009377777576446533
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,10240,768,0.018241778016090393
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,4096,6144,0.013947554760509066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3072,6144,0.01717333330048455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,7168,2560,0.02245866590076023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,10240,3072,0.013070222404268054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,2560,1536,0.006055111272467508
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,64,256,0.0022373333987262514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,8192,16384,0.026514665948020086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2048,12288,0.017435555656750996
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,10240,16384,0.57259734471639
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,5120,1536,0.00981955561372969
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1024,65536,0.08728977706697251
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,1024,3072,0.017609778377744887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2048,3072,0.00979555563794242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,7168,2560,0.011935999823941125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,256,2048,0.006730666591061487
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,12288,4096,0.014208889669842191
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,512,4096,0.016586666305859882
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,768,2560,0.056286222404903836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,4096,1024,0.015172445111804538
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,6144,10240,0.03184533450338576
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,12288,65536,0.1466533342997233
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,512,1024,0.0032231110251612137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,1536,1536,0.007801777786678738
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1024,4096,0.006990222467316522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,32,1024,0.005927111125654644
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1024,6144,0.009188444250159794
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,6144,512,0.005936000082227919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,64,16384,0.015232000086042615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,10240,1536,0.008815111385451423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1024,1024,0.011746666497654386
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,768,1536,0.006451555424266391
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,512,5120,0.008332444561852349
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,4096,3072,0.00757422215408749
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,64,512,0.002542222125662698
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,768,3072,0.015186667442321777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,64,16384,0.017087111870447796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,16384,3584,0.014361777239375643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,6144,3584,0.14659644497765437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,7168,2048,0.008073777788215214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3584,4096,0.06076000134150187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,8192,512,0.055697778860727944
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2560,3584,0.025622222158643935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,1024,3072,0.015511110424995422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,128,8192,0.024085332949956257
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,512,8192,0.02126577827665541
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,3072,128,0.005780444376998478
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,5120,2048,0.008061333662933773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,256,2560,0.007270221908887227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,5120,2560,0.018960888187090557
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,256,12288,0.016543999314308167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,1536,768,0.03440711233350966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,256,65536,0.0587386687596639
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,10240,5120,0.017301332619455125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2560,16384,0.06000355879465739
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,65536,32,0.009925333162148794
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,3072,5120,0.03016622198952569
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,10240,6144,0.02051377793153127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,4096,3584,0.011048888994587792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,256,64,0.0018497777895794974
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,512,512,0.005589333259397083
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,2560,2560,0.0069724445541699724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,3072,5120,0.02574133376280467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,16384,4096,0.0221413340833452
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,128,6144,0.018024888303544786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,3072,768,0.005643555687533484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,5120,1536,0.006656888872385025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,512,256,0.0059262220230367445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,16384,3072,0.013501333693663279
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,768,3072,0.007274666594134436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,768,2048,0.007315555380450354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,8192,32,0.009723555710580613
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2560,16384,0.020283554991086323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,3584,4096,0.008033778104517195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,512,10240,0.03918577896224128
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,4096,12288,0.017479999197853934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,16384,512,0.05728444125917223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,8192,256,0.0028782222006056043
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2560,7168,0.014084445105658637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,7168,2048,0.06836889187494914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,8192,8192,0.015405333704418607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,2048,7168,0.02482755482196808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,7168,2048,0.02167555524243249
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,32,7168,0.010055999788973067
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,32,16384,0.005623111294375525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,65536,5120,0.04905866583188375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,128,1024,0.006280000011126201
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,768,6144,0.009343999955389235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1024,2560,0.0069520001610120135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,10240,5120,0.3003591166602241
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,128,12288,0.005629333357016246
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1024,5120,0.0052639999323421055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,1536,3584,0.02398577829202016
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,4096,768,0.005550222264395819
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,12288,512,0.006677333265542984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3584,1024,0.007567110988828871
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,64,5120,0.0086666668454806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,1024,16384,0.030396444929970637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,8192,7168,0.026994667119450037
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,10240,8192,0.042698668109046094
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,4096,32,0.005240889059172736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,2048,10240,0.012740444805887012
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,32,3072,0.00490133340160052
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,64,256,0.006238222122192383
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,2560,12288,0.015596444408098856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,7168,2560,0.008007999923494127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,65536,512,0.01442488862408532
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,3072,1024,0.005913777897755305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,65536,5120,0.9449813630845812
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,32,16384,0.007681777907742395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,8192,512,0.008956444760163626
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,65536,2048,0.07719288931952582
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,3584,7168,0.010517333116796283
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,10240,2560,0.02990844514634874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,512,64,0.002480000050531493
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3584,10240,0.038206222984525896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,256,12288,0.059939556651645236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,7168,8192,0.18014221721225313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,3072,3584,0.01688088807794783
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,512,4096,0.0207315550910102
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,12288,3072,0.014474666780895658
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,8192,8192,0.01640533407529195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,768,7168,0.020977778567208186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,128,2048,0.006628444625271692
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,768,2048,0.004283555679851108
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,65536,10240,0.18886310524410674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,8192,2560,0.026124444272783067
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,16384,512,0.003998222036494149
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,6144,7168,0.022729777627521094
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,6144,8192,0.03317066695955064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,7168,16384,0.025946666797002155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,12288,4096,0.036116444402270846
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,32,32,0.006984000404675801
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,5120,4096,0.010816888676749336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1536,6144,0.010955555571450127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,3584,1536,0.015849777393870883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,8192,8192,0.03467199868626065
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,12288,2048,0.02551022171974182
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,256,12288,0.01330933305952284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2048,12288,0.01756977703836229
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,768,8192,0.010965333216720156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,65536,12288,0.2766062153710259
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,2560,2048,0.006720888945791457
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,8192,1536,0.0063440000845326324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1024,6144,0.006824000014199152
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,2560,3584,0.016480889585283067
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,3072,128,0.007706666986147563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,512,768,0.0074924445814556545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,8192,5120,0.016529776983790927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,12288,256,0.0063031112982167145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,65536,8192,0.2081448899375068
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2048,5120,0.009275555610656738
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,128,5120,0.005576888720194499
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,512,2048,0.006647999915811751
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,12288,12288,0.050322665108574756
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,4096,16384,0.07252622312969632
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,10240,16384,0.04035910964012146
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,32,10240,0.005963555640644497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,5120,12288,0.23776266309950086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,256,7168,0.020011555817392137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,256,2048,0.0069351109365622205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,7168,10240,0.07185688945982191
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,7168,6144,0.02381244467364417
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,32,1536,0.008300444318188561
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,1024,5120,0.017200888858901132
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2560,256,0.002850666642189026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,3584,5120,0.010013333625263637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,16384,128,0.006592000110281839
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1024,12288,0.013090666797426013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,8192,3584,0.010420444111029306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,128,7168,0.005598222215970357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,3072,12288,0.016732444365819294
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,3584,12288,0.19698222478230795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,6144,16384,0.03640444411171807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,5120,512,0.00564177789621883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,1024,768,0.01651822196112739
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,16384,128,0.006262222097979651
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2048,2048,0.0045422220395671
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,6144,8192,0.026013332936498854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,768,8192,0.010479999913109673
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,4096,12288,0.03271555569436815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,12288,8192,0.15981954998440212
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2048,8192,0.04026222229003906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,5120,2048,0.0069955554273393415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,768,4096,0.008005333443482717
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,768,1536,0.004179555508825514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,1024,10240,0.030202666918436687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,4096,64,0.0058319999112023254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,12288,512,0.0070062221752272705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,2560,4096,0.007958222594526079
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,8192,65536,0.19272532727983263
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,1024,1024,0.006057777752478917
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,6144,512,0.020000888241661918
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,6144,7168,0.045496000183953174
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,2048,4096,0.04318933354483711
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,2560,2560,0.020006222857369315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1536,16384,0.023268444670571223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,768,5120,0.01997244523631202
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,2048,3072,0.01550044450494978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,6144,2048,0.021652445197105408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,7168,1024,0.005422222117582957
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,2048,8192,0.04002844293912252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1536,1536,0.0063484443558586975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,7168,1024,0.007982222570313348
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,7168,768,0.005976000179847081
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,128,128,0.0022391111900409064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2560,1024,0.005656888915432825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1024,64,0.0026071110947264563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,12288,7168,0.030893332428402368
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,8192,65536,0.08066933022605048
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3072,64,0.0022159999029503926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,8192,65536,0.1363191074795193
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,4096,1536,0.01928444372283088
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,5120,10240,0.03679644399219089
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,6144,1536,0.01760444376203749
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,32,1024,0.009744000103738572
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,2560,256,0.0151555554734336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,6144,10240,0.01632355484697554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,65536,2048,0.024525332782003615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,8192,5120,0.015768888923856948
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,8192,2048,0.024912888805071514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,3072,768,0.005959110955397288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,1536,10240,0.023686221904224817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,3584,2048,0.019359999232821994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2048,8192,0.08906755844751994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,3072,1536,0.006296000132958095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,1024,512,0.005636444522274866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,3072,65536,0.25457334518432617
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,3072,1536,0.006351999938488007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,2048,16384,0.030930668115615845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,512,10240,0.009486222432719337
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,12288,3072,0.01611111064751943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,7168,1024,0.005412444472312927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,2048,3584,0.007647111184067196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,768,1536,0.006298666612969504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,6144,3072,0.02162044412559933
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,12288,1024,0.007965332931942409
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,12288,7168,0.023665777511066858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,256,256,0.0058871110280354815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,8192,8192,0.024653333756658766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,4096,1536,0.006568888823191325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,8192,256,0.0057751110030545135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,65536,2560,0.04342933495839437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,768,10240,0.040954666005240545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,768,6144,0.02758488886886173
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,16384,65536,0.16799733373853895
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,256,10240,0.007877333296669854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1536,1536,0.0064586665895250105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,12288,512,0.00666133314371109
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,512,1024,0.006339555399285422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2048,256,0.002453333387772242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2560,128,0.017847999930381775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,65536,1024,0.3108168972863091
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,65536,512,0.19557510481940377
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,16384,768,0.008020444048775567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,65536,8192,0.16300977600945368
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,6144,2560,0.007984889050324758
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,8192,3072,0.008060444560315874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,2560,8192,0.011576889289749993
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3584,8192,0.11047199699613784
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,8192,65536,0.08959377474255031
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,3072,8192,0.04563822348912557
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,65536,6144,0.15504088666703966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,768,10240,0.011691555380821228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,512,768,0.015056888262430826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,4096,2560,0.02032533288002014
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2560,16384,0.0341360006067488
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,128,65536,0.1979466676712036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,3072,3072,0.02038755516211192
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,7168,16384,0.047691557142469615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,12288,7168,0.03399466805987888
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,8192,1536,0.015896000795894198
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,16384,32,0.018271999226676095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,3072,2048,0.015572445260153877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3584,7168,0.014985778265529208
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,768,8192,0.020655999581019085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,5120,2560,0.010108444425794814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,6144,128,0.014064000712500678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2048,12288,0.026320000489552815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,8192,2560,0.03635111120012071
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,768,2560,0.00534400012758043
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2560,1024,0.0075937774446275495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,6144,7168,0.1437057786517673
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,10240,3072,0.012811555630630918
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,65536,1024,0.024200000696712073
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1024,256,0.0025271111064487034
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,1024,128,0.0052915554907586836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,3072,10240,0.040927112102508545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,256,4096,0.011388444238238864
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,128,128,0.01072266697883606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,128,128,0.0021662221600612006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,768,4096,0.018912000788582694
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,10240,6144,0.043198221259646945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,16384,1024,0.0073902226156658595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,4096,8192,0.03981511129273309
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3072,1024,0.006302222195598815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,512,2048,0.004335999902751711
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,12288,10240,0.04602666695912679
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,4096,65536,0.07800533374150594
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,7168,4096,0.01260444439119763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3584,64,0.005616888817813661
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1536,256,0.005606222069925732
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1536,7168,0.011341333389282227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,12288,7168,0.13138488928476968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,128,128,0.015541333291265698
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2560,7168,0.013609777722093793
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,512,16384,0.016186666157510545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,768,1024,0.007348444726732042
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,10240,3584,0.015344889627562629
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,256,65536,0.016525333126386006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,128,16384,0.13320888413323298
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,10240,128,0.015396444333924187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,16384,1024,0.1276764472325643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,5120,3584,0.025104888611369666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,1024,65536,0.08704621924294366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2048,5120,0.008912000391218398
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,128,512,0.011396444506115384
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,4096,8192,0.030100444952646892
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,128,12288,0.1009955538643731
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,512,512,0.005793777604897817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,10240,8192,0.02922933300336202
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,10240,16384,0.04200444287723965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,768,65536,0.5036533143785265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,256,8192,0.010696888797812991
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2560,3584,0.007655111451943715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,128,1024,0.007646222081449296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1024,12288,0.04900622367858887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,5120,10240,0.0555937753783332
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1024,3072,0.013771555489963956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,65536,16384,0.22694132063123915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,128,16384,0.01682133310370975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,64,1024,0.0031759999692440033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,256,64,0.0022684445397721394
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,5120,128,0.014959999256663851
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,8192,256,0.002874666617976295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,65536,7168,0.31765156322055393
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,256,1024,0.005919111271699269
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3072,7168,0.21059466732872856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,5120,6144,0.02828622195455763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,16384,512,0.007353777686754863
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,3584,7168,0.21981688340504965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,256,32,0.003645333151022593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,12288,512,0.011777777638700275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,768,5120,0.05864622195561727
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,1536,1024,0.01632000009218852
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2560,1024,0.009708444277445475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,16384,4096,0.04283110962973701
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,768,12288,0.01569422251648373
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,4096,2560,0.015423110789722867
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,6144,65536,1.6187369028727214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,12288,64,0.0032835555159383347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,2048,7168,0.024041778511471216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,256,5120,0.011317333413494958
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,768,2048,0.03247911069128249
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3584,2560,0.006866666591829724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,10240,2048,0.009741333623727163
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,10240,10240,0.10278311040666367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,3072,512,0.014488000008794995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,5120,5120,0.03359555535846286
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,4096,128,0.005255110975768831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,512,12288,0.026025777061780293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,5120,256,0.005507555686765247
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,8192,65536,0.2608444425794813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,10240,512,0.006301333506902059
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,512,65536,0.07378311289681329
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,32,10240,0.013020444247457715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,2560,7168,0.03005689051416185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1536,128,0.002171555534005165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,4096,12288,0.017530666457282174
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,128,256,0.005274666680230035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,512,1536,0.0041928887367248535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,3072,7168,0.009692444569534725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,128,64,0.006310222049554189
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2048,512,0.013733333183659447
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1024,128,0.006615110983451207
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,1536,1024,0.006015111174848344
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,16384,8192,0.11567021740807427
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,12288,1536,0.009705777797434065
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,5120,10240,0.264686213599311
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,12288,4096,0.016948444975747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,65536,512,0.014500444134076437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,512,2560,0.014451555079884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,16384,1536,0.02980533242225647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,32,65536,0.05570755402247111
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,5120,3584,0.020855110552575853
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,1536,65536,1.3002995385064018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,512,2048,0.013833777772055732
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,512,32,0.005602666487296422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,2560,3584,0.016629333297411602
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,16384,10240,0.04258222381273905
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1536,3584,0.0076657773719893555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,12288,32,0.037414222955703735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,8192,8192,0.021564443906148274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,3584,2048,0.016662221815851
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3584,3584,0.009717333647939894
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,256,6144,0.019849777221679688
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,32,128,0.00590844452381134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,16384,64,0.002141333288616604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,6144,2560,0.008242666721343994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,6144,10240,0.02311022248533037
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,256,7168,0.007780444290902879
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,5120,128,0.020353777541054618
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,8192,3072,0.0236462222205268
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1024,2048,0.02474044428931342
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,1024,4096,0.01719377769364251
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,512,7168,0.010052444206343757
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,768,3584,0.007641777396202087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,6144,2560,0.007274666594134436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,768,4096,0.011392888923486074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,512,1024,0.0032951111594835916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,7168,768,0.0032559999575217566
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2560,32,0.008616000413894653
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,10240,12288,0.03357955482270982
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1536,16384,0.020776000287797716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,5120,10240,0.015708444847000968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,12288,7168,0.03480622172355652
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,5120,12288,0.020865778128306072
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,512,2560,0.005025777965784073
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,65536,64,0.011110222174061669
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3584,12288,0.016879111528396606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,16384,128,0.015490666031837463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1024,8192,0.010758221977286868
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2048,1024,0.007679999702506595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2560,64,0.0020515556550688213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,5120,2560,0.007314666277832455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,5120,3584,0.02951822347111172
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,6144,12288,0.04829688866933187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,7168,5120,0.04246755441029867
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,8192,7168,0.042437334855397545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,12288,1536,0.02845422095722622
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,12288,1536,0.010524444282054901
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,65536,1536,0.020868443780475195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,1536,3072,0.007291555404663086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,8192,1536,0.09503200319078232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,5120,65536,0.3265519936879476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,16384,10240,0.05548888776037428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,768,1536,0.01983022193113963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,3584,2560,0.006996444529957241
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,10240,7168,0.03087555699878269
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,512,768,0.0031271111220121384
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,128,1536,0.01627377834584978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,128,4096,0.004659555438492033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2048,16384,0.02086488902568817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,8192,64,0.008984000318580205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,12288,12288,0.05028088887532552
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,4096,5120,0.05539466606246101
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,512,8192,0.010728889041476779
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3072,3584,0.008634666601816813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3584,256,0.015288889408111572
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,8192,16384,0.07822399669223361
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,768,65536,0.062105774879455566
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,64,1536,0.0038008888562520346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,64,16384,0.008240888516108194
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3072,7168,0.013273778061072031
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,768,5120,0.009266667068004608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3584,1536,0.0042257776690853965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,32,512,0.002495999965402815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,16384,768,0.008298666940795051
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3072,12288,0.017858665850427415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,128,256,0.011367110742463006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,128,1024,0.002838222309947014
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,512,5120,0.01647911138004727
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2560,2048,0.006654222392373615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,3072,3072,0.016328889462682936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2560,1536,0.004048000193304486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1536,10240,0.04592355423503452
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,8192,32,0.006979555719428592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,4096,2048,0.007164444360468123
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,768,2048,0.006638222270541721
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,6144,1024,0.0037724446091387006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,768,4096,0.01937599976857503
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,6144,768,0.003212444484233856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3584,512,0.005557333429654439
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1024,64,0.005913777897755305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,256,1536,0.008081778056091731
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,65536,3072,0.03168977631462945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,5120,768,0.004087999877002505
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,12288,12288,0.0503039989206526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,512,2048,0.006641777853171031
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,128,10240,0.02236000034544203
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,768,64,0.0032382222513357797
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,512,3072,0.005017777697907554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,256,1536,0.0063075555695427795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,128,128,0.005311111195219888
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1024,16384,0.019959110352728102
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,8192,512,0.015849777393870883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,10240,256,0.0340942210621304
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,768,768,0.0035422220826148987
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,7168,1536,0.014654222461912366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,7168,16384,0.0451493329472012
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,3072,1024,0.005991999887757831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,1024,128,0.005643555687533484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3584,3072,0.009232000344329411
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,768,128,0.005934222290913264
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,1024,256,0.005635555419656966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,12288,3584,0.048743999666637845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,8192,2560,0.009463110731707679
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3072,65536,0.3611448870764838
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,6144,4096,0.030799110730489094
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,2048,256,0.0052879999081293745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,32,128,0.006244444598754247
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2560,64,0.00527022240890397
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,7168,3072,0.010288889209429422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,768,16384,0.020795555578337777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,3072,1024,0.016526222229003906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,1024,256,0.005225777626037598
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,768,7168,0.00813422186507119
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,6144,8192,0.047990222771962486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2048,7168,0.0177182224061754
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,12288,2048,0.01297688898113039
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1024,512,0.005888000130653381
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,128,10240,0.033152000771628484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,5120,3072,0.008295999632941352
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,5120,2560,0.010991111397743225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,256,10240,0.011711111499203576
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,6144,32,0.002159111201763153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,3584,12288,0.016906667086813185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,768,65536,0.06199377775192261
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3072,6144,0.011169777976142036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,128,12288,0.021957332889238994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,10240,1024,0.006903111106819577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,6144,1024,0.017906667457686532
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,6144,65536,0.31129956245422363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1536,5120,0.008300444318188561
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,10240,7168,0.01923733287387424
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,3072,5120,0.020260444945759244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2560,128,0.002543111021320025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,6144,12288,0.0651786658498976
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,512,3072,0.010095111197895473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3072,512,0.005953777581453323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,8192,1536,0.039938665098614164
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,2048,3584,0.020242666204770405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,768,64,0.007989333735571967
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,128,10240,0.012093333734406365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,65536,12288,3.9967113071017795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,5120,10240,0.044491555955674916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,512,512,0.008611555728647444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,12288,7168,0.02645422187116411
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,1536,7168,0.019648000597953796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1536,6144,0.09379200140635173
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,32,12288,0.03638933433426751
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,768,16384,0.10307821962568496
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,768,10240,0.05730933613247342
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,10240,65536,0.20987733205159506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,12288,2048,0.09817777739630805
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,65536,768,0.06557777855131361
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,5120,2048,0.006758222149478064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,12288,64,0.005877333382765452
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,4096,4096,0.009582222335868413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,256,3072,0.0069617778062820435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,5120,10240,0.017581333716710407
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1536,64,0.006238222122192383
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,8192,8192,0.016145777371194627
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,65536,7168,0.1703235573238797
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,128,1024,0.0028453332682450614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,7168,256,0.01682844426896837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,6144,6144,0.026668444275856018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2048,2560,0.028789334826999243
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,65536,16384,5.895053439670139
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,4096,65536,0.07058311171001859
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,7168,128,0.015110222829712762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2048,128,0.00553955551650789
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3072,10240,0.01623288955953386
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,16384,65536,0.1432808902528551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,4096,256,0.007996444072988298
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,256,3584,0.011443555355072021
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,12288,10240,0.040878223048316106
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3584,16384,0.022946667340066697
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,12288,4096,0.2995964421166314
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,12288,1536,0.009382222261693742
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,8192,7168,0.01516888870133294
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,10240,256,0.006611555400821898
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,8192,7168,0.014633778068754407
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,7168,16384,0.7368879848056369
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,10240,1024,0.011680000358157687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,1536,2560,0.024382222029897902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,768,16384,0.04520533482233683
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,1536,16384,0.05936533212661743
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,16384,3584,0.17977778116861978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,1536,128,0.008770667016506195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,10240,12288,0.04957688848177592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1024,16384,0.03321866525544061
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,512,2048,0.006613333192136552
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,128,6144,0.014816888504558139
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,12288,512,0.008473777936564552
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1024,256,0.005675555517276128
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1536,7168,0.010050666828950247
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,6144,512,0.008084444536103142
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,6144,8192,0.031066666046778362
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,2560,128,0.014879999889267815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,4096,8192,0.08009066846635607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,2560,1024,0.01457155578666263
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2560,768,0.0038933331767717996
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,128,512,0.002537777854336633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,64,2560,0.005076444397370021
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,16384,16384,1.8773937225341797
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,16384,3072,0.06549244456821017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,3072,10240,0.015164444843928019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2560,1536,0.006581333362393909
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,256,6144,0.010424888796276517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,3072,65536,0.17724533875783285
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1024,1024,0.006271111054552927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,10240,1536,0.008428444465001425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,768,128,0.005580444302823808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,7168,1024,0.0196426659822464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,768,10240,0.025714667307005987
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,7168,3072,0.017501332693629794
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,3072,768,0.01755288905567593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3584,2048,0.007426666716734569
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,256,768,0.005960888746711943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,256,5120,0.006616888774765863
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,3584,768,0.017610667480362784
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,768,768,0.017272889614105225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1024,32,0.00834222220712238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,4096,1024,0.015285332997639975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,768,128,0.005223999834722943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,6144,256,0.01719466679626041
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3584,6144,0.011722666521867117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,6144,512,0.024655999408827886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,32,3584,0.007649777664078607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,3072,1536,0.027973333994547527
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,12288,1536,0.00943022221326828
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,3584,256,0.016806221670574613
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,5120,8192,0.05529866615931193
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2560,3584,0.007061333292060428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,64,5120,0.00872622181971868
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,10240,5120,0.01809955636660258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,64,10240,0.004942222187916438
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,4096,3584,0.008345777789751688
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,4096,512,0.005591111050711737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,3584,5120,0.008580444587601556
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,512,2560,0.006865777903132969
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,6144,4096,0.011067555182509951
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,5120,128,0.01497244503762987
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,7168,256,0.006629333313968446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1536,6144,0.016194666425387066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,16384,512,0.006674666785531574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,8192,1024,0.008231110870838165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,2048,3072,0.00814222213294771
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,16384,10240,0.07445866531795926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,16384,16384,0.09109333488676284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,768,3584,0.012533333566453723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,10240,256,0.0174319992462794
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,6144,3584,0.03881511092185974
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,32,2048,0.006640888750553131
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1536,1536,0.008984888593355814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,6144,4096,0.05866666634877523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,65536,5120,0.0529093345006307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,16384,1536,0.029827555020650227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,512,65536,0.06200355291366577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,256,16384,0.05234933230612013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,8192,1536,0.006483555667930179
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3584,128,0.005982222242487802
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,128,6144,0.03755555550257365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,2560,512,0.013853333062595792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,12288,1536,0.026233777403831482
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,12288,2560,0.011824888487656912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,7168,16384,0.04436888959672716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,2560,10240,0.014269333746698169
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,8192,1536,0.006594666590293248
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,1536,16384,0.02089511189195845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3584,65536,0.08845333258310954
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,256,10240,0.029374221960703533
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,10240,256,0.002567111204067866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,4096,8192,0.025144888295067683
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2048,3072,0.005943111247486538
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,8192,768,0.0074177781740824384
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,64,2048,0.00869066682126787
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,6144,3072,0.007670222057236566
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,10240,4096,0.03465155429310269
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,2048,8192,0.022316444251272414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,8192,1024,0.004576888763242298
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,2560,3584,0.017179555363125272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,6144,12288,0.021741333934995864
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,4096,3584,0.009479111267460717
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,512,12288,0.11574933263990615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1024,3072,0.005663110978073544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,12288,16384,0.03970400161213345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,65536,5120,0.10986844698588054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,5120,16384,0.02383022175894843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,3072,7168,0.2032239966922336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,256,4096,0.017920000685585868
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,5120,10240,0.016534222496880423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,7168,65536,0.13061866495344374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,8192,512,0.019644444187482197
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,1536,512,0.0053137776752312975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,768,512,0.00738666703303655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,4096,3584,0.019290667441156175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,4096,10240,0.05023555623160469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,768,10240,0.012551111479600271
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,32,7168,0.00574311117331187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,10240,256,0.017658667431937326
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,768,2048,0.006769777586062749
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,65536,2048,0.03752800159984165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,8192,10240,0.026657778355810378
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3072,3072,0.0068195557428730865
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,64,64,0.005351999981535806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,768,65536,0.062260442309909396
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,768,256,0.014911111858155994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,4096,7168,0.011615110768212212
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,65536,16384,0.13076178232828775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,3584,16384,0.03606399893760681
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,32,512,0.002595555658141772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,7168,256,0.006888000087605582
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,256,10240,0.08326311243904962
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,768,10240,0.015075556106037565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,5120,3584,0.02273066673013899
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1536,12288,0.02498488956027561
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,16384,12288,0.08086578051249187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,256,2048,0.004543999830881755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,2048,768,0.00564177789621883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,32,128,0.0020355555332369274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2048,2560,0.011446221835083432
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,16384,3584,0.014348444011476306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,5120,12288,0.10576355457305908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3072,4096,0.008288000192907121
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,10240,12288,0.08063199785020617
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,16384,768,0.006934222247865465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,8192,12288,0.030866665972603693
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,65536,128,0.010009778042634329
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3584,8192,0.015993777248594496
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3072,2560,0.010418666733635796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,10240,512,0.010106667048401302
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3584,5120,0.011149333582984077
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,5120,4096,0.024637333220905725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,3072,2560,0.007000000112586551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2560,4096,0.008403555386596257
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,64,4096,0.005998222364319696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,5120,12288,0.06358400318357679
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,512,5120,0.02091377807988061
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,6144,10240,0.19364355670081246
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,1024,12288,0.014226666755146451
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,128,768,0.013810666898886362
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,65536,1536,0.033168001307381526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,768,64,0.0021608887861172357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,6144,12288,0.030025776889589097
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,2560,1536,0.006272888845867581
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,16384,6144,0.021355556117163763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,2048,16384,0.03566666775279575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,7168,512,0.01625866691271464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,5120,1024,0.0061048890153567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,768,7168,0.009159999589125315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,1024,5120,0.008785777621799046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,16384,128,0.0027173333283927706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,512,64,0.0021786666992637846
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,32,5120,0.008801778157552084
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,16384,10240,0.07740355862511529
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,16384,2560,0.012368000215954252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,65536,7168,0.127691560321384
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,4096,10240,0.014875556031862894
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2560,10240,0.015823110938072205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1024,2048,0.009355555805895064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3072,256,0.002597333242495855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,2048,6144,0.019444444113307528
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,4096,12288,0.03154666556252374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,65536,3072,0.04186222288343641
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,2048,768,0.016968000266287062
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,2560,8192,0.03851022322972616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3072,1024,0.0036346668170558084
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3072,768,0.03102133340305752
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,768,1024,0.003760888758632872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1536,128,0.009659555223253038
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,10240,768,0.025302223033375208
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,768,6144,0.006653333289755716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,64,6144,0.009362666971153682
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,2048,6144,0.00904088881280687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,2048,10240,0.017423111531469557
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,12288,3584,0.015778667396969266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,65536,128,0.02188266648186578
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,7168,3584,0.02924088968171014
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,7168,10240,0.033342222372690834
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,768,4096,0.024570667081409033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,12288,7168,0.04362044400639004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,4096,3584,0.008083555433485243
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,3584,2048,0.0069573331210348345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,768,256,0.002524444419476721
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1536,2560,0.004911111046870549
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,16384,512,0.006609777609507243
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,32,7168,0.011061333119869232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,256,256,0.012455110748608908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,768,3584,0.007737778127193451
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,3072,12288,0.01902844508488973
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1024,1024,0.005947555518812603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,128,5120,0.016199111938476562
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,6144,6144,0.013743999931547375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,4096,3072,0.02003466586271922
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,5120,512,0.021274665991465252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,8192,768,0.013391999734772576
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,2560,4096,0.017636444833543565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,3072,16384,0.023510222633679707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,64,32,0.0021928888228204516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,768,3584,0.017269333203633625
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,4096,512,0.005575110928879843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,1024,2048,0.006315555423498154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,1024,2048,0.014208000567224292
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,4096,256,0.002520888836847411
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,768,4096,0.014142221874660917
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,12288,16384,0.065702219804128
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,6144,2048,0.010517333116796283
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,512,2048,0.006163555714819167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,4096,1536,0.006320000108745363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,256,256,0.0021564445147911706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,128,65536,0.058837334314982094
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,16384,768,0.007084444165229797
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,5120,6144,0.02568622264597151
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1536,2048,0.006594666590293248
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1024,65536,0.17545599407619902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,12288,12288,0.03141155507829454
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,8192,65536,0.13746044370863172
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1024,128,0.006405333264006509
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,4096,65536,0.09793155723147923
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,3584,5120,0.02682488825586107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1024,1024,0.006622222148709827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,12288,2048,0.0343191126982371
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,8192,10240,0.030210667186313208
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,128,3072,0.01738577749994066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3072,32,0.005922666854328579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,1536,8192,0.03418755531311035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,16384,8192,0.03192977772818671
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,512,7168,0.007012444237867991
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,3072,128,0.014846222268210517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,256,768,0.014119999276267158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1536,6144,0.010218666659461128
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,512,2048,0.014468444718254937
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,64,65536,0.04931644598642985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2048,16384,0.015046222342385186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,512,7168,0.016640888320075143
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3584,12288,0.019256888164414298
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,65536,64,0.006993778049945831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,32,32,0.00629333323902554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,8192,6144,0.0417502224445343
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,128,2048,0.004604444321658877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1024,512,0.002951111023624738
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,5120,5120,0.024449777272012498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2048,12288,0.24762399991353354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,128,2048,0.00699733363257514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,12288,10240,0.058427552382151283
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,7168,2560,0.012440000143316058
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,512,5120,0.006628444625271692
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,512,512,0.01477599971824222
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,7168,5120,0.013942221800486246
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,512,768,0.006583999842405319
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,10240,256,0.009026666482289633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,7168,4096,0.055143998728858105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,7168,2048,0.009740444521109263
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,512,12288,0.013166222307417126
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,1536,65536,0.07457955678304036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,8192,10240,0.018636445204416912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,12288,10240,0.02962933315171136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,8192,5120,0.05249511202176412
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,8192,64,0.0022408889813555614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,16384,10240,0.04726666543218824
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,65536,1024,0.018951111369662814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,512,1536,0.008195555872387355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,8192,128,0.006960888703664144
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,128,6144,0.009017777939637503
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,12288,16384,0.04041422075695462
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,512,6144,0.05839822027418348
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,4096,256,0.03371289041307237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2560,256,0.0057591112951437635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,3584,1024,0.01532444523440467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,4096,65536,0.15919466813405356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,7168,12288,0.029432889488008287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,5120,2048,0.0069857777820693115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,16384,3072,0.05010399884647793
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,5120,1536,0.006295111030340195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3072,10240,0.01573422220018175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,256,768,0.0058755555914507965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,8192,2048,0.009206222163306342
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,12288,256,0.006228444476922353
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,16384,7168,0.05561511384116279
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,10240,64,0.0021582223061058256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,4096,6144,0.011049778097205691
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,768,32,0.0021679999513758556
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3584,2048,0.0069679998689227635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,128,5120,0.006272000157170826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,6144,65536,0.08007288641399808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,32,1536,0.003532444437344869
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,5120,256,0.02183644473552704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,8192,8192,0.05997688902748955
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3584,3584,0.010900444454616971
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3072,4096,0.008632000121805403
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,2560,2048,0.014816888504558139
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2560,10240,0.023222222924232483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,768,7168,0.010011555420027839
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,7168,2048,0.007889777421951294
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,5120,6144,0.10292088985443115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,10240,3584,0.36280356513129336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,3584,7168,0.05068711108631558
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,3072,3584,0.008303110798199972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,32,1536,0.00350222239891688
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,65536,6144,0.06706844435797797
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2560,10240,0.0226239992512597
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1536,10240,0.010665777656767102
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,7168,6144,0.012458666331238218
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,512,5120,0.013100444442696042
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,8192,1536,0.06076177623536852
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3584,512,0.005761777775155173
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,8192,1536,0.012125333150227865
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,65536,512,0.06494044595294528
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,768,512,0.005903111149867375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2560,512,0.005921777751710679
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,128,512,0.005594666633341048
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,1024,6144,0.020112888680564035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,5120,2560,0.09385244713889228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1536,768,0.0028773333049482773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,6144,7168,0.028205331828859117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,8192,768,0.017921777235137094
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,256,128,0.007647111184067196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,64,16384,0.005883555445406172
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,768,8192,0.010760000182522668
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,256,256,0.005559111220969095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,768,3584,0.007654222349325816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,12288,7168,0.030489779180950586
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,5120,5120,0.009397333694828881
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1024,5120,0.008991999758614434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,10240,256,0.003499555504984326
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,1536,12288,0.05091199941105313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,6144,512,0.008475555313958062
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,65536,4096,0.08741866879993015
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,10240,10240,0.039650665389166936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,3584,7168,0.03067022230890062
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,768,32,0.0022515555222829184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3584,5120,0.17048888736300996
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,1536,65536,0.06468177504009671
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,8192,5120,0.012464889221721225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,7168,65536,0.20098488860660127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,7168,2048,0.03360889024204678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,8192,8192,0.01964977714750502
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,65536,1536,0.07942755354775323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,7168,1536,0.08324978086683485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,256,256,0.006258666515350342
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,32,2048,0.004299555387761858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1024,4096,0.0069679998689227635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,32,10240,0.008061333662933773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,12288,5120,0.09888088703155518
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,8192,7168,0.01999022232161628
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,12288,7168,0.0750160018603007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,6144,16384,0.1577599975797865
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,768,10240,0.0370506677362654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,128,4096,0.007865777446164025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3072,4096,0.009221333596441481
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,1536,2560,0.0188773340649075
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1536,10240,0.01387288918097814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,65536,768,0.02016533414522807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,512,3072,0.01484266585773892
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,768,1536,0.008616888688670265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,128,4096,0.03236711025238037
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,32,3072,0.005559999909665849
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,65536,3584,0.0869235528839959
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,768,6144,0.018272888329293992
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,64,1536,0.0038035557501845886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,2048,3584,0.015819556183285184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,128,64,0.0052782222628593445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,256,256,0.017089777522616915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,768,6144,0.01789777808719211
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3584,64,0.021664889322386846
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,3072,3072,0.007133333219422235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,10240,6144,0.03157511022355821
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,12288,1024,0.006275555739800136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2560,4096,0.009921777579519484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,8192,32,0.005241777747869492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,1536,10240,0.04754844307899475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3584,1536,0.005040888984998067
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1536,3072,0.00794488853878445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,3584,3584,0.021384888225131567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,768,10240,0.011767110890812345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,512,8192,0.010657777388890585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,512,1024,0.008332444561852349
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,768,32,0.0021724444296624926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,65536,3072,0.06806488831837972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2048,3072,0.007353777686754863
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,10240,7168,0.019891555110613506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,128,8192,0.005619555711746216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,10240,8192,0.026176000634829204
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,3584,5120,0.03930577635765076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,64,2048,0.004232888834344016
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,65536,128,0.02978488802909851
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,7168,1536,0.026660444007979497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,5120,3072,0.007292444507280986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2048,4096,0.0079644446571668
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,2048,5120,0.017901332841979135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,3584,128,0.014755555325084262
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1024,16384,0.012248000337017907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,5120,8192,0.029519110918045044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,16384,64,0.00830666638082928
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,65536,5120,0.3203679985470242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,128,4096,0.007827555967701806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2048,2560,0.006989333364698622
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,256,65536,0.06173421939214071
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,768,7168,0.019294222195943195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2048,10240,0.015318221516079374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,1536,3584,0.03388799892531501
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,128,8192,0.010681777364677854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,256,65536,0.07560977670881483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,7168,128,0.015211555692884656
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,5120,8192,0.016201777590645682
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,10240,7168,0.026572444372706946
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,1024,256,0.011074666347768573
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,5120,2560,0.011053333679835001
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,10240,2560,0.02700888944996728
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,2048,8192,0.03887288769086202
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,768,1024,0.0038844446341196695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,7168,6144,0.014486221803559197
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,1024,12288,0.02574133376280467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1024,64,0.0021768889079491296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,1536,256,0.020264888803164165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,256,1024,0.0062871111763848206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,32,2560,0.004576000074545543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,16384,10240,0.05784889062245687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,16384,2560,0.018765333626005385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,65536,1536,0.09095199902852376
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,128,1024,0.006287999865081575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,512,64,0.0058462222417195635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1536,8192,0.011044444309340583
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3584,1024,0.005873777800136142
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,6144,768,0.006243555496136348
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,1536,3072,0.007280000381999546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,10240,512,0.0041884444653987885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,4096,4096,0.010408889088365767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,12288,2560,0.014070222775141398
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,65536,64,0.016227554943826463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,32,5120,0.004867555366622077
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,768,64,0.005577777822812398
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,12288,12288,0.031545778115590416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,10240,768,0.019222221440739103
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,3072,3584,0.01109688894616233
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,7168,256,0.016862221890025668
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,16384,8192,0.046815110577477344
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1536,128,0.002291555619902081
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,16384,6144,0.12685333357916936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,2560,4096,0.00795377790927887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,16384,8192,0.04831466741032071
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,3584,512,0.005559111220969095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,3072,3584,0.02700533303949568
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1024,512,0.00592533333433999
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,8192,256,0.006950221955776215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,12288,8192,0.5043324364556206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,3584,12288,0.04247200157907274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,128,32,0.005935110979610019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,128,3584,0.007650666766696506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,512,8192,0.005607111172543631
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,8192,10240,0.05674488676918877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,6144,768,0.017804443836212158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,5120,4096,0.011672889192899069
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,8192,128,0.015526221858130561
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2048,128,0.0031946665710873077
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,768,4096,0.018795554836591084
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3584,4096,0.04319466816054451
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,65536,2560,0.03157866663402981
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,256,128,0.005371555685997009
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,16384,6144,0.020255999432669747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,65536,1024,0.15004177888234457
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,256,65536,0.021627555290857952
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,128,16384,0.015814221567577787
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,32,12288,0.013082666529549493
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,6144,7168,0.012821333275900947
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,2048,16384,0.01992177797688378
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,1536,3072,0.02236000034544203
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,512,6144,0.010416888528399997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,64,768,0.002817777709828483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,2560,256,0.014872888724009195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,3072,512,0.015232000086042615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,10240,256,0.017239999439981248
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,3584,8192,0.012632889052232107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2560,7168,0.012219555675983429
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,256,2048,0.006249777972698212
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,12288,3072,0.027500443988376196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,32,32,0.005614222337802251
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3072,256,0.0042444442709287005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,256,1536,0.006524444454246097
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,1536,65536,0.06891466511620416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3584,1536,0.005991999887757831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2560,1536,0.022614222433831956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,65536,512,0.05722577704323662
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,10240,16384,0.05012089014053345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,10240,8192,0.021890666749742296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,10240,512,0.018923555811246235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1536,1024,0.003165333428316646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,256,1536,0.009672000176376766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,1536,12288,0.013784888717863293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,7168,65536,0.22210489379035103
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,8192,65536,0.09110133515463935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,12288,3584,0.035049776236216225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,128,3072,0.013982221484184265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,16384,128,0.0064239998658498125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,3072,4096,0.020939555433061387
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,5120,10240,0.020729778541458976
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,512,3584,0.0053075556125905775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,7168,1024,0.008008889026112026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,4096,2048,0.005980444451173146
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2560,2048,0.0063075555695427795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,128,1536,0.004679111142953237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,10240,4096,0.056833777162763804
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1024,10240,0.013211555778980255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1024,32,0.005584888988071018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3072,768,0.004211555752489301
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3584,3584,0.007874666816658443
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,6144,512,0.00630666646692488
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,10240,32,0.011115555961926779
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,2560,512,0.005362666729423735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,512,16384,0.008038221961922115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,12288,2560,0.029818667305840388
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3584,65536,0.11118843820359971
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,4096,128,0.02602311140961117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3072,128,0.00702311098575592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,10240,16384,0.08606221940782334
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,4096,65536,0.13784710566202799
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,10240,1536,0.015084443820847405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,768,5120,0.0210826661851671
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2048,16384,0.019111999207072787
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,6144,1024,0.05537955628501045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,8192,16384,0.04495733314090305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2560,12288,0.017328888177871704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,16384,1536,0.010835555692513784
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,256,3072,0.007357333269384172
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,12288,1024,0.0580657786793179
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,65536,1536,0.021807110971874658
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,10240,6144,0.01763199932045407
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,7168,5120,0.02782755427890354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,5120,3072,0.02256533337963952
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,8192,12288,0.06635822190178765
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2560,128,0.002179555594921112
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,7168,4096,0.0127697777416971
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1536,1536,0.006308444258239534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1536,1536,0.006659555352396435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,5120,1536,0.011742221812407175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1024,2048,0.0063573333124319715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1024,5120,0.0063582224150498705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,128,2048,0.007979555262459649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,3584,7168,0.02381244467364417
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2048,3072,0.01038933296998342
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,256,5120,0.005664888769388199
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,128,8192,0.005304000029961268
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,512,6144,0.005642666584915585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,65536,64,0.004247111164861255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,5120,10240,0.04833066794607374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,2048,6144,0.023269333773189124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,65536,12288,9.788475884331598
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1536,65536,0.10011911392211914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1024,2560,0.013444444371594323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,6144,3072,0.01584533353646596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,16384,3072,0.0747528870900472
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,7168,768,0.006637333167923822
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,7168,6144,0.03901866740650601
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3584,12288,0.0270115551021364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3584,5120,0.011795555551846823
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,128,7168,0.011094222466150919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,6144,768,0.003760888758632872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,8192,3072,0.012785777449607849
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,7168,6144,0.018744000130229525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,8192,512,0.00943999985853831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,12288,65536,0.3094240029652913
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,4096,5120,0.02460533380508423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,8192,2048,0.006966222491529252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1024,32,0.002199999988079071
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2048,3072,0.007363555332024892
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,5120,512,0.0029484445436133277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,8192,3584,0.010103999740547603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,512,6144,0.005963555640644497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,6144,1536,0.007148444652557373
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,5120,128,0.005620444400442972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,16384,2048,0.011902222202883827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1536,128,0.0022533333135975734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,10240,7168,0.1885662211312188
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,16384,1024,0.008604444563388824
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,3072,12288,0.036715557177861534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,4096,1536,0.004843555390834808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,6144,128,0.0022275555464956495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,4096,768,0.006631111105283101
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,3584,12288,0.05408089028464424
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3584,4096,0.012149333126015134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1536,6144,0.018509333332379658
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1536,3072,0.007321777443091075
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,768,6144,0.0210515558719635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,768,2560,0.006981333096822103
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,64,6144,0.00665866666369968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,256,1536,0.015467555986510383
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,256,256,0.005952000204059813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,256,512,0.002534222271707323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1024,1536,0.004560889055331548
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2048,1024,0.0042204442951414315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,10240,16384,0.0856328871515062
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,32,2048,0.005511999958091312
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2048,512,0.005559111220969095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,16384,3072,0.09494844410154556
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,4096,3584,0.0185208883550432
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,8192,7168,0.017143110434214275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,4096,256,0.01585422290696038
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,65536,512,0.013096888860066732
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,512,8192,0.01997599999109904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,512,1024,0.006248000181383557
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,4096,128,0.006243555496136348
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,2048,1024,0.005933333188295364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,10240,128,0.03254933489693536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3072,8192,0.012124444047609964
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,7168,5120,0.01312800000111262
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,3584,4096,0.02938488788074917
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,4096,16384,0.021136000752449036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,8192,2560,0.010727110836240979
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,4096,2048,0.005657777604129579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,5120,256,0.006616000086069107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,8192,128,0.015678221980730694
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,5120,128,0.014867555763986377
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2560,768,0.003573333223660787
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,16384,128,0.009367110828558603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,16384,6144,0.02381333377626207
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,1536,256,0.01515111161602868
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,10240,2560,0.01015377789735794
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,7168,4096,0.20228444205390084
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,7168,3072,0.016301333904266357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3584,16384,0.21799643834431967
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,4096,8192,0.013912889692518445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1024,16384,0.01889155473974016
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,3072,12288,0.03772355450524224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,65536,10240,2.3902053833007812
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,768,3072,0.011131555669837527
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,16384,512,0.01038933296998342
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,5120,6144,0.025272000167104933
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,768,2048,0.00692622239391009
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,768,4096,0.008338666624493068
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,65536,1024,0.027561777167850073
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3584,3072,0.007654222349325816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,64,4096,0.008368000388145447
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,12288,128,0.015784889459609985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1536,128,0.00590844452381134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1536,16384,0.24534755282931855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,512,1024,0.030192000998391047
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,128,3072,0.011224889092975192
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,128,256,0.011739555332395764
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,2048,512,0.022682666778564453
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,10240,256,0.016207999653286405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,6144,12288,0.02756799923049079
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1536,8192,0.021349334054523047
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,2560,2560,0.019242667489581637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,256,32,0.0029351111087534162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,65536,3584,0.18038754993014863
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,5120,5120,0.11007644732793172
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,10240,128,0.00222044438123703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,1024,5120,0.017613333132531907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,64,64,0.0022168887986077205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,7168,7168,0.04955377843644884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2560,64,0.0028124445428450904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,12288,2560,0.012831110921170978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,5120,128,0.01517511076397366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2048,10240,0.02260622216595544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,512,65536,0.03278222348954942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,2560,10240,0.026408000124825373
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,7168,128,0.005639111002286275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,3072,3584,0.025723555021815833
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,1024,7168,0.02751111156410641
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,16384,512,0.022843556271659002
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,128,3072,0.007330666813585494
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,256,512,0.0029502221279674103
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,1536,1536,0.006268444574541516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,128,768,0.0028568889117903183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3584,1536,0.006577777779764599
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,65536,512,0.05745333433151245
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,768,5120,0.006637333167923822
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,2560,4096,0.021253334151373968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,256,256,0.013739555246300168
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,768,10240,0.0269777774810791
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,256,128,0.012453333371215396
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3584,10240,0.018030222919252183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,3584,1536,0.01962844365172916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,5120,1024,0.05470400055249532
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,512,65536,0.08606577581829494
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,4096,3072,0.0076453329788313965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,65536,2560,0.0846231116188897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,768,6144,0.02014400064945221
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,16384,5120,0.020464888877338834
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,10240,512,0.006617777877383762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,256,256,0.0022622222701708474
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,5120,7168,0.013168888787428537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3072,4096,0.009571555587980483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,6144,65536,0.07656800084643893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1536,32,0.0021688888470331826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,16384,256,0.00656444455186526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3072,10240,0.014017777310477363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,10240,3072,0.020780444145202637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,256,2048,0.01346933344999949
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,4096,7168,0.1298124392827352
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,7168,1024,0.014196443888876172
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,5120,10240,0.018366222580273945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,7168,7168,0.020709334148301017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2048,512,0.0035662220584021676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,2560,128,0.014857777290874057
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,32,32,0.005598222215970357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2048,2048,0.008627555436558193
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,12288,3584,0.03440444336997138
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,12288,1024,0.007409777906205919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,512,3584,0.00795911086930169
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,256,65536,0.07047555843989055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,256,768,0.007606222397751278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,65536,6144,1.8525928921169703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,7168,2560,0.03736977775891622
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,12288,5120,0.03560622202025519
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,65536,768,0.030618667602539062
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,2048,1536,0.006229333165619109
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,64,3072,0.004864888886610667
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3584,3584,0.008285333712895712
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,128,512,0.0059368887709246735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,512,8192,0.04576622115241157
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,65536,64,0.0041724443435668945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,128,10240,0.01994311147265964
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,512,768,0.0028684443483750024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,6144,768,0.016862221890025668
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1024,768,0.006229333165619109
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,12288,5120,0.022985777921146814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,8192,256,0.01684177749686771
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,64,10240,0.02532711128393809
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,65536,5120,0.11028711001078288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,32,4096,0.00828888929552502
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,12288,6144,0.04310755597220527
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3584,32,0.005601777798599667
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3072,3584,0.009611555271678502
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,2048,256,0.025732444392310247
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,4096,768,0.016186666157510545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3584,768,0.007113777928882175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,10240,65536,0.15522310468885633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,32,1536,0.00692622239391009
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,2560,2560,0.01535022258758545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,65536,768,0.0144195556640625
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,6144,12288,0.06708533234066434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,32,4096,0.007976888782448238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,5120,6144,0.015103111664454142
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,16384,10240,0.09308355384402806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,8192,16384,0.04584533307287428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,8192,65536,0.44300177362230087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,7168,1536,0.007227555745177799
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,8192,3072,0.08489866389168634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,2048,65536,0.07852800024880303
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,512,12288,0.03163733416133457
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2560,10240,0.012416889270146688
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,768,512,0.005928888916969299
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,1536,4096,0.007991111112965478
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,256,3584,0.016171556380059984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,16384,65536,0.32240088780721027
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,8192,512,0.003400000019205941
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,10240,16384,0.07679377661810981
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,16384,128,0.010336000058386061
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,512,1024,0.02105866703722212
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,768,768,0.0031902222997612427
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,4096,1024,0.01724088854259915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,10240,3072,0.03214577833811442
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,8192,7168,0.030738668309317693
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,12288,65536,0.24695110321044922
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1024,1024,0.0035031110876136353
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,12288,32,0.0026977778308921387
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,65536,64,0.0041884444653987885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,128,16384,0.027824888626734417
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,8192,2048,0.014848000473446317
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,10240,1024,0.09007466501659817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,32,10240,0.023387556274731953
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,768,768,0.0033528889632887314
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,10240,7168,0.022808000445365906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,7168,3584,0.010654222634103564
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,3584,16384,0.13590221934848362
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,7168,4096,0.012797333300113678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,4096,2048,0.025359110699759588
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,1536,12288,0.019704000817404855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2048,64,0.00645955569214291
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,256,256,0.006103111224042044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,64,2560,0.007191999918884701
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,6144,8192,0.5380746523539225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,10240,6144,0.01757422255145179
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,768,5120,0.02493688960870107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1536,12288,0.01683822274208069
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,16384,5120,0.2333137724134657
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,4096,3072,0.007664889097213745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,16384,4096,0.029434667693244085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,12288,4096,0.01865600049495697
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,6144,12288,0.054690665668911405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,256,10240,0.021951999929216173
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1536,32,0.005274666680230035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,7168,65536,0.2434799936082628
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1536,4096,0.007284444239404466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,7168,10240,0.03503288825352987
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,256,7168,0.11352445019616021
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,256,6144,0.05043288734224108
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,128,6144,0.009297777381208207
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,256,768,0.014158222410413953
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,7168,65536,0.528946664598253
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2048,5120,0.028916445043351915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,7168,16384,0.02473066747188568
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,64,6144,0.009308444129096137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,768,1024,0.00591822216908137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1536,1024,0.0034871109657817413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,7168,4096,0.026102221674389307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,2048,512,0.005560889012283749
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,512,1024,0.014480888843536377
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,128,3072,0.018415111634466384
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,1536,3072,0.015840889679061044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,5120,256,0.006680888848172293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,768,512,0.005345777918895085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,16384,2048,0.027577777703603108
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,256,3072,0.007275555696752336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1536,768,0.021839110387696162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,3072,4096,0.02780533168050978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,16384,3584,0.019740444090631273
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,768,128,0.005335999859703912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,256,1536,0.0038764443662431505
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,12288,65536,0.5121297836303711
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2048,768,0.01109866632355584
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1024,1536,0.012166221936543783
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,3072,65536,0.48212623596191406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,3584,7168,0.03760800096723769
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,7168,8192,0.5438435342576768
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,8192,2048,0.00831911133395301
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,128,768,0.002887111157178879
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,12288,3584,0.054452445771959096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1024,4096,0.014440000057220459
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,1024,768,0.0326275560590956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,16384,8192,0.025881777207056682
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,65536,768,0.022181333767043218
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,256,768,0.014463999205165438
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,16384,7168,0.0333555539449056
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,3584,7168,0.05155644483036465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,32,1024,0.009170666337013245
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,5120,3584,0.010355555348926121
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,6144,5120,0.010102222363154093
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,128,7168,0.01755822201569875
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,12288,2048,0.029158221350775823
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,8192,8192,0.021384000778198242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,7168,3072,0.010178666975763109
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3584,32,0.0022337778160969415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2048,10240,0.01679377754529317
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1024,2560,0.004881777697139316
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,5120,3072,0.02481511069668664
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,16384,256,0.006329777754015393
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,5120,1536,0.007939555578761632
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2048,10240,0.016727111405796476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,10240,65536,2.6799102359347873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,65536,16384,0.13425866762797037
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,12288,8192,0.033681776788499616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,6144,16384,0.8432728979322645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,16384,768,0.030972444348865088
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,512,32,0.0074133334888352295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,768,4096,0.021706667211320665
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,10240,4096,0.2632986704508464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,3584,8192,0.03368533319897122
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,32,8192,0.005581333405441708
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2048,32,0.005568000177542369
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3072,32,0.002056888822052214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,2048,128,0.02029333346419864
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,768,2048,0.016474665866957772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,512,3584,0.007418666448858049
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,3584,65536,0.215719117058648
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3584,12288,0.3931253221299913
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,64,6144,0.009530666801664565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,1024,4096,0.007684444387753804
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,12288,5120,0.0806479983859592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,5120,32,0.029521776570214167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,5120,128,0.002165333264403873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,512,64,0.00666844430896971
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,16384,2560,0.016075554821226332
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,768,12288,0.01499022212293413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,6144,64,0.005679111099905438
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2560,3072,0.009110222260157267
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,16384,128,0.020205333828926086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1536,16384,0.013827555709415011
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1536,8192,0.01108977778090371
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,7168,3072,0.029096888171301946
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,7168,7168,0.10182844267951117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,128,1536,0.003903999924659729
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,3072,6144,0.028558221128251817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1024,768,0.0038933331767717996
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,4096,16384,0.1155431138144599
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,512,8192,0.034330666065216064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3584,10240,0.017864889568752713
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,256,16384,0.05215733249982198
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,12288,65536,0.2235999902089437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,8192,5120,0.026761778526835974
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,7168,1536,0.01963200006220076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,7168,3072,0.0237137774626414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,512,65536,0.29067378573947483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,65536,3584,0.040523555543687605
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,8192,7168,0.016851555969980028
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2560,8192,0.01239999963177575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,5120,1024,0.0039057777159743835
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,1536,8192,0.02027466727627648
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,2560,1536,0.014899555179807874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,128,10240,0.00924711094962226
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,512,10240,0.012443555725945367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2048,32,0.005720888988839255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,2048,3584,0.02383822202682495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3072,16384,0.01847822301917606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,3072,10240,0.03664355476697286
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,32,16384,0.015506666567590503
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,32,6144,0.005236444373925527
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,12288,16384,0.041984889242384166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,1536,1024,0.016200888488027785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,512,1536,0.014871110518773397
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,6144,128,0.009230222139093611
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,2048,256,0.00519822206762102
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,2560,4096,0.024238222175174292
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,64,10240,0.012727110750145383
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,3584,1536,0.015508444772826301
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,4096,1536,0.006537777682145436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,256,7168,0.011327111058764987
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,16384,1024,0.008410666551854875
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,1536,16384,0.02095288866096073
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,8192,2560,0.007678222325113084
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,16384,1536,0.009830222361617619
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,1024,768,0.01716977854569753
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,8192,2560,0.010097777677906884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,32,32,0.0018808889306253856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,3072,4096,0.027793778313530818
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,6144,256,0.0053342220683892565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,512,5120,0.017616000440385606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,1536,768,0.017248888810475666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,128,2048,0.013229333692126803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,12288,768,0.008715555899673039
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,768,3584,0.016209777858522203
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,3584,3584,0.007633777956167857
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,512,3072,0.0052684446175893145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,10240,3584,0.23348267873128256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,64,3584,0.004932444542646408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2048,8192,0.014499555031458536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,10240,65536,5.93356196085612
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,256,3072,0.016507556041081745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,768,256,0.007990222010347579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,768,12288,0.013218666944238873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,65536,128,0.036200000180138484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,1024,4096,0.016543999314308167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1536,1536,0.0051893335249688895
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,16384,3072,0.01366311146153344
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,10240,512,0.01309866706530253
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,6144,3584,0.01757333344883389
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,4096,256,0.005569777968857024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,4096,256,0.005671999934646819
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,8192,5120,0.016410667035314772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,10240,10240,0.02823466724819607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,512,1024,0.009685333404276106
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,3072,12288,0.05709422296947903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1536,16384,0.03450399968359206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,16384,16384,0.04512622290187412
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,256,6144,0.010402667025725046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,12288,7168,0.04773422082265218
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,6144,256,0.008150222400824228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,128,1536,0.006236444330877728
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,12288,65536,0.16492622428470188
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2560,10240,0.012431110772821637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,1024,3584,0.016209777858522203
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,10240,4096,0.017644445101420086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,2560,16384,0.03438399897681342
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,12288,4096,0.044031110074785024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,6144,512,0.01629066632853614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,128,3072,0.007308444215191736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,5120,16384,0.046302222543292575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,8192,512,0.0032275555034478507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,12288,16384,0.05974577532874214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2048,2560,0.010277333358923594
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,256,10240,0.011765333513418833
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,65536,2048,0.044460445642471313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,768,5120,0.01587199999226464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,12288,7168,0.04552888870239258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,8192,1024,0.006631111105283101
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2560,2048,0.008989333278603023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,512,128,0.005280888742870755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,2560,4096,0.01864177816443973
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,1536,12288,0.01314755529165268
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,1536,7168,0.024784889486100938
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,32,8192,0.010419555836253695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,2560,512,0.005604444278611078
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,1024,4096,0.03100888927777608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,2560,768,0.038791110118230186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3584,128,0.008019555773999956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,768,512,0.012871111432711283
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1536,512,0.0052791109515561005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2560,512,0.011138666835096149
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,512,1536,0.013478221992651621
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,16384,2560,0.01850222216712104
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,512,4096,0.009754666851626502
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,12288,10240,0.04338488976160685
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,16384,5120,0.04038044479158189
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1536,32,0.006745777610275481
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,6144,768,0.0052684446175893145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,32,4096,0.004613333278232151
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,4096,7168,0.013117333253224691
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,65536,16384,5.182738833957248
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1024,2560,0.02033955521053738
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1536,8192,0.010900444454616971
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,5120,2560,0.024569777978791132
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,12288,2048,0.009482666850090027
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,1024,12288,0.043640888399548
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,1536,5120,0.008360000120268928
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,65536,10240,0.1947937806447347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,12288,1536,0.009780444204807281
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,768,512,0.002560888934466574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,3072,512,0.017664889494578045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,6144,10240,0.04497155547142029
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,65536,3584,0.03579466541608175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1536,256,0.002475555572244856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,6144,10240,0.19688355922698975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,12288,16384,0.272328880098131
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,8192,1536,0.006683555742104848
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,12288,1536,0.011579555769761404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,4096,128,0.0052382221652401825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,4096,4096,0.02342844506104787
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,1536,5120,0.008422222402360704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,5120,12288,0.01793955597612593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2048,2048,0.004233777936961916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,12288,8192,0.08144711123572455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,10240,12288,0.08677066697014703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,7168,3072,0.029046220911873713
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,512,10240,0.03530222177505493
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,128,12288,0.013215999636385175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2560,7168,0.04638222191068861
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,8192,1024,0.006334222025341458
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,10240,1536,0.008239111138714684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,4096,2560,0.016194666425387066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,256,10240,0.005640000104904175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,8192,3584,0.010761777559916178
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1536,4096,0.008009778128729926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,65536,3584,0.09185244639714558
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,10240,2560,0.011500444677140979
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,10240,1536,0.033619556162092425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3072,768,0.005945777727497949
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,12288,7168,0.027473777532577515
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,8192,1536,0.006955555743641323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,12288,65536,0.19970756106906465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,3072,128,0.017879999346203275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,4096,65536,0.08775999810960557
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,3584,1536,0.019761777586407132
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,2560,65536,0.0648604432741801
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,64,768,0.007099555598364935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,5120,3072,0.008078222473462423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,3584,512,0.005932444499598608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,6144,6144,0.044453332821528115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1536,3072,0.011733333269755045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,3584,3072,0.10454044739405315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,6144,4096,0.041859553919898145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,4096,512,0.014783110883500842
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3584,7168,0.12193866570790608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,7168,512,0.005943111247486538
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,1024,10240,0.011821333732869891
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,3072,256,0.01531733406914605
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,128,2048,0.005058666484223472
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2048,512,0.005984888722499211
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,128,8192,0.012415111064910889
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,3072,256,0.005388444496525659
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,12288,4096,0.0170471105310652
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,4096,1536,0.00962488849957784
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,64,768,0.002970666728085942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,128,768,0.005677333308590784
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,2048,3584,0.024496888120969135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,256,4096,0.018068444397714403
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,2560,4096,0.017443555924627516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,512,12288,0.017216889394654166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,512,512,0.0028444443725877335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,10240,12288,0.04960888955328199
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,16384,1024,0.009742221898502773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,256,5120,0.008987555901209513
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,32,4096,0.004583999928500918
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3584,12288,0.019913777709007263
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,32,6144,0.009323555562231276
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,1536,8192,0.02774577836195628
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,7168,3072,0.024365334047211543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,512,2048,0.008671110702885522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,32,7168,0.010109333528412713
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,65536,768,0.015603555573357476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,2560,3072,0.01147555559873581
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,7168,128,0.005600888695981767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,6144,8192,0.01688088807794783
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,2048,3584,0.007532444265153673
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,512,6144,0.009683555199040307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,4096,3584,0.023461333579487268
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,4096,16384,0.021866665946112737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,16384,3072,0.025052444802390203
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,32,768,0.0029013332807355454
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,65536,5120,0.05755822526084053
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,256,1536,0.006313777632183499
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,4096,1024,0.015501333607567681
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,4096,10240,0.4003564516703288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1024,1024,0.003920888735188378
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2560,512,0.005631111148330901
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,16384,65536,0.2998640007442898
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3584,2560,0.006698666761318843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,5120,2048,0.010474666953086853
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,5120,128,0.016511999898486666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,4096,5120,0.009369778136412302
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,6144,3584,0.028785778416527644
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3072,1024,0.005984888722499211
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,8192,6144,0.0333315564526452
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,256,5120,0.008977777428097194
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,128,65536,0.05791999896367391
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,1536,512,0.005550222264395819
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,256,12288,0.024403555525673762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,6144,1024,0.00599911105301645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,8192,2048,0.03689422210057577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,6144,64,0.0052897776994440295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,12288,256,0.0063697778516345555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,2048,3584,0.055467555920283
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,10240,8192,0.030332446098327637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,32,12288,0.10053066412607829
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,65536,2560,0.02832711074087355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,3072,6144,0.022436444958051045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,10240,4096,0.013732444081041547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,12288,512,0.00905600024594201
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,16384,10240,0.0664657751719157
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,16384,1024,0.007118222614129384
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,768,16384,0.0355697770913442
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,768,10240,0.011824000212881299
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,128,768,0.002874666617976295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2560,512,0.0029235554652081597
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,128,5120,0.005695111221737332
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,768,12288,0.03734222385618422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,6144,256,0.0022622222701708474
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,2048,1024,0.013777777552604675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,2048,5120,0.017148445049921673
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,12288,128,0.008347555167145198
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,3072,16384,0.0325546662012736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1536,768,0.006305777778228124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,256,4096,0.015039111177126566
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,3584,4096,0.02754666739039951
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,5120,1024,0.005650666852792104
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,32,3072,0.014829332629839579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,65536,5120,0.21207822693718803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,7168,8192,0.04886933167775472
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,256,16384,0.018546667363908555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,5120,3072,0.00980799976322386
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,512,2048,0.008986666798591614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,3584,4096,0.027775999572541978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3072,2048,0.007701333198282454
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,768,7168,0.00960533320903778
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,65536,64,0.048376000589794584
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1536,1536,0.0053093334039052325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,8192,65536,3.480127122667101
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,1024,768,0.005616000129116907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1536,768,0.006020444548792309
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,1024,16384,0.016598222984208
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1536,1024,0.008019555773999956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,8192,1536,0.007304000357786815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1024,2560,0.009448889229032729
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,12288,7168,0.04889422323968676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3584,2560,0.008552889029184977
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,16384,65536,0.17257244057125518
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,2048,3072,0.007696888513035244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,5120,16384,0.041350222296184964
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,2560,1024,0.014560888210932413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,6144,7168,0.031018667750888403
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,65536,2560,0.14169243971506754
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,3584,6144,0.036017778846952654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,768,5120,0.016586666305859882
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,1536,1536,0.006314666734801398
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3584,128,0.005947555518812603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,65536,2560,0.044567998912599355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1536,12288,0.013520888984203339
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,4096,10240,0.05072089036305746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,256,10240,0.024030221833123103
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,5120,1024,0.006711111300521427
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,3072,5120,0.012126222252845764
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,4096,12288,0.017318222257826064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,6144,4096,0.018917333748605516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,2560,65536,0.06510133213467069
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,4096,10240,0.0345164438088735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,8192,1024,0.006249777972698212
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,4096,6144,0.03770666652255588
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,16384,6144,0.23202133178710938
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,128,65536,0.05741777684953478
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,768,3584,0.006260444306664997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,6144,5120,0.03240088952912225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,1536,6144,0.010699555277824402
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,7168,16384,0.023754666248957317
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,8192,256,0.03750844465361701
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,12288,5120,0.03167555398411221
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,16384,1024,0.008728000024954477
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,2048,6144,0.02032177812523312
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,6144,16384,0.04316799839337667
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,3584,5120,0.053054223457972206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,5120,6144,0.018984888990720112
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,3584,6144,0.01312800000111262
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,8192,10240,0.030423998832702637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2048,32,0.0021688888470331826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,4096,3584,0.008259555531872643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,16384,6144,0.04119644562403361
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,7168,256,0.0028488888508743713
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,2560,2048,0.0092275556590822
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,1536,5120,0.04480266571044922
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,768,64,0.005965333431959152
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1536,5120,0.010423111418883005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,5120,4096,0.011376000112957425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,10240,3584,0.12015022171868218
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,768,5120,0.008629333641793992
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2048,2560,0.018175111876593698
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,512,7168,0.019191111127535503
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,12288,8192,0.28103555573357475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,5120,6144,0.014711111783981323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,7168,64,0.005254222287072076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2560,65536,0.174566215938992
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,3584,4096,0.024996444582939148
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,1024,512,0.005951111101441913
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,1536,16384,0.019189332922299702
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,16384,16384,0.050534221861097545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,128,6144,0.004892444445027245
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,768,5120,0.007631999750932057
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,16384,1024,0.007984889050324758
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,3584,3584,0.00740444411834081
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1024,7168,0.01037599974208408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,512,5120,0.008738666772842407
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,256,2560,0.007280000381999546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,12288,7168,0.020439111524158053
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,3584,8192,0.01553422212600708
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1536,16384,0.01924088928434584
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,12288,768,0.09350666734907363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2560,3584,0.007068444457319047
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,16384,3072,0.014112888111008538
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,2560,768,0.005576888720194499
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,512,256,0.002157333410448498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,2048,10240,0.05085244443681505
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,3584,10240,0.09017778105205959
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,16384,512,0.006887110984987683
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,2048,16384,0.030108445220523413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,16384,512,0.00700444479783376
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,7168,256,0.006934222247865465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,512,5120,0.005261333452330695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,12288,6144,0.03635911146799723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,10240,12288,0.0948444472418891
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1536,128,0.0069448889957533936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,10240,1024,0.01880088945229848
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,64,4096,0.004534222185611725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,64,5120,0.014139556222491793
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,16384,7168,0.025222222010294598
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,512,8192,0.015503999259736804
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1024,768,0.003580444388919406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,16384,64,0.006700444552633498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,10240,128,0.010813333094120026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,256,128,0.0063528890411059065
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,65536,12288,0.1300719976425171
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,7168,1536,0.007308444215191736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1024,128,0.0028764444092909494
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,8192,4096,0.011028444601429833
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,1024,7168,0.009677333136399588
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3072,1024,0.005020444591840108
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,768,65536,0.062053336037529834
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,8192,256,0.01719199948840671
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,768,8192,0.011516444385051727
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,12288,8192,0.026564444104830425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,512,5120,0.005640888793600931
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,2560,10240,0.05879111422432793
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,65536,12288,0.18270933628082275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,6144,768,0.01829866733815935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,768,32,0.006409777949253718
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,10240,3584,0.029905776182810467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,128,12288,0.03987110985649957
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,256,1024,0.003346666693687439
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,7168,32,0.005625777774386936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,5120,128,0.00371111101574368
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,5120,12288,0.03538844320509169
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,8192,65536,6.645869361029731
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,64,128,0.003984000119898054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,768,8192,0.0129457778400845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,8192,1536,0.019313777486483257
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,4096,32,0.0026168889469570587
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3584,5120,0.016121778223249648
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,256,5120,0.08452622095743816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1024,8192,0.020655999581019085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,5120,7168,0.027092445227834914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,256,6144,0.005295111073387994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3072,6144,0.02076088885466258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,8192,2560,0.010187555518415239
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,6144,64,0.005585777676767773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,5120,768,0.005678222411208683
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1024,2048,0.006616888774765863
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,512,768,0.005921777751710679
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,5120,1536,0.022711111439598933
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,256,3584,0.005952000204059813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,7168,1536,0.024115555816226538
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,7168,64,0.005680888891220093
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,3072,8192,0.042669332689709134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,64,3584,0.004964444372389051
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,5120,32,0.006411555740568373
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,4096,768,0.01753155555990007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,256,1024,0.006026666611433029
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,64,5120,0.005655999812814925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,32,6144,0.015551111764378019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,3072,6144,0.021734222769737244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,768,4096,0.008040889269775814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3072,1024,0.003491555651028951
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,128,12288,0.014138667119873894
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,1024,3072,0.016653334101041157
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,16384,4096,0.04117155406210158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,12288,3072,0.03830755419201321
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,128,3584,0.004967999955018361
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,7168,65536,0.13758488496144614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,128,7168,0.01755911111831665
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,65536,768,0.0173333336909612
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,7168,2048,0.025789333714379206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,1024,7168,0.026659554905361597
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,4096,128,0.00527555536892679
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,5120,32,0.007034666836261749
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,64,2048,0.004957333207130432
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,65536,2560,0.07229333453708224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,8192,5120,0.012051555845472546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,3584,1536,0.018423111902342904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,4096,256,0.00259911103381051
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,6144,12288,0.019496000475353666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,5120,4096,0.023609777291615803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,4096,768,0.017902221944597032
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,3072,6144,0.009031111167536842
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,6144,32,0.007696000238259633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,2560,1024,0.007375111182530721
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2560,5120,0.008058666355080074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,7168,3072,0.023875556058353845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,6144,6144,0.02036088870631324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2048,6144,0.008850666383902231
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,4096,4096,0.009545777406957414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,2560,2048,0.006607999818192587
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,1536,6144,0.026603556341595117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1024,16384,0.018561777141359117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,12288,5120,0.022427555587556627
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,12288,2048,0.012854222622182635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,128,512,0.002568888788421949
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3584,16384,0.025311110748185053
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,2048,1024,0.006673777682913675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,256,3584,0.007842666572994655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,512,2560,0.009384888741705153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,512,32,0.0028320000403457214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,7168,10240,0.02945599953333537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2048,7168,0.01109777804878023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,7168,8192,0.01868799991077847
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,4096,2048,0.01851111153761546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,5120,16384,0.030786666605207656
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2048,10240,0.010120000276300643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,6144,6144,0.12627377774980333
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,128,65536,0.010310221877362993
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,16384,128,0.016537777251667447
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,65536,12288,0.1873422198825412
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,1024,6144,0.00905866672595342
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,10240,1536,0.02409511142306858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,64,512,0.00254488881263468
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1536,16384,0.026925333672099646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,768,2048,0.011782222323947482
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,8192,1024,0.006350222147173352
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,16384,256,0.006274666637182236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,8192,1536,0.006986666884687211
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,4096,6144,0.010388444695207808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,128,65536,0.07948799928029378
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,8192,3072,0.01830577850341797
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1024,512,0.0090524446633127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,128,8192,0.010454222559928894
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3584,3584,0.021595555875036452
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2048,65536,0.06247644291983711
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,1024,128,0.006281777802440856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,6144,12288,0.021847999758190576
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2560,6144,0.012129777835475074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3584,10240,0.017717333303557504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,5120,5120,0.011811555259757571
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,7168,2048,0.019261333677503798
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1536,512,0.013032000097963544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,32,2560,0.010144888526863521
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,16384,10240,0.05279111199908786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2048,7168,0.01274222218328052
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,256,10240,0.01483022173245748
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,12288,4096,0.03672711054484049
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,12288,1536,0.02139733400609758
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,32,1536,0.003690666622585721
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,8192,4096,0.2103795607884725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,12288,6144,0.025829333398077223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,2560,512,0.00638488887084855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,512,128,0.002165333264403873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,10240,2560,0.088664001888699
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,8192,768,0.004127999974621667
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,65536,128,0.00490044429898262
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2560,3072,0.00654666663871871
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,1536,3584,0.016338666280110676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,4096,512,0.002700444517864121
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2560,3072,0.007631999750932057
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,8192,10240,0.04412800073623657
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,256,3584,0.005935110979610019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,512,512,0.01240711079703437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,768,3072,0.02940088841650221
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,3584,3072,0.016837333639462788
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2048,512,0.011744889120260874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,6144,4096,0.008302222523424361
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,2560,128,0.005605333381228977
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,65536,4096,0.08548355764812893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,5120,512,0.01685066686736213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,12288,3072,0.013040888640615674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,512,10240,0.09239200088712905
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,7168,128,0.015152000718646579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,1536,8192,0.011408000356621213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,6144,7168,0.07461688915888469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,512,2048,0.014179555906189812
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3584,3072,0.00977955593003167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,7168,10240,0.016435555285877652
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,12288,8192,1.068394660949707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,64,256,0.0032479998966058097
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,65536,3072,0.0858888891008165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,16384,5120,0.4004373285505507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,3584,2560,0.020211555891566806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,7168,256,0.00962222201956643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,7168,2560,0.009017777939637503
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,1024,1024,0.005952000204059813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,7168,2560,0.009296000003814697
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,1024,12288,0.037046220567491316
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,128,1536,0.003866666720973121
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,65536,10240,13.493470085991753
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,8192,12288,0.05164444446563721
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,65536,1024,0.025011556016074285
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,12288,65536,0.14741244581010607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,16384,12288,0.5575137668185765
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,6144,6144,0.05921510855356852
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,768,1024,0.007656888829337225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,65536,512,0.05801510810852051
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,12288,32,0.003526222374704149
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1024,10240,0.018725333942307364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,5120,256,0.007269333634111617
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,4096,512,0.012814222110642327
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,128,3584,0.007672000262472365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,128,3584,0.014462222655614218
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,16384,4096,0.018220444520314533
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,10240,12288,0.058224002520243325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3584,65536,0.07233333587646484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,256,10240,0.020984888076782227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,2560,5120,0.00831911133395301
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,5120,16384,0.021911111142900255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,512,1536,0.013613333304723104
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,10240,32,0.008715555899673039
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,128,10240,0.012531555361217923
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,2560,2560,0.006948444578382704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,3584,16384,0.036079999473359846
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,2560,1536,0.0064702220261096954
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,65536,4096,0.039749334255854286
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,256,1536,0.0038497779104444715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,4096,4096,0.010308444499969482
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,7168,32,0.0021297778520319196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,6144,10240,0.016975111431545682
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,3584,6144,0.02626488937271966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,10240,3584,0.029846221208572388
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,4096,3584,0.027441778116756018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,12288,512,0.006978666616810693
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,128,512,0.0058746664888328975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2048,128,0.0028533333291610083
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3072,1536,0.006613333192136552
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3584,512,0.0031928889867332247
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,256,5120,0.050033777952194214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,2560,6144,0.009689778089523315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,6144,5120,0.013563555147912769
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,32,8192,0.010673777924643623
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,5120,16384,0.056926223966810435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,10240,2560,0.03091288937462701
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,6144,128,0.005595555735958948
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1024,2560,0.004936000125275718
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,1024,1536,0.0059164443777667145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,10240,1024,0.03858400053448147
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,4096,4096,0.011768889096048145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,7168,8192,0.014881778094503613
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,10240,4096,0.025232889586024817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3584,3584,0.007651555869314406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,16384,768,0.019281778070661757
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,3584,4096,0.01038577738735411
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,4096,10240,0.027512888113657635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,1536,1536,0.01447111037042406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,12288,32,0.06315733326805963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,8192,10240,0.036200887627071805
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,4096,6144,0.06648888852861193
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1024,16384,0.020288000504175823
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,512,768,0.005961777849329843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,12288,64,0.062048892180124916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,7168,1536,0.021124444074100916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,4096,4096,0.010072888599501716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3072,3072,0.006607111129495833
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,10240,2048,0.00980088859796524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,256,3584,0.006279110908508301
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,10240,5120,0.01572355628013611
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,5120,4096,0.014946666028764514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,12288,768,0.018203554881943595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,7168,12288,0.03955555624432034
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,512,16384,0.031174222628275555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1024,5120,0.03440711233350966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,768,7168,0.01917866700225406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,1536,1024,0.0057813334796163775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,12288,7168,0.2603813277350532
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,768,3584,0.008379555410808986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,32,5120,0.004934222333961063
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1024,6144,0.0101724440852801
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,7168,2048,0.011310222248236338
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,8192,6144,0.03951199849446615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,7168,4096,0.013586666848924426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,512,65536,0.058509336577521436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2560,32,0.0021679999513758556
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,12288,2560,0.012424888710180918
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,16384,6144,0.03360266817940606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,10240,2048,0.011934222446547614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,6144,6144,0.039782222774293684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,512,8192,0.02197422252760993
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,10240,3584,0.03309066759215461
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,256,3072,0.007798222204049428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,1024,256,0.005244444641802046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,1024,65536,0.08178310924106173
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,768,768,0.015828443898095023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,2048,7168,0.03538577755292257
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,7168,512,0.01724088854259915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,256,512,0.009115555220180089
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,4096,1536,0.004187555362780889
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,1536,7168,0.03450311223665873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1536,256,0.005607111172543631
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,512,768,0.0058417779703934984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,2560,768,0.016576000385814242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,7168,512,0.01724177764521705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,256,4096,0.00796088907453749
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,65536,16384,4.68922381930881
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,12288,6144,0.059614221254984535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,5120,8192,0.02320266597800785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1536,4096,0.023260444402694702
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3072,12288,0.018600889378123812
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,256,2560,0.007336888876226213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,256,32,0.005496888938877318
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,4096,10240,0.07656355698903401
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,512,256,0.005233777893914117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2560,6144,0.011157333023018308
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,256,1024,0.0031617778456873367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,16384,12288,0.06820888651741876
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,512,65536,0.3177439901563856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,128,8192,0.016168889072206285
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,7168,1024,0.006631111105283101
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,2560,256,0.014855111638704935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,8192,8192,0.021401777863502502
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,64,64,0.006963555183675554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,4096,16384,0.026850667264726426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,16384,6144,0.09545333517922296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2560,32,0.00564533347884814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,65536,5120,0.1100471085972256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,6144,2048,0.008006222546100616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1024,64,0.0058986664646201665
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,12288,128,0.010421333213647207
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,6144,512,0.042261332273483276
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,8192,3072,0.06574844651752047
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,1024,8192,0.052016887399885386
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,256,128,0.0058577776783042485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1024,64,0.005947555518812603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1024,10240,0.024271999796231587
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,7168,8192,0.015200000670221118
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,16384,64,0.002559111143151919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,4096,8192,0.06266399886873034
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,8192,5120,0.012297777665985955
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,2048,10240,0.04739288820160759
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,65536,256,0.008423999779754216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,512,1536,0.006227555374304454
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,7168,768,0.006243555496136348
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,4096,1024,0.00848444468445248
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3072,128,0.005727111051479976
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,4096,5120,0.013044444223244986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,768,6144,0.008289777570300633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,1024,3072,0.009696889254781935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1024,4096,0.005970666805903117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,512,12288,0.023123555713229712
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,128,1024,0.011735999749766456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3584,3072,0.007382222347789341
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,128,3072,0.023982221881548565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,6144,10240,0.033486220571729876
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2560,4096,0.008340444829728868
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,3072,8192,0.012096888489193387
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3584,128,0.003937777959638172
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,256,6144,0.009649777577983009
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,1536,12288,0.016301333904266357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,4096,3584,0.009336888790130615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,65536,3072,0.5757768948872884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,4096,768,0.009869332942697737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,768,10240,0.0557253360748291
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,8192,65536,0.08171021938323975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,512,3584,0.015209777487648858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1024,12288,0.011405333048767514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,128,8192,0.010760000182522668
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1024,768,0.005967999911970562
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3072,1536,0.005666666560702854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,16384,8192,0.04287199841605293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,65536,3072,0.06006044149398804
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,7168,3072,0.007740444607204861
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,6144,12288,0.03721244467629327
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2560,1536,0.004300444490379757
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,1536,768,0.0070017774899800616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,7168,512,0.008353778057628209
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,128,512,0.0026008888251251644
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,65536,512,0.014225777652528552
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,65536,8192,0.39133156670464414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,2048,5120,0.01882577770286136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2560,2048,0.00794577764140235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3072,65536,0.07071644730038114
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,256,128,0.012120889292822944
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,768,8192,0.02759644389152527
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,6144,65536,0.23823110262552896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,12288,1536,0.010557333628336588
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1024,64,0.00227111101978355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2048,4096,0.0236124445994695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,256,1024,0.006551110910044775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,32,1536,0.004330666528807746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,3584,8192,0.10871289173762004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,512,65536,0.07396711243523492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,7168,10240,0.024665777881940205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,6144,16384,0.0551546679602729
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,12288,65536,0.14484000205993652
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,3584,128,0.01585155559910668
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,16384,2560,0.03515911102294922
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,512,10240,0.009111999637550777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,512,2048,0.004934222333961063
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,256,7168,0.010755555497275459
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,6144,3584,0.0587866637441847
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,768,16384,0.010585777461528778
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,768,6144,0.0210515558719635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,2048,6144,0.030172444052166406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3584,256,0.00598933340774642
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,1024,5120,0.01145511120557785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,3072,512,0.005576888720194499
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,4096,3072,0.034916443957222834
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1024,1536,0.005281777845488654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,512,10240,0.030452443493737116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,1536,256,0.01514844430817498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,4096,4096,0.010943999720944298
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,8192,3072,0.03970311085383097
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,512,2048,0.01723288827472263
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,6144,2048,0.006945778098371293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,6144,1536,0.012454222473833295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,768,16384,0.017529777354664274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,256,512,0.005214222189452913
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,3072,512,0.013773333695199756
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2560,2048,0.027273777458402846
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,1024,7168,0.030679110023710463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,5120,256,0.009339555270142024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,65536,8192,0.12066577540503608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,512,64,0.007954667011896769
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,7168,2048,0.021298666795094807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,128,4096,0.0041955556306574075
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,5120,4096,0.00794400026400884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,256,3072,0.005537777725193236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,8192,1024,0.004638222356637319
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,64,128,0.005682666682534748
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,16384,5120,0.027024888330035742
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,1024,768,0.005562666803598404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,16384,2560,0.02612000041537815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,6144,7168,0.02016444504261017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,3584,12288,0.016813332835833233
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,2560,2560,0.0158595558669832
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,1024,10240,0.01260444439119763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,2560,6144,0.014171555638313293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,512,12288,0.20756267176734078
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2560,1536,0.006587555425034628
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2560,128,0.011710222396585675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,5120,1536,0.01683022247420417
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,65536,768,0.019261333677503798
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3584,1536,0.006621333460013072
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,768,16384,0.019944889677895438
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,3072,1536,0.01759199963675605
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1024,6144,0.010390222072601318
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,2560,7168,0.0391591125064426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,10240,16384,0.05277688966857063
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,1536,2560,0.006933333145247565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,256,16384,0.02788355615403917
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,10240,65536,0.20410488711463082
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,5120,128,0.015549333559142219
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,10240,128,0.006360888895061281
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,32,16384,0.00591822216908137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,1536,2048,0.014480888843536377
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,3072,5120,0.0329386658138699
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,2048,4096,0.007618666523032718
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,7168,1536,0.006582222051090664
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,2048,768,0.005960000058015187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,512,5120,0.030049777693218652
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,64,12288,0.10212177700466579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2048,4096,0.008897778060701158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,512,4096,0.007679111427730984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,4096,65536,0.10784977674484253
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,8192,8192,0.019366222951147292
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2048,10240,0.016128889388508268
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,6144,7168,0.022846221923828125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,1536,128,0.014473777678277759
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,256,2560,0.0226239992512597
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,768,128,0.014434667097197639
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2560,5120,0.009911999934249455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,6144,768,0.005564444594913059
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2048,16384,0.03460266523891025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1024,64,0.006601777755551868
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,32,3072,0.00793688909875022
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3584,1024,0.004537777768241035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3584,1536,0.011053333679835001
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,512,2048,0.045921779341167875
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,5120,65536,0.14221244388156468
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,8192,1024,0.01038933296998342
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,128,16384,0.017156443662113614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,32,1024,0.0028391112056043413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,32,512,0.0024888888001441956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,16384,3584,0.018602665927675035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,65536,1536,0.021851556168662176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2048,65536,0.06239289045333862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1536,4096,0.014475555883513557
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,7168,7168,0.060589333375295006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,12288,512,0.00517511119445165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,256,3072,0.010888888604111142
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,7168,768,0.02221600049071842
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,3072,8192,0.024159111910396155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,5120,32,0.004496888981925117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,65536,7168,0.07468355364269681
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,2560,512,0.013789333403110504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,2560,2560,0.015789333317014907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,3072,4096,0.02606311109330919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2560,3584,0.00792711145348019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,256,256,0.005945777727497949
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,6144,5120,0.040638221634758845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,3072,2048,0.006583111153708563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,12288,2560,0.014188443620999655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,3584,16384,0.022409778502252366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,256,12288,0.04014844364590115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3584,12288,0.020755555894639757
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,7168,6144,0.028559999333487615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,7168,1024,0.006229333165619109
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,64,3584,0.010672888822025724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,1024,256,0.007659555309348636
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,8192,5120,0.13580089145236546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,65536,5120,0.05549155672391256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2560,3072,0.010588444769382477
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,65536,512,0.01646133263905843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,3072,5120,0.009357333183288574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,3584,256,0.005576888720194499
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,12288,2560,0.013393777940008374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,16384,2560,0.030322667625215318
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1536,4096,0.0074888889988263445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,128,2560,0.013441777891582914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,12288,1536,0.024067555864651997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,16384,12288,0.06333244509167142
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,7168,32,0.0027191111197074256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,768,16384,0.12479733096228705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,12288,3584,0.0363031095928616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,16384,1024,0.009763555394278632
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,768,12288,0.008913777768611908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,7168,32,0.006981333096822103
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,256,10240,0.006236444330877728
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1024,3072,0.007254222200976477
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,5120,256,0.005566222386227713
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2048,512,0.019848888119061787
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,2048,128,0.01443911095460256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,128,768,0.0028453332682450614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,512,768,0.0028151110228565005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3072,128,0.012735111018021902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,128,65536,0.07402755816777547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2560,5120,0.06463822391298082
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,6144,768,0.007285333342022366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,512,1024,0.0032382222513357797
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,65536,3072,0.09451378054089016
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,5120,256,0.003925333420435588
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,5120,768,0.005929777605666055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,65536,3072,0.15444178051418728
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,65536,1536,0.03568889035118951
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,5120,3584,0.026558222042189703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1024,32,0.006266666783226862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,4096,12288,0.0557875567012363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1536,2048,0.021163556310865615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,4096,3072,0.007322666545708974
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,16384,1536,0.010727110836240979
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,8192,2048,0.03603022297223409
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,768,6144,0.009120889008045197
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,64,32,0.00590844452381134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,6144,65536,0.1247146659427219
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1024,65536,0.057617776923709445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,6144,32,0.006675555474228329
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,12288,3584,0.018914666440751817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,256,5120,0.008586666650242275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,2048,65536,0.1585208839840359
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,16384,32,0.002181333386235767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,512,2048,0.0069875551594628235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,128,4096,0.015152000718646579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,65536,7168,0.0743502246008979
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1536,5120,0.014705777168273926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,128,1536,0.006973333656787872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,12288,3072,0.013524444566832649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,10240,16384,0.04144444399409824
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2560,768,0.0028613333900769553
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3072,768,0.007685333490371704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,16384,256,0.017542221479945712
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3584,2048,0.007287999822033777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,65536,256,0.030951112508773804
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,4096,6144,0.013375110924243927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,10240,65536,0.1874764495425754
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,768,3584,0.008038221961922115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,5120,7168,0.013780444032616086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,128,2048,0.006721777634488211
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3072,256,0.005769777629110549
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2048,8192,0.01402933398882548
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,2560,1024,0.014348444011476306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,12288,3584,0.05637066894107395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,64,1536,0.003814222084151374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,10240,64,0.006173333360089197
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,128,3072,0.013372444444232516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,3584,6144,0.01197866681549284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,2048,3072,0.0158595558669832
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,1024,3072,0.008054222497675154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,8192,2048,0.010590222146775989
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,64,12288,0.005614222337802251
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3072,8192,0.019925332731670804
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,8192,7168,0.045019557078679405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,3072,12288,0.017312000195185345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,1536,512,0.013246222502655454
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,512,512,0.005930666708283954
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,4096,2048,0.023863999380005732
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,512,12288,0.01500266624821557
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,128,1536,0.006108444597986009
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,768,65536,0.02386844489309523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2048,2560,0.017080000705189176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,768,3072,0.005255110975768831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,768,2560,0.005232000102599462
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,32,3584,0.004257777912749184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,256,256,0.012465777496496836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,256,7168,0.01793333391348521
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,65536,128,0.03543466660711501
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,16384,65536,2.8322249518500437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,65536,7168,2.0460684034559464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,12288,1024,0.024471999870406255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,768,10240,0.0138586668504609
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,64,8192,0.011395555403497485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,3072,65536,0.9154159757826063
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,5120,2048,0.027097778187857732
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,64,10240,0.012110222544935016
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,256,64,0.0021688888470331826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,4096,32,0.0024275556206703186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,5120,1536,0.0069653333889113525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,8192,256,0.01791822248035007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2048,4096,0.010433778166770935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,12288,6144,0.22109244929419625
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1024,256,0.005631111148330901
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1024,6144,0.018384888768196106
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,6144,32,0.019552888141738046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,32,8192,0.029303110308117334
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3072,3072,0.010634666515721215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2048,1024,0.00832711077398724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3072,3584,0.017907554904619854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,768,4096,0.011056889262464313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,7168,32,0.0046968890560997855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,128,5120,0.01551644504070282
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3584,32,0.006005333529578314
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,8192,2048,0.009092444347010719
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3584,16384,0.031049778064092
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,6144,768,0.009389333426952362
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,512,65536,0.02789155642191569
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,5120,2560,0.006635555376609166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,256,5120,0.00830844458606508
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,6144,2560,0.025176000263955858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,12288,65536,11.259788513183594
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,6144,16384,0.040511111418406166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,10240,1024,0.0069448889957533936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1024,128,0.003998222036494149
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,8192,6144,0.020671111014154222
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,16384,3584,0.025781333446502686
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,65536,16384,0.3528764512803819
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,512,32,0.0018924443672100704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,12288,12288,0.9111466937594943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,8192,4096,0.011305777562989129
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3072,3072,0.024819556209776137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,2560,768,0.00565155554148886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,5120,4096,0.023385778069496155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,6144,768,0.00378577783703804
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,8192,2048,0.1590613391664293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2048,256,0.0057546666098965546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,7168,4096,0.010655110908879174
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,128,768,0.005980444451173146
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,768,16384,0.08391377660963271
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,256,3584,0.005249777601824866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,12288,12288,0.0628124475479126
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,64,12288,0.03660088777542114
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,32,3584,0.004967111266321606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,65536,256,0.015304888288180033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,10240,2560,0.012720888687504662
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,7168,1536,0.006372444331645966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,5120,512,0.015909334023793537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,6144,1536,0.006313777632183499
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,7168,128,0.0022799999763568244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,65536,8192,0.18214755588107637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,8192,256,0.017602667212486267
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,128,2560,0.014164444473054675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,8192,16384,0.10728355248769124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,1024,10240,0.012455110748608908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2048,1536,0.03377066718207465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3072,1536,0.006671111202902264
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,512,768,0.005707555760939916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,256,10240,0.15744354989793566
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,512,7168,0.010794666906197866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,256,65536,0.016806221670574613
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,1536,3072,0.019281778070661757
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,128,65536,0.04872266782654656
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,768,3584,0.025923555095990498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,4096,3584,0.007702222300900354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,10240,3584,0.022848000129063923
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,65536,32,0.0052728888889153796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,12288,5120,0.04086577892303467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,128,768,0.0028799999919202593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,10240,3072,0.07983733548058404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,128,1024,0.0057546666098965546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,6144,2560,0.009630222287442949
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3584,768,0.020639111598332722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,5120,12288,0.020344000723626878
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,128,1536,0.014829332629839579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,16384,256,0.006579555571079254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,32,768,0.005917333480384614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,3584,1536,0.005948444621430502
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,7168,64,0.011857777833938599
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1536,6144,0.010377777947319878
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,8192,1024,0.004279110994603899
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,12288,768,0.0069848886794514125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,12288,3584,0.3946959972381592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,64,8192,0.01072355525361167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,7168,2560,0.007332444190979004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,768,512,0.005991999887757831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,32,12288,0.014105778601434497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,4096,3072,0.04082489013671875
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,6144,64,0.03295733200179206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,256,4096,0.005988444305128521
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,65536,3072,0.09363022115495469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,128,10240,0.012139555480745105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,512,16384,0.008489777644475302
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,2048,16384,0.019912888606389362
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,128,32,0.00589155571328269
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,3072,5120,0.031193776263131037
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,5120,65536,0.15830666489071316
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,512,2560,0.006956444846259222
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,768,2560,0.01481244464715322
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,4096,128,0.022824888428052265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,1536,7168,0.014459555347760519
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,6144,1024,0.018253333038753934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,10240,8192,0.04511733187569512
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2048,64,0.0029306666304667792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,7168,5120,0.03624711102909512
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,5120,7168,0.014331556028789945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,768,256,0.011009777585665384
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,3072,16384,0.23082399368286133
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,768,2560,0.024471110767788355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,768,128,0.0021608887861172357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1536,1536,0.006606222026877933
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3584,768,0.034274667501449585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,10240,65536,0.4909493128458659
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,4096,12288,0.018230222993426852
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,7168,32,0.0022133334229389825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,3584,2560,0.025200888514518738
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,1024,2560,0.006675555474228329
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,3072,3584,0.007349333001507654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,128,6144,0.005987555616431766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,256,2560,0.004954666727119022
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,10240,1024,0.0064693333374129395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,12288,512,0.006640888750553131
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,3072,12288,0.050985776715808444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,64,32,0.0018844445132546956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,768,128,0.015146666102939181
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,65536,6144,0.12709244092305502
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,512,65536,0.0794408917427063
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,12288,2048,0.06491288873884413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,1536,512,0.013438222308953604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,256,128,0.012048888537618848
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,5120,256,0.005423111220200856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,512,128,0.013210666676362356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,1024,2048,0.006656888872385025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1024,2048,0.004217777815130022
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,6144,5120,0.024664888779322307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,6144,256,0.014275555809338888
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1536,768,0.003527111063400904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,128,3584,0.014688000082969666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,65536,4096,0.09991200102700128
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,512,512,0.0055440002017551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,4096,2048,0.013530666629473368
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,7168,12288,0.023747555083698694
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,64,1024,0.00342399999499321
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1536,768,0.013199999928474426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,2560,7168,0.029286222325430974
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,128,128,0.012423999607563019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,6144,128,0.002567111204067866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,16384,3584,0.04116444455252753
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,256,256,0.0053226666318045715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,64,256,0.005588444570700328
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,12288,2560,0.030396444929970637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2560,768,0.003967111309369405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,32,10240,0.004906666775544484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,3072,1024,0.014529777897728814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1536,2048,0.00701333334048589
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,768,2048,0.0064586665895250105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,3072,3072,0.007360888852013483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,512,768,0.005973333285914526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,12288,6144,0.01812977757718828
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,2560,768,0.016383111476898193
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,128,128,0.005196444276306364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,16384,6144,0.023445333043734234
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,768,256,0.005936000082227919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3072,10240,0.3083226680755615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,6144,128,0.015120888749758402
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,1536,1024,0.01648266613483429
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,5120,1024,0.00590133335855272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,2560,3072,0.01870577699608273
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,6144,3584,0.026649778087933857
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,4096,256,0.0053119998839166425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,512,5120,0.04853777752982246
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,32,768,0.007602666815121968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,7168,12288,0.030647112263573542
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1024,3584,0.0059164443777667145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2560,4096,0.008249777886602614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,10240,12288,0.03931822379430135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,128,3584,0.009676444033781687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,1024,65536,0.06266399886873034
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,16384,768,0.008469333251317343
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,10240,1536,0.11741244792938232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,5120,65536,0.33403555552164715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,4096,12288,0.01963644391960568
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1024,32,0.0021786666992637846
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3584,64,0.00591111100382275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,1024,2560,0.015199111567603217
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,5120,1024,0.006991110742092133
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,12288,2560,0.012002666791280111
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,1536,7168,0.009700444837411245
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1536,1024,0.005968889014588461
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3584,3072,0.016598222984208
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,5120,512,0.006619555668698416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,768,1536,0.01459822224246131
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,768,32,0.0063075555695427795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,6144,65536,0.12380533748202854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,3584,7168,0.010449777874681683
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,5120,4096,0.015957333975368075
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,10240,10240,0.04282844397756788
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,65536,7168,0.09538400173187256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,5120,2560,0.023005333211686876
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,32,64,0.005630222045713001
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,128,65536,0.9290559556749133
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3072,4096,0.00925866680012809
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,7168,6144,0.046874665551715426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,6144,64,0.0026497778793176017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,12288,12288,0.04907555712593926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,5120,32,0.010581333604123859
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,12288,3584,0.013702222042613559
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,1536,7168,0.010004444254769219
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,3072,12288,0.02239111065864563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,768,7168,0.019833778341611225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,3072,8192,0.015293333265516492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3584,128,0.0023431111541059283
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,3072,1536,0.018578666779730055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,3584,128,0.021810667382346258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,6144,1024,0.0052435555391841466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,256,1024,0.019620445039537217
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1024,2048,0.005620444400442972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,16384,10240,0.031883554326163396
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,65536,32,0.007017778025733099
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,768,5120,0.01965688831276364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,2560,10240,0.026406221919589575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,7168,5120,0.011699555648697747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,2048,8192,0.023440000083711412
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,512,8192,0.01998933321899838
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,512,128,0.005271111097600725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,65536,256,0.1779928869671292
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,10240,2560,0.010055111514197456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,128,3072,0.007261333366235097
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,7168,2048,0.008072000410821702
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,65536,768,0.016189333465364244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,3072,1024,0.005933333188295364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,2560,16384,0.04163022173775567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,768,4096,0.00925866680012809
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,12288,5120,0.3710177739461263
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,8192,3072,0.029716445340050593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,65536,32,0.0042453333735466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,7168,16384,0.02375288969940609
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1536,6144,0.030886222918828327
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,4096,7168,0.01977511081430647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,4096,65536,0.07022044393751356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1024,5120,0.009347555538018545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,768,2560,0.004957333207130432
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,512,1536,0.015300444430775113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,1024,2560,0.015849777393870883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,4096,128,0.01181688904762268
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3072,3072,0.0070773329999711775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,16384,12288,0.036455111371146307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2560,8192,0.03311644328965081
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2048,6144,0.010030222435792288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,3072,2560,0.049313777022891574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,32,7168,0.010056888891590966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2048,1024,0.0052800000541739995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,10240,256,0.0074577778577804565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,7168,128,0.00555288874440723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,64,2048,0.004318222403526306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,512,64,0.005603555589914322
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,256,64,0.0021546667234765156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,65536,3072,0.19632000393337676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,5120,10240,0.01717422240310245
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,12288,3072,0.013795555465751223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,512,4096,0.07603110869725545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,10240,128,0.006280888699822956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,16384,16384,0.07306044631534152
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3072,1536,0.0070364442136552595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,7168,6144,0.016344888342751395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,2560,256,0.014863999353514777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,1024,256,0.01478488908873664
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,32,8192,0.010401777923107147
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2048,4096,0.014081777797804939
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,32,64,0.0018079999006456798
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,10240,65536,0.5256649123297797
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,3584,6144,0.009298666483826106
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,10240,3584,0.050323555866877236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1024,8192,0.010857777463065254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,12288,32,0.008979555633332994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,10240,8192,0.05901866488986545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,8192,1536,0.021577777134047613
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,7168,5120,0.2516417768266466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,6144,8192,0.03160444564289517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,4096,12288,0.03418488966094123
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,65536,1536,0.02217688825395372
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2048,3072,0.007315555380450354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3072,7168,0.011128889189826118
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,12288,16384,0.7155973116556803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,128,768,0.0059262220230367445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,16384,3072,0.01796266602145301
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,64,2560,0.022849778334299724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,4096,3584,0.007747555772463481
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,10240,256,0.0063031112982167145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,32,768,0.007996444072988298
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,512,6144,0.009733333355850643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,7168,10240,0.020067556036843192
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,6144,4096,0.012082666986518435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3584,6144,0.012191111014948951
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,256,5120,0.008363555702898238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,16384,6144,0.05228711167971293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,64,1536,0.006596444381607904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,7168,12288,0.03944622145758735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,3584,128,0.00552977787123786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,6144,7168,0.03843733337190416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,1536,65536,0.06272266970740424
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,3072,10240,0.015296889675988091
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,512,1024,0.005578666511509154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,8192,12288,0.02051466703414917
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,2048,4096,0.01684177749686771
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,7168,4096,0.015045333239767285
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1536,3584,0.007603555917739868
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,128,128,0.0052382221652401825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,7168,16384,0.027504000398847792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,16384,1024,0.008344888687133789
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1536,7168,0.10618399911456639
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,4096,1024,0.0032177778581778207
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,6144,2048,0.021653332644038733
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,10240,2048,0.033088889386918806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,12288,128,0.002834666727317704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,16384,6144,0.03136000037193298
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1536,1024,0.004150222159094281
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,32,3072,0.006928000185224745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,12288,1536,0.015235554840829639
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,32,256,0.005562666803598404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,256,4096,0.009693333672152625
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,12288,65536,1.8654276529947917
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,12288,128,0.006342222293217977
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,7168,3584,0.10043111112382676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,7168,64,0.02204977803760105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,5120,6144,0.01348533315791024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,16384,512,0.022691556149058874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,16384,128,0.01236088905069563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1024,2560,0.007299555672539606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,16384,768,0.06234755780961779
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,32,12288,0.011751111182901593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,10240,8192,0.2441493405236138
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,512,16384,0.03502666619088914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,10240,1536,0.044900443818834096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,1024,10240,0.02372177773051792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,7168,7168,0.3108853234185113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2560,12288,0.01757955551147461
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,256,10240,0.010958222051461538
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,128,1024,0.006116444451941385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,8192,8192,0.022340445054901972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,8192,1536,0.006575111299753189
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,6144,5120,0.012812444733248817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,512,8192,0.005649777750174205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,2560,7168,0.03013955553372701
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1536,6144,0.00867377801073922
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1536,768,0.005879111174080107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,128,768,0.0028293333533737394
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,3072,4096,0.1251173284318712
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,512,512,0.012415111064910889
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,8192,6144,0.01939288940694597
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2560,1536,0.0038515557017591265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,3072,5120,0.01965688831276364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,5120,6144,0.0776124464140998
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,2560,65536,0.06764622529347737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,65536,128,0.005646222167544895
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,2560,65536,0.0889973309304979
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,12288,7168,0.03724888960520426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,1536,16384,0.03571555680698819
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,16384,4096,0.022537777821222942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2048,256,0.006868444383144379
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,6144,6144,0.014150222142537435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,2048,16384,0.022328888376553852
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,768,3584,0.015649777319696214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,16384,3584,0.02010844482315911
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,16384,7168,0.0255057778623369
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,1024,5120,0.01585244470172458
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3584,3584,0.008286221987671321
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,2048,1536,0.016937777400016785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,768,1024,0.01314311143424776
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,12288,6144,0.04637599984804789
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3584,16384,0.05392000079154968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,64,6144,0.004905777672926585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2560,65536,0.06112088759740194
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,8192,4096,0.010612444745169746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2560,2048,0.006605333338181178
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,16384,256,0.01756355497572157
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,4096,2048,0.009522666533788046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,8192,4096,0.013413333230548434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,7168,5120,0.01313333296113544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,7168,1024,0.019658666517999437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,12288,3584,0.023264888260099623
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3584,256,0.005953777581453323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,6144,2048,0.02176711128817664
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,7168,128,0.008020444048775567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,6144,768,0.005889777921968036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,3584,12288,0.0372613337304857
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1024,1536,0.010953777366214328
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,512,2048,0.008967999782827165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,7168,3072,0.012375999655988483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,768,10240,0.01184800018866857
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,65536,3584,0.059952000776926674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,6144,768,0.006016888966162999
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2048,12288,0.026391999589072332
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,6144,512,0.005564444594913059
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2048,6144,0.009735999835862054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,10240,65536,0.30399378140767414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,10240,1024,0.02141600019401974
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,512,2560,0.014467555615637036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,3072,12288,0.06002489063474867
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,6144,3072,0.01090844472249349
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,256,2048,0.013769778112570444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,1536,256,0.0053075556125905775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2560,256,0.007266666326257918
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,4096,6144,0.013266666895813413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3584,5120,0.009786666267448002
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,128,4096,0.007664889097213745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2048,2048,0.016057777735922072
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1536,64,0.005551110953092575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,12288,3584,0.11072266764110988
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,64,3072,0.004528000122971005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2560,10240,0.04119288921356201
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,6144,65536,0.08247111241022746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,256,65536,0.058331555790371366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,10240,4096,0.016280000408490498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,32,128,0.002135111019015312
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2048,2048,0.010079999764760336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1536,10240,0.01238044434123569
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,32,5120,0.009001777403884465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,10240,2048,0.010332444475756751
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,256,6144,0.010088000032636855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,32,512,0.009467555416954888
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,4096,1024,0.005717333406209946
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,5120,5120,0.009026666482289633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,16384,2048,0.03115733464558919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,16384,512,0.09969866938061184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,8192,5120,0.017687999539905123
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,768,10240,0.024393777052561443
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,6144,2560,0.007999111380841997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,16384,4096,0.019300444258583915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,3584,768,0.017196445001496207
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,256,7168,0.010450666977299584
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,512,6144,0.005962666538026597
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,256,8192,0.010416888528399997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3072,3072,0.02678311036692725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,16384,8192,0.030951112508773804
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,768,5120,0.01734133395883772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,10240,16384,0.040975110398398504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,64,256,0.0022088889446523455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3584,512,0.0034053333931499054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,64,10240,0.08299466636445788
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,5120,5120,0.011010666688283285
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3072,10240,0.016208888755904306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,128,5120,0.015185778339703878
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,2560,3072,0.022006221943431433
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,65536,12288,0.21745422151353624
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,256,768,0.005969777703285217
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,5120,7168,0.028951111767027114
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,1536,2048,0.006707555717892117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,256,32,0.0021964444054497615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,10240,65536,0.18984533680809867
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,512,256,0.002687999978661537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,8192,10240,0.07383555836147733
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,3072,7168,0.023376888699001733
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,8192,8192,0.04878044459554884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,10240,256,0.026486222942670185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,5120,1536,0.006968888971540663
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,768,2048,0.015513777732849121
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3072,32,0.005576000031497743
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1536,1536,0.003929777691761653
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,512,1536,0.006296000132958095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,256,2560,0.01365244471364551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,512,64,0.005611555443869696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,5120,7168,0.025644444757037695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,65536,3072,0.04807555675506592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,10240,2560,0.030729777283138696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,5120,2048,0.0069226668112807805
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,1536,10240,0.014631110760900708
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,1536,256,0.014911111858155994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,8192,5120,0.014527999692493014
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,32,8192,0.017337777548366122
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,4096,32,0.0052666668262746595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,32,12288,0.013398222625255585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3584,128,0.0053484443989064955
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,7168,3072,0.01184800018866857
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,32,1536,0.007643555601437886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,65536,7168,0.136172440316942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3072,256,0.006986666884687211
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,3584,128,0.005325333525737126
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2560,16384,0.01795022189617157
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,32,256,0.0025457777082920074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,32,128,0.00563466673096021
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,7168,4096,0.026865777042176988
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,1536,7168,0.009729777773221334
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1536,65536,0.03578133384386698
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3584,8192,0.014918223023414612
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,1536,2560,0.006986666884687211
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,1536,512,0.005598222215970357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,4096,5120,0.03379022081693014
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,128,16384,0.015569777952300178
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1024,5120,0.008994667066468133
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,3584,3584,0.008013333711359236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,7168,768,0.00590844452381134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,3584,4096,0.007733333441946242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,128,512,0.010693333215183683
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,6144,10240,0.01719999975628323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1024,3072,0.005623999983072281
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,128,512,0.01883288886811998
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,4096,2048,0.006621333460013072
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,4096,256,0.005959110955397288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,32,7168,0.0151555554734336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,6144,128,0.0028364445186323593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,12288,7168,0.25340623325771755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,65536,10240,0.2023031049304538
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,1536,10240,0.01181244436237547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3072,5120,0.011499555574523078
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,4096,7168,0.013529777526855469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,8192,512,0.005375999957323074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,65536,128,0.1993431117799547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,1536,65536,0.06276088953018188
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,5120,3072,0.009002666506502364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1536,2048,0.004354666504595014
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,8192,5120,0.014490667316648694
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,256,7168,0.03466755482885573
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,128,6144,0.009031111167536842
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,8192,64,0.0023768889821237987
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,4096,7168,0.024726221958796184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1024,1024,0.018287110659811232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,4096,3072,0.007242666350470648
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,4096,128,0.005616000129116907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,7168,7168,0.014140443669425117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,5120,6144,0.2031573322084215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,12288,16384,0.15564000606536865
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,5120,1536,0.0047848886913723415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,1536,10240,0.011910222470760345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,4096,1536,0.0196426659822464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,7168,7168,0.013366222381591797
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,10240,10240,0.04309422108862135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,8192,8192,0.025241777300834656
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,8192,32,0.00562755556570159
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,10240,2560,0.011397333608733283
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1024,768,0.00351555562681622
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,512,3584,0.007323555648326874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,512,128,0.005573333137565189
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,4096,16384,0.02256533337963952
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,2048,6144,0.03429422113630507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,512,6144,0.025474665893448725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1024,1024,0.0058906666106647914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,7168,8192,0.023170666562186346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,768,12288,0.015033778217103748
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,65536,12288,14.17200893825955
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,7168,3072,0.016484444340070088
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,12288,10240,0.061885330412122935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,5120,1536,0.00658044425977601
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,65536,1536,0.45504622989230686
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1536,32,0.0024871110088295406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,768,2560,0.009309333231714036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,128,1024,0.005567111074924469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2560,5120,0.023006222314304773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,16384,4096,0.01923466722170512
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,5120,512,0.005893333504597346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3072,1024,0.018497778309716117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,5120,768,0.009436444275909001
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,10240,12288,0.030777778890397813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2048,7168,0.011476444701353708
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,5120,1024,0.005894222193294101
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,5120,12288,0.021681777305073206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,8192,6144,0.045349333021375865
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,10240,2560,0.01238844460911221
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,8192,512,0.04747733473777771
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,6144,64,0.0052168890833854675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,2560,16384,0.02093066606256697
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,4096,1024,0.005894222193294101
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,128,2560,0.013078221844302284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,32,2048,0.004201777693298128
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,2560,5120,0.029713776376512315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,256,1536,0.006269333263238271
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,12288,768,0.005586666779385672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,256,16384,0.027112000518374976
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,2560,128,0.01479200025399526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,16384,1536,0.013032000097963544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,256,4096,0.00832622249921163
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,16384,256,0.007324444750944774
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,256,1536,0.006218666831652324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,2560,16384,0.020634666085243225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,256,6144,0.009328000247478485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3584,64,0.0018533333722088074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,768,512,0.005210666606823604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,5120,65536,0.11745333671569824
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,16384,64,0.002483555633160803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1024,6144,0.008309333688682979
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2048,1536,0.0038648889296584656
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,128,2048,0.004222222086456087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,7168,128,0.002165333264403873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,2560,768,0.006616888774765863
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2560,3584,0.008003555238246918
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,512,3072,0.014793778459231058
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,7168,6144,0.016221332881185744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,65536,128,0.009672889278994666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1024,4096,0.007948444121413762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,768,65536,0.024200000696712073
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,256,1024,0.006246222390068903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,8192,2048,0.0069679998689227635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,512,6144,0.009332444104883406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,4096,2048,0.006583111153708563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,4096,6144,0.012839111189047495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,3584,3072,0.02202400068442027
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,4096,16384,0.06943466928270128
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,8192,128,0.005607999861240387
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,65536,2560,0.09198044406043158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,4096,3072,0.01942133406798045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,6144,6144,0.07996977700127496
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,32,128,0.0021555556191338436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,256,128,0.0022435554613669715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1536,2560,0.013991999957296582
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,256,128,0.0052915554907586836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,10240,3072,0.013838222457302941
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,65536,4096,0.09905066755082871
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,2560,256,0.027469333675172593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,3072,3584,0.017280000779363845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,6144,128,0.006588444527652528
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,768,1024,0.004641777939266629
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,768,2560,0.00742133292886946
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,2048,12288,0.015202666322390238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,1024,7168,0.016389333539538913
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1536,12288,0.014339556296666464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,128,8192,0.0105591110057301
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,1536,16384,0.019683554768562317
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,4096,1536,0.019337778290112812
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,256,768,0.014461333552996317
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,65536,128,0.03287555442916022
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,12288,65536,0.23451111051771376
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,3072,3584,0.018985778093338013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,12288,3072,0.0637484457757738
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,7168,6144,0.027662222584088642
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,8192,16384,0.6387235323588053
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,256,2560,0.010423111418883005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,768,3072,0.01126488877667321
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1024,6144,0.026263111167483862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,7168,256,0.00527022240890397
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,4096,5120,0.027314666244718764
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,6144,5120,0.06912889083226521
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,768,1536,0.014870221416155497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,7168,512,0.003301333429084884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,2048,768,0.005950222412745158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,768,128,0.005568000177542369
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,4096,3584,0.008573333422342936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,12288,2048,0.011583999627166323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,128,65536,0.06174311372968885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,12288,5120,0.04163022173775567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,5120,1536,0.006315555423498154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,4096,1024,0.005281777845488654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,10240,32,0.007768889268239339
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,16384,512,0.006714666469229593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,6144,1536,0.025199999411900837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,2560,5120,0.011964444484975604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1024,64,0.0028719999310043124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,2560,12288,0.03893688983387417
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,1536,256,0.015444444285498725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,128,12288,0.02330044408639272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,65536,6144,0.13115999433729383
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,3584,128,0.005975111077229182
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,1024,1024,0.005894222193294101
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,64,1536,0.00674222202764617
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,12288,512,0.006648889018429651
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1024,1536,0.0042071110672420925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,5120,2560,0.020005333754751418
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,6144,3584,0.010009778042634329
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,2560,768,0.017066667477289837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,512,1024,0.009495110975371467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3584,65536,0.0662319991323683
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,7168,12288,0.019345778557989333
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,128,512,0.009093333449628618
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,1536,2048,0.00665155549844106
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,4096,1024,0.004581333448489507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,128,64,0.0021866667601797315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,5120,3072,0.04063466522428725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,12288,2560,0.029870222012201946
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,1024,3072,0.007668444679843054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,4096,12288,0.027467555469936792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,256,7168,0.017909333109855652
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,4096,768,0.01689511040846507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,4096,5120,0.011852444046073489
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,64,12288,0.02513511147763994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3584,12288,0.01863822175396813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,512,8192,0.14221244388156468
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,5120,8192,0.03977333174811469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,2560,2048,0.02207999924818675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,256,768,0.005568888866239124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,65536,128,0.09475466940138076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,7168,256,0.01685422162214915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,4096,8192,0.021037333541446265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,5120,2560,0.00738933351304796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,8192,5120,0.01459377838505639
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1024,1024,0.007625777688291337
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,10240,64,0.006566222343179915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,8192,256,0.004643555730581284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,65536,2560,0.7810728814866809
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,8192,65536,4.92265150282118
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,512,4096,0.007975111405054728
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,64,10240,0.0069297779765393995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,3072,768,0.015988444288571674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,64,1536,0.006272000157170826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,512,128,0.010115555591053432
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,3584,2048,0.006809777683681912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,12288,4096,0.017646221650971305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1024,16384,0.019223110543357003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,7168,4096,0.01277511070171992
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,12288,10240,0.06173421939214071
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,7168,7168,0.0888248880704244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,512,3072,0.03813866774241129
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,768,10240,0.022335110439194575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,16384,3072,0.04776444368892246
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,7168,7168,0.5224711100260416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,512,12288,0.05540800094604492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,768,3584,0.00904622260067198
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,512,128,0.008046222229798635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,16384,12288,0.08586577574412028
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,16384,10240,3.1312728457980685
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,7168,12288,0.08136533366309272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,8192,2560,0.00980444418059455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,10240,5120,0.05032799972428215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,5120,16384,0.06611733304129706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,512,5120,0.01756800048881107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,5120,256,0.00666222224632899
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,65536,2560,0.050917334026760526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,6144,1536,0.018609777092933655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,12288,1536,0.034471998612085976
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1024,6144,0.009727111293209923
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,65536,2048,0.027082666754722595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,5120,768,0.005602666487296422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,128,1024,0.011784888803958893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,16384,128,0.00665866666369968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,7168,1024,0.010379555324713389
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,16384,256,0.009916444619496664
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,16384,4096,0.06677599747975667
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,10240,128,0.002941333378354708
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,10240,2560,0.04740622308519152
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,12288,16384,0.05332622263166639
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,768,512,0.005929777605666055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,4096,7168,0.07500622007581922
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,1536,16384,0.03183999988767836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,1024,3584,0.01553600033124288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,6144,1024,0.014575110541449653
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,16384,3584,0.014785778191354541
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,3584,2560,0.006928000185224745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,65536,7168,0.07743022176954481
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,2560,2048,0.016207110550668504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,512,8192,0.02749866743882497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,4096,12288,0.023519999451107446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,768,12288,0.013487111363146039
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3072,2048,0.009038222332795462
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,1536,1536,0.018606222338146634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,2560,4096,0.017156443662113614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,16384,4096,0.11646489302317302
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2560,16384,0.01830044388771057
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,10240,65536,2.9574559529622397
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,2048,4096,0.008716444174448649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,10240,5120,0.05931199921502007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,128,8192,0.022292445103327434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,4096,10240,0.05081511206097073
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,65536,128,0.019711110326978896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,8192,512,0.017913776967260573
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,8192,8192,0.11480800310770671
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,12288,768,0.015617777903874716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,5120,3072,0.012796444197495779
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1024,5120,0.008833777573373582
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,10240,2560,0.026887110537952844
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,8192,8192,0.025892444782786902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,6144,8192,0.016594666573736403
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,768,512,0.012830221818553077
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,768,3584,0.005642666584915585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,5120,3584,0.037957333856158786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,12288,2560,0.013331555657916598
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,32,2560,0.008833777573373582
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,4096,12288,0.34482399622599286
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,12288,3072,0.01552177800072564
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,10240,7168,0.019511111908488803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,1536,2048,0.01981511049800449
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,2560,10240,0.014292443792025248
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,32,6144,0.04978222317165799
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1536,32,0.0052560000783867305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,32,256,0.002573333266708586
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,768,7168,0.01036800030204985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,64,10240,0.005664888769388199
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,768,64,0.0019066666977273093
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,5120,768,0.017543110582563613
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,8192,512,0.006009777800904379
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1536,1024,0.006309333360857434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,12288,6144,0.021104888783560857
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,8192,3072,0.0288355549176534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,768,5120,0.09715822007921006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,7168,2560,0.009302222066455418
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,3072,8192,0.0307262208726671
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,5120,4096,0.011376000112957425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,12288,65536,0.13702578014797634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3584,7168,0.03919911053445604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1536,3584,0.0062977779242727495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,6144,1024,0.006270222365856171
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,65536,1024,0.019694222344292533
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,16384,6144,0.06479022237989637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,1536,768,0.009424000150627559
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,256,256,0.005255110975768831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2560,128,0.006976000136799282
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,7168,16384,0.08039822181065877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,16384,2560,0.01240888900227017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,5120,256,0.032348444064458214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,2048,2048,0.006322666588756774
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,2560,1536,0.005965333431959152
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,256,8192,0.021399999658266704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3584,128,0.005613333235184352
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,5120,2560,0.019667555888493855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,65536,512,0.015138667490747241
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,5120,10240,0.016955556141005624
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,65536,4096,0.06868533293406169
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,16384,5120,0.037455999188952975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,256,1536,0.006614222294754452
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,128,4096,0.011121778024567498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1024,1024,0.006285333385070165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,8192,6144,0.020049777295854356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,1536,10240,0.025939555631743536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,128,10240,0.005952888892756567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,6144,768,0.006030222194062338
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,64,6144,0.014512888259357877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,12288,10240,0.03491200009981791
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,64,128,0.0052933332820733385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,256,3584,0.007298666569921706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,2560,6144,0.026300445199012756
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,4096,3584,0.03100444542037116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,7168,4096,0.01997511088848114
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,16384,10240,0.052609778112835355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,6144,8192,0.019263111882739596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,8192,7168,0.016943110360039603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,12288,12288,0.3611777888403998
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,65536,16384,0.14346578386094835
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,5120,256,0.007623111208279927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,2048,256,0.017280888226297166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,8192,12288,0.0351128876209259
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2048,1024,0.007692444655630324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,16384,64,0.010093332992659675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,65536,3072,0.04287822379006279
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,512,1024,0.005677333308590784
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,16384,16384,0.04362666606903076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,256,8192,0.010699555277824402
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,5120,5120,0.0938933359252082
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,128,12288,0.04221422142452664
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,16384,1024,0.07413066758049859
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,256,2560,0.014448889427714877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,3072,16384,0.033047109842300415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2048,65536,0.10163378053241306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,8192,1536,0.006531555619504716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2048,256,0.00253866674999396
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,64,1536,0.003548444559176763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,16384,1024,0.018716444571812946
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,128,1536,0.006610666712125142
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,256,7168,0.0052933332820733385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,16384,64,0.0432293348842197
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,6144,2560,0.029645333687464397
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,256,16384,0.032482667101754084
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,768,768,0.006238222122192383
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,768,16384,0.015469332536061605
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,768,2560,0.007247111035717859
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,5120,5120,0.012161778079138862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,32,65536,0.00904977818330129
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2560,64,0.0019120000716712738
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,256,1536,0.00629688882165485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,1536,16384,0.10015200244055854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,64,256,0.002430222307642301
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,512,12288,0.10894755522410075
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,2048,512,0.014112888111008538
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,16384,1536,0.009685333404276106
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,65536,2560,0.033041778537962176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1536,10240,0.014561777313550314
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,512,16384,0.01627111103799608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,3584,12288,0.06382489204406738
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,512,2048,0.004655999855862724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2560,512,0.005967999911970562
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,16384,6144,0.019926221834288705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,7168,512,0.0064382221963670515
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2048,12288,0.017176000608338248
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,16384,32,0.046055111620161265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,10240,512,0.04251111216015286
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,32,512,0.005954666684071223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2048,10240,0.015509333875444202
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2048,3584,0.014064888159434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,128,3072,0.004967111266321606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,256,4096,0.007985778152942657
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,32,3584,0.008019555773999956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,1536,10240,0.1197626723183526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,8192,128,0.00979644474056032
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1536,128,0.003624888757864634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,1024,65536,0.062203552987840444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2048,16384,0.03198666705025567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,128,1024,0.0032631111227803757
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2048,7168,0.01073066641887029
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1536,7168,0.01144266708029641
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,1536,10240,0.011792889071835412
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,5120,1536,0.01664888858795166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,65536,4096,0.12152355247073704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,512,6144,0.014584889014561972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,256,3072,0.004923555586073133
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,3584,2048,0.007368000017272101
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,1024,256,0.01682133310370975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,3584,7168,0.027631999717818365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,3584,128,0.0052471111218134565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,256,16384,0.006673777682913675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,512,5120,0.009367110828558603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,128,256,0.0022115556316243275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3584,12288,0.0897128913137648
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1536,4096,0.007063999772071838
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,32,4096,0.005964444329341252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1536,3584,0.0069022224181228215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3584,8192,0.13492000102996826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1024,7168,0.010527110762066312
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,256,1024,0.005948444621430502
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,64,10240,0.005643555687533484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,32,1024,0.0063386667105886675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,12288,16384,0.06624711222118802
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2048,768,0.006046222315894232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,12288,10240,0.18073066075642905
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,4096,1024,0.017533333765135873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,12288,16384,0.04051822092798021
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,3072,65536,0.1261733373006185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,32,1024,0.005931555396980709
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,6144,768,0.008952889177534316
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,4096,7168,0.03616177704599168
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,768,768,0.015465777781274585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,12288,256,0.017504889104101393
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2048,7168,0.013552889227867126
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,2048,7168,0.0110871113008923
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,7168,512,0.0083137775460879
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,8192,256,0.005932444499598608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2560,3072,0.012438221938080259
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,7168,2560,0.010407110883129967
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2560,16384,0.03707999984423319
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,512,2048,0.014840000205569796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,4096,12288,0.03127022253142463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,256,3072,0.009715555442704095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,512,4096,0.0075866662793689305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2560,512,0.005586666779385672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,2048,3072,0.00700800038046307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3584,65536,0.06832355260848999
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,12288,65536,4.710687001546224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,256,16384,0.037738667594061956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,10240,3072,0.022505778405401442
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,2048,3584,0.03703022334310744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,2560,12288,0.016311999824311998
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2048,256,0.005949333310127258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1536,4096,0.008026666939258575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,1024,3072,0.007477333148320516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,1024,768,0.015533333023389181
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1536,256,0.0052720002002186244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,8192,5120,0.014223999447292753
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,768,2048,0.015951111912727356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,6144,6144,0.013979555832015144
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,4096,5120,0.0314008891582489
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,1024,8192,0.01793511046303643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,4096,65536,0.06751110818650988
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,768,8192,0.010727110836240979
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,1024,128,0.014481777946154276
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,6144,256,0.005584888988071018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,16384,3072,0.03151022063361274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,64,16384,0.006001777946949005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,16384,10240,0.05099466774198744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2560,12288,0.029009779294331867
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,16384,6144,0.031203554736243352
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,256,2048,0.006585777633719974
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,65536,2048,0.08566577566994561
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,256,512,0.0053280000057485365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,1024,7168,0.009351111120647853
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,768,128,0.0052888890107472735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3072,16384,0.02124711043304867
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,128,256,0.012049777640236748
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1024,3072,0.03443377878930833
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,16384,512,0.0041866666740841335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,64,10240,0.005970666805903117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,7168,768,0.005246222019195557
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,2560,3584,0.008008889026112026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,16384,3072,0.013168888787428537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,6144,6144,0.036440001593695745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3584,7168,0.012259555359681448
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,16384,12288,0.07966933647791545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,4096,3072,0.017277333471510146
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3072,3584,0.08342933654785156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,512,256,0.0022400000856982335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,128,256,0.005252444495757421
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,768,3584,0.00756888919406467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2048,65536,0.0452497767077552
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,7168,2560,0.010107555323176915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,5120,8192,0.01626755628320906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,2560,512,0.00526311124364535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,3072,2048,0.015384000208642749
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,2560,8192,0.03904799951447381
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,3072,8192,0.044863998889923096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,4096,1536,0.005945777727497949
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,128,3584,0.01147555559873581
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,16384,256,0.0028568889117903183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,10240,6144,0.20077244440714517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2560,128,0.0058755555914507965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,16384,3584,0.06395644611782499
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,256,8192,0.03316088848643833
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,12288,4096,0.026093333959579468
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2560,6144,0.010447999669445885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,4096,768,0.005576888720194499
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,5120,65536,0.21087822649214003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,6144,256,0.005568000177542369
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3584,256,0.0021537778278191886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1536,8192,0.019069333871205647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,1536,3584,0.0192239996459749
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,1536,7168,0.013034666577974955
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,65536,3584,0.05381155676311917
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,256,768,0.00591822216908137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,128,5120,0.015511110424995422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,8192,6144,0.08538666698667739
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,4096,512,0.003546666767862108
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,12288,4096,0.01443733274936676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,768,512,0.00666222224632899
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3584,3072,0.007168889045715332
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,10240,7168,0.033575110965304904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,7168,8192,0.03258488906754388
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,6144,10240,0.10297599765989517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1536,6144,0.012092444631788464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3072,65536,0.06669955783420138
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,10240,256,0.017557332913080852
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,7168,3584,0.03061422374513414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3584,2560,0.006404444575309753
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,2560,5120,0.019243554936514962
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,16384,64,0.017436444759368896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,64,128,0.002199999988079071
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,12288,1024,0.007810666329330868
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,768,3584,0.008047999607192146
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,256,1536,0.006241777704821692
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,10240,16384,1.2383129331800673
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,1536,6144,0.029545777373843726
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,256,4096,0.012423111332787408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,2560,8192,0.028744889630211726
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3584,4096,0.019741333193249173
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,128,8192,0.01865866614712609
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,256,128,0.005385777602593104
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,12288,256,0.007805333369308048
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,128,10240,0.01219911128282547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,1024,7168,0.021433777279324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,1536,2048,0.00703111125363244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,2048,512,0.006027555714050929
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,10240,6144,0.02208799951606327
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,3072,2048,0.018968888454967074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2560,32,0.005712000032265981
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,7168,10240,0.044600887431038745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3584,512,0.0025600000388092464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3584,8192,0.015592000550693937
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,10240,4096,0.026034666432274714
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,10240,65536,0.27983376714918345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,5120,1024,0.01779822177357144
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,4096,16384,0.02288177775012122
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,6144,7168,0.01555377741654714
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3072,32,0.007992888490358988
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,1024,6144,0.00902133352226681
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,1024,8192,0.020966221888860066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,1024,8192,0.011726222104496427
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,10240,1536,0.06535733408398099
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,64,64,0.0021048889805873237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,1024,4096,0.00812711069981257
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,32,7168,0.005172444300519096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,1024,8192,0.03148711058828566
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,128,3072,0.0047191112405723995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,3072,7168,0.011515555282433828
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,5120,768,0.007166222565703922
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,6144,128,0.005647111270162795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3072,7168,0.019910222954220243
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,8192,1024,0.0121760004096561
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,256,2048,0.006628444625271692
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,12288,6144,0.049673779143227466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,7168,8192,0.020065777831607394
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,7168,128,0.0022008888837363985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2560,12288,0.014306666122542487
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,5120,512,0.00830488900343577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,64,128,0.005551110953092575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,1024,4096,0.008021333151393467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3584,768,0.005559999909665849
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,6144,768,0.011108444796668159
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,256,32,0.00554666668176651
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,128,8192,0.010681777364677854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,2560,7168,0.07828888628217909
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,512,512,0.002528888897763358
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,4096,1024,0.005595555735958948
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,12288,8192,0.03949066665437486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,65536,5120,0.09148266580369736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,6144,512,0.002871111035346985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,7168,32,0.007839110990365347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,10240,10240,0.05849422348870171
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,10240,512,0.020304888486862183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,5120,65536,0.3126888804965549
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,10240,6144,0.03696889016363356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,16384,16384,0.09271644221411811
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,65536,4096,0.05229511194758945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,512,12288,0.007308444215191736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,2048,768,0.007261333366235097
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,12288,768,0.006592888798978593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,128,6144,0.008643555144468943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,65536,7168,0.1324497726228502
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,2560,1024,0.016516443755891588
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,768,12288,0.02881244487232632
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2048,2560,0.055120001236597695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,64,12288,0.005306666509972678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,65536,12288,0.13094133800930446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,10240,7168,0.023789332972632513
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,65536,6144,0.13192000654008654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,3072,768,0.010971555279360877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,3072,10240,0.01573333309756385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,6144,128,0.005601777798599667
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,256,5120,0.016215999921162922
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2048,7168,0.016552888684802584
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,16384,2048,0.041618665059407554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,4096,1536,0.004298666699065102
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,5120,1536,0.017272889614105225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3584,12288,0.01723911033736335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,5120,65536,0.1377991173002455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,12288,2048,0.01016088906261656
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,1536,2560,0.02717599935001797
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,65536,2048,0.044496887260013156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,12288,256,0.006266666783226862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,12288,7168,0.08699022399054633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,16384,8192,0.047641777329974704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,64,65536,0.029640889830059473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2560,8192,0.029040889607535467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,5120,256,0.0052533331844541765
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2560,256,0.0028817777832349143
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,12288,65536,0.22384354803297254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,1024,768,0.005624889085690181
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,32,4096,0.007672000262472365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,4096,1024,0.005720888988839255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,7168,7168,0.02053333322207133
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,12288,8192,0.03068888849682278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,5120,2048,0.007366221812036302
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,3072,1536,0.006248000181383557
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,5120,3584,0.01753422286775377
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,6144,128,0.019070222973823547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,4096,2048,0.021459556288189355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,1536,12288,0.05552355448404948
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,8192,3584,0.03285600079430474
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,10240,16384,1.3245342042711046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,1536,8192,0.03578222129080031
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,2560,2048,0.01870133313867781
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,2560,12288,0.01625688870747884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,128,3584,0.00815822184085846
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,512,6144,0.009107555780145857
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,512,4096,0.005709333138333426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,256,3072,0.007407111426194509
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,6144,2048,0.01458577811717987
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,5120,32,0.0021724444296624926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,256,65536,0.12745332717895508
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,512,2048,0.009195555415418413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,8192,3584,0.012238221863905588
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,65536,7168,0.0769253306918674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,3072,768,0.01668355531162686
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,128,2560,0.005060444275538127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,10240,3072,0.012224888636006249
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2560,3584,0.011729777687125735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,128,6144,0.016891555653678048
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,256,16384,0.01623288955953386
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,2048,7168,0.03267200125588311
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,8192,1536,0.007607999775144789
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,16384,5120,0.024262222978803847
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,768,16384,0.035382221142450966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1536,4096,0.010699555277824402
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,65536,7168,0.6066604190402561
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,64,6144,0.004567999806669024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,16384,12288,0.038233776887257896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,256,512,0.011896889242861005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,16384,4096,0.02463111115826501
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,6144,65536,0.07448710997899373
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,3584,16384,0.06795377863778009
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,256,2560,0.006715555571847492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,32,5120,0.008655110994974772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,8192,10240,0.038535111480289035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,7168,12288,0.02275733318593767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,1024,256,0.01512888901763492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,128,4096,0.009150221943855286
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,4096,6144,0.1716302235921224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,256,6144,0.009348444640636444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,1024,1536,0.006360888895061281
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,7168,3072,0.02713777787155575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,2048,2048,0.01551466683546702
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,256,5120,0.021695110532972548
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,5120,7168,0.015999999311235216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,768,512,0.006609777609507243
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,5120,2560,0.007924444145626491
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,1536,3584,0.007710222568776872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,8192,2560,0.008641777767075432
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,8192,10240,0.03007377849684821
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,128,128,0.011679111255539788
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,3584,5120,0.02090311050415039
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1536,7168,0.010468444062603844
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,128,64,0.008356444537639618
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,7168,12288,0.04342400034268697
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,768,2048,0.022478222846984863
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3072,2560,0.015544889701737298
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,7168,16384,0.03902133305867513
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,1536,512,0.013450666434235044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3584,3584,0.008630221916569604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,1024,5120,0.017063111066818237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,128,128,0.005223999834722943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,4096,7168,0.02404622236887614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,512,128,0.005391110976537068
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,4096,8192,0.016546666622161865
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,4096,5120,0.024681778417693243
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,3072,4096,0.02485688858562046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,65536,5120,0.056939555539025195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,768,5120,0.005586666779385672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,256,8192,0.018538667096032035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,65536,8192,0.6812088754442004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,16384,4096,0.2407617833879259
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,64,1024,0.004291555533806483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3072,3072,0.01145333300034205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,10240,6144,0.18464799722035727
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,7168,512,0.007308444215191736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3072,1536,0.008626666333940294
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,128,1024,0.003940444439649582
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,16384,256,0.006602666858169768
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2048,12288,0.018166222506099276
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,5120,5120,0.009124444590674506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,65536,768,0.06362755431069268
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,3584,512,0.015379556351237826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,6144,12288,0.0717866685655382
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2560,128,0.002562666725781229
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,65536,8192,0.09245689047707452
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,16384,6144,0.047520889176262744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,256,16384,0.0258559998538759
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,768,7168,0.019273777802785236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,65536,10240,0.10753955443700154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,12288,8192,0.03449155555831061
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,6144,128,0.005576888720194499
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,7168,12288,1.3580079608493383
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,5120,1024,0.025752888785468206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2048,128,0.003536888708670934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1536,10240,0.04086755712827047
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3072,128,0.003576000117593341
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1536,2048,0.015213333898120456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,6144,10240,0.027154665854242113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,12288,512,0.020625778370433383
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,5120,128,0.005912888795137405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,64,65536,0.0625128878487481
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,768,1536,0.011893333660231696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,5120,3584,0.03852355480194092
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1536,3072,0.007296889192528195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2560,65536,0.11104977793163723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,7168,1536,0.01182666669289271
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,256,128,0.002181333386235767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,16384,1024,0.02660622199376424
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,65536,2048,0.15631911489698622
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,64,512,0.002489777695801523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,65536,256,0.013352000051074557
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,512,64,0.005580444302823808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,128,5120,0.009315555294354757
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,12288,768,0.025973333252800837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,3072,16384,0.020831111404630873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1536,3584,0.0069013333155049225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,256,3072,0.025943999489148457
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2560,1024,0.013234666652149625
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,128,32,0.0069146665434042616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,2048,12288,0.060179558065202504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,2048,256,0.015814221567577787
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,4096,3072,0.01847111185391744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,3072,256,0.01685688893000285
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,4096,7168,0.01296533313062456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,10240,32,0.0031991110493739447
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3584,6144,0.06572177675035265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,2560,10240,0.012829333543777466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1024,12288,0.009756444229020013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,256,2560,0.005394666559166378
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1024,12288,0.016564443707466125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,5120,128,0.002263111165828175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,6144,512,0.008245333201355403
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1024,65536,0.028036445379257202
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,7168,1024,0.016799110505315993
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1024,7168,0.07498577568266127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,128,1536,0.012073777616024017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3584,512,0.0031795555518733132
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,2560,1024,0.006028444402747684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,16384,768,0.0221733334991667
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,3072,10240,0.014356444279352823
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,768,12288,0.017094223035706412
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,512,2560,0.005254222287072076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,3072,128,0.014491554763582019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,1024,3072,0.0076142218377855085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,768,12288,0.013076444466908773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,16384,3584,0.02012977831893497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,64,768,0.005968889014588461
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2560,3072,0.00721688899728987
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,1024,16384,0.018851555056042142
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,768,7168,0.0069226668112807805
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,128,2048,0.013759111364682516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,4096,6144,0.012243555651770698
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,12288,2560,0.011178666518794166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,6144,512,0.015507555670208402
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,1536,8192,0.01608622239695655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,16384,2048,0.0287715560860104
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,8192,4096,0.010372444159454769
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3584,2048,0.005894222193294101
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,16384,512,0.007265778051482306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2560,16384,0.020456888609462313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,65536,256,0.012416000167528788
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1536,1024,0.003167111012670729
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,3584,256,0.0052666668262746595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,512,5120,0.008289777570300633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,4096,2560,0.021282666259341772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,7168,6144,0.018248000078731112
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,768,1024,0.005564444594913059
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,1536,4096,0.007931555310885111
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,16384,64,0.009356444080670675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2560,10240,0.01458666721979777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,512,3072,0.0069679998689227635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,12288,128,0.0021457777669032416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2048,1024,0.005150222116046482
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,5120,128,0.0069155556460221606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,7168,7168,0.08396622207429673
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1536,1024,0.005954666684071223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,16384,3072,0.017643555998802185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,3072,3072,0.007613333563009898
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1536,3072,0.00554666668176651
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,128,6144,0.05006133185492622
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,2560,5120,0.02308711078431871
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,512,7168,0.010483555495738983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,5120,256,0.006572444405820634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,768,768,0.0076657773719893555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,128,768,0.006269333263238271
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,12288,128,0.003077333379122946
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,12288,64,0.002550222186578645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,256,5120,0.02700800034734938
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,12288,128,0.0022133334229389825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,3072,512,0.014182221558358936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,512,3584,0.006918222126033571
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,65536,5120,0.05875644418928358
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,2048,128,0.005559111220969095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3584,12288,0.018066666192478605
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,16384,3072,0.014856000741322836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,7168,2560,0.007387555307812161
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,256,2560,0.0069102222720781965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1024,65536,0.030156443516413372
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,65536,4096,0.11300978395673965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,32,12288,0.007333333293596904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,16384,3584,0.033097777101728655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3072,65536,0.25101155704922146
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3072,8192,0.01181333346499337
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,3072,2560,0.03945422172546387
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,32,3072,0.00831911133395301
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,16384,512,0.015527110960748462
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,512,768,0.005953777581453323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,7168,256,0.005603555589914322
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,512,16384,0.01724177764521705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2560,8192,0.01310755560795466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,256,3072,0.004875555634498596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,65536,1536,0.030664887693193223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,768,128,0.011065777805116443
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,5120,5120,0.06455822123421563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,768,10240,0.012439111040698158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,16384,2048,0.3027617666456434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,12288,5120,0.19262666172451445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,12288,64,0.010716444088353051
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,5120,512,0.005975111077229182
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,8192,16384,0.046943998999065824
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,6144,5120,0.03106133474244012
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,1024,8192,0.010695999695195092
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,65536,32,0.007319110963079665
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,7168,10240,0.017873777283562552
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,64,1536,0.004256000121434529
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,65536,5120,1.5032791561550563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,16384,256,0.01146133326821857
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,768,5120,0.040001776483323835
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,1536,2048,0.036185777849621244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3072,2048,0.011783999701340994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,16384,7168,0.02629511058330536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,7168,32,0.0019253332995706135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,768,768,0.005650666852792104
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,5120,65536,0.10714844200346206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,16384,12288,0.03650133477316962
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,7168,2048,0.008551110823949179
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,256,256,0.002504000026318762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,512,128,0.005255110975768831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,768,1536,0.003924444317817688
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,16384,3072,0.08516177866193984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,16384,32,0.02474400069978502
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,5120,1536,0.006513777706358168
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,256,256,0.005626666463083691
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,4096,512,0.014882667197121514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,5120,2048,0.017581333716710407
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,5120,768,0.005991111199061076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,512,1024,0.0032115555885765287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,8192,8192,0.030243555704752605
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,4096,2048,0.005538666827811136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,256,768,0.006262222097979651
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,2048,1024,0.018772444791264005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,8192,5120,0.025620443953408137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,128,3584,0.005664000080691443
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,10240,5120,0.02083288960986667
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,8192,12288,0.03670755690998501
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,2560,5120,0.008653333617581261
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,32,1536,0.0063484443558586975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,7168,64,0.004565333326657613
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,1536,512,0.005591111050711737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3584,1024,0.009745777481132084
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,3584,3584,0.007420444654093848
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,12288,8192,0.05420533484882779
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,12288,12288,0.036630223194758095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,65536,16384,0.3055608802371555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,5120,65536,0.1936195558971829
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,4096,7168,0.2751413451300727
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,10240,3584,0.01721333298418257
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1536,1024,0.025097777446111042
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,3072,2560,0.007626666790909237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,6144,5120,0.03894666830698649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,5120,4096,0.029127111037572224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,3072,8192,0.052354666921827525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,6144,1536,0.015412444869677225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,12288,128,0.01726222203837501
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,16384,3072,0.03142755561404758
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1024,3584,0.007637333538797166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,6144,1536,0.006656888872385025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,512,512,0.002550222186578645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,2048,4096,0.01148088855875863
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3072,128,0.002504000026318762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,10240,12288,0.043848001294665866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,32,65536,0.007623111208279927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,512,4096,0.008360000120268928
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,8192,256,0.005931555396980709
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,10240,1024,0.0199297782447603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,10240,128,0.003535110917356279
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,768,10240,0.00979911122057173
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,1536,768,0.005995555470387141
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3584,2048,0.007368889119890001
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,12288,65536,0.22632000181410047
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,1536,512,0.00565244464410676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,2560,128,0.005583999885453119
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,65536,12288,0.2606666617923313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,32,65536,0.059664891825781934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,128,12288,0.014303111367755465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,32,3584,0.036313778824276395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,768,12288,0.013935110635227628
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,8192,6144,0.09829066859351264
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,16384,2560,0.0428071121374766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,12288,4096,0.020629333125220407
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,7168,10240,0.03891288903024461
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,5120,512,0.019267555740144517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,10240,2048,0.025881777207056682
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,12288,7168,0.059880885812971324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,7168,128,0.009152000149091085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,4096,1536,0.008536888493431939
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,8192,1024,0.0849262211057875
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,12288,32,0.007747555772463481
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,65536,2560,3.0473367902967663
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,256,2560,0.005060444275538127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,12288,3072,0.05156888895564609
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,1024,3584,0.007679999702506595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,10240,64,0.006629333313968446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,16384,256,0.0062871111763848206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,2048,10240,0.012159999873903064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,512,5120,0.02060622142420875
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,65536,2048,0.02684088879161411
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1536,1024,0.005616000129116907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3072,128,0.002789333255754577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,256,3072,0.010812444819344414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2560,128,0.007598222129874759
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,12288,12288,0.06749866406122844
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,3072,4096,0.007874666816658443
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,768,1536,0.006649777707126405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,10240,7168,0.02347822156217363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,7168,6144,0.012614222036467658
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,12288,1536,0.025088888075616624
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,2560,1536,0.006257777826653586
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3072,3072,0.012916444076432122
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3072,3584,0.00738844441043006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2048,5120,0.008961777720186446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,3072,5120,0.01145155562294854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,32,4096,0.007762666377756331
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1024,2048,0.009653333160612319
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,2048,1024,0.016864000095261466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,2048,3584,0.007813333637184566
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,768,768,0.005706666658322017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,4096,3072,0.010019555687904358
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,12288,65536,0.1531120008892483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1024,1536,0.006474666711356904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,5120,2560,0.006983111302057902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,65536,768,0.41826844215393066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,128,2560,0.007090667055712805
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,2048,512,0.006261333409282897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,4096,1024,0.016711110870043438
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,12288,10240,0.03102133340305752
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,65536,1536,0.04871911141607496
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,12288,16384,0.058055109447903104
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,8192,65536,0.09120088815689087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,768,65536,0.06468088759316339
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2560,256,0.0039048890272776284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,1536,1536,0.017847110827763874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,16384,7168,0.02603911028967963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,2048,1536,0.01650399963061015
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,2560,3584,0.010746666954623329
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,6144,8192,0.034825778669781156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,12288,4096,0.03567022085189819
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2560,4096,0.00831200016869439
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,8192,12288,0.024746666351954143
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3072,8192,0.01517599986659156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,128,2048,0.006592000110281839
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,2560,4096,0.029476443926493328
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,3072,65536,0.15190311272939047
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,7168,64,0.002579555536309878
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,32,10240,0.006042666733264923
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,16384,256,0.00628977765639623
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,8192,5120,0.012108444339699216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,128,32,0.005965333431959152
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,6144,64,0.005351999981535806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,12288,768,0.005225777626037598
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,768,128,0.007903999752468532
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,2048,10240,0.016307555966907077
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,16384,8192,0.031481779283947416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,65536,4096,0.038513776328828596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,256,128,0.012086222569147745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,16384,4096,0.10751733514997695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,65536,16384,0.1636364459991455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,3072,3072,0.00756533361143536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,256,64,0.0019155556543005838
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,4096,2048,0.006659555352396435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1536,16384,0.0192248887485928
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,3584,768,0.030615998639000788
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2048,2560,0.00702133360836241
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,3584,3072,0.019662222928471036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,768,2048,0.004313777718279096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,7168,64,0.005569777968857024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,7168,3584,0.050478223297331065
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,128,16384,0.017667555146747164
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,8192,2048,0.025913778278562758
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,12288,512,0.020795555578337777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,65536,6144,0.06747821966807048
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,65536,2560,0.16626755396525064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,5120,2048,0.006983111302057902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,256,1536,0.017871111631393433
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,65536,1536,0.019274666905403137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2560,16384,0.017657778329319425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,7168,256,0.002218666589922375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,4096,1024,0.0032275555034478507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,8192,4096,0.06110666857825386
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2560,3584,0.00794222205877304
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,32,10240,0.01959911154376136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,2048,1024,0.007296889192528195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3072,5120,0.009321778184837764
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,8192,6144,0.013468444347381592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,512,128,0.0021502222451898786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,4096,1024,0.017685333887736004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,768,512,0.006248888870080312
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,3072,12288,0.016551110479566786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,1536,3584,0.01588711142539978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,4096,8192,0.3059111171298557
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3584,1536,0.008463111188676622
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,768,6144,0.01715555621518029
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,64,3072,0.004644444419278039
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,256,65536,0.11053066783481175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,768,2560,0.023223999473783705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,65536,128,0.010775110787815519
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2560,64,0.0032239999208185407
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,12288,2560,0.03188622328970168
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,16384,1536,0.009727111293209923
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,16384,1536,0.010462221999963125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,6144,3584,0.007992888490358988
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2560,32,0.0058471109304163195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,768,8192,0.006919999917348226
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,128,512,0.005435555759403441
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,128,7168,0.04145866632461548
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3072,3584,0.01016177816523446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,2048,2560,0.01536799967288971
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,3072,1536,0.006248000181383557
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,6144,32,0.0059075554211934405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,4096,10240,0.017236444685194228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,6144,2048,0.009749333063761393
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,2560,2048,0.06739911105897692
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,12288,5120,0.03222666515244378
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,128,256,0.0021866667601797315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,12288,5120,0.018135999639829
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,65536,1024,0.022106667359670002
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,3072,6144,0.027738667196697656
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,5120,65536,0.1475813388824463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,512,128,0.005924444645643234
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2048,10240,0.02680622206793891
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,2560,8192,0.04208977686034309
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,32,3584,0.004566222015354368
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,7168,768,0.005955555372767978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,7168,5120,0.017435555656750996
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,65536,3072,0.03674844569630093
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,6144,1536,0.020594666401545208
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,6144,768,0.005604444278611078
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,768,256,0.01480888823668162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3584,128,0.0035013332962989807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,1536,65536,0.07908799913194445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,12288,64,0.018237334158685472
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,8192,16384,0.02718488872051239
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,12288,1024,0.007009777757856581
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,10240,768,0.0069475554757648045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,1024,4096,0.018613333503405254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,128,2048,0.006589333216349284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3584,8192,0.023551111419995625
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,128,512,0.006973333656787872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,16384,2560,0.029900444878472224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,3584,12288,0.03018222252527873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,7168,4096,0.03290311164326138
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,768,5120,0.011345778074529437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,64,3072,0.004864888886610667
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,16384,16384,0.6481955316331651
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,1024,12288,0.013402666482660504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,6144,3584,0.025408888856569927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,1536,7168,0.03975111246109009
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1536,512,0.003177777760558658
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,8192,12288,0.020419556233617995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,768,12288,0.01498044364982181
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,128,128,0.005331555588377847
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,2048,12288,0.026929777529504564
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,512,128,0.012720888687504662
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,5120,7168,0.013799111048380533
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3584,256,0.0022239999638663395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,3072,1536,0.01581333412064446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3072,1024,0.03219822380277846
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,16384,128,0.006578666468461354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3072,512,0.004146666576464971
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,64,7168,0.005618666609128316
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,32,2560,0.004537777768241035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,7168,512,0.0059279998143513995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,2560,65536,0.0841582218805949
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,65536,64,0.016144000821643405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,128,2560,0.005342222336265776
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,65536,12288,0.13058311409420437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,8192,65536,1.2191990746392143
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,10240,16384,0.04389422138532003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,512,3072,0.01924888955222236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,2048,16384,0.020035554965337116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,65536,6144,0.11155022515190972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,5120,12288,0.0192266669538286
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,12288,4096,0.017242666747834947
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,7168,2048,0.00777422222826216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,1024,16384,0.04539555642339918
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,4096,2048,0.005255110975768831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,128,4096,0.015431111057599386
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,12288,12288,0.08895111083984375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,768,4096,0.00629333323902554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,6144,2560,0.007698666718271043
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,10240,7168,0.02255200015174018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,7168,2560,0.008434666527642144
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,4096,65536,1.4040106667412653
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,16384,512,0.022668444448047217
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,512,128,0.006000888844331105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,7168,1024,0.005640888793600931
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,768,32,0.007650666766696506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,1024,2048,0.007365333537260692
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1024,128,0.002263111165828175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,7168,2048,0.019725332657496136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,4096,512,0.01518133282661438
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,7168,65536,2.810021294487847
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,6144,128,0.003943110919660992
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,128,6144,0.006562666760550604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,65536,1024,0.05290666553709242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,8192,5120,0.012512000070677863
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2048,768,0.0042275554604000514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1536,1536,0.0039022221333450745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,2048,2048,0.025394666526052687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,65536,128,0.054479112227757774
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,12288,1024,0.009401777552233802
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1024,1536,0.00796088907453749
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2048,1536,0.008046222229798635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,512,1024,0.01477688882086012
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,5120,32,0.007993777592976889
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,7168,65536,0.13562844859229192
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3072,512,0.008355555435021719
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,1024,512,0.013120888835854001
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,32,1024,0.005954666684071223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,5120,256,0.0166631109184689
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,1024,8192,0.05634044276343452
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,7168,7168,0.01609688831700219
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2048,12288,0.2167724503411187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,512,768,0.005922666854328579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,5120,4096,0.007995555798212687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,256,3584,0.005706666658322017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,7168,12288,0.06342222293217976
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,4096,256,0.016923555069499545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,16384,512,0.0070266665683852295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,4096,256,0.00254844439526399
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,256,3072,0.021727111604478624
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3584,1536,0.006028444402747684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,768,256,0.014891554911931356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2560,5120,0.008364444805516137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,32,3584,0.004609777695602841
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,12288,5120,0.020768000019921195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,256,1024,0.004643555730581284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,1536,1024,0.01592088904645708
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,2048,16384,0.38496889008416074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,16384,1536,0.013080000049538083
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,2560,3072,0.0070568886068132185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,8192,2048,0.007274666594134436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,6144,3584,0.09158755673302545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,4096,32,0.01423377792040507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3072,768,0.01573244399494595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,768,32,0.00217333332531982
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,6144,16384,0.06667733192443848
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,65536,12288,0.2633964485592312
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,128,2560,0.004951999833186467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1024,3584,0.02250755495495266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,8192,12288,0.03168088859981961
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,128,8192,0.005611555443869696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,16384,10240,0.05309777789645725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2048,3072,0.006709333509206772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,4096,512,0.003621333175235324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,256,4096,0.007677333222495184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,16384,8192,0.0959048867225647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,16384,65536,0.34986223114861387
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,2560,16384,0.026403556267420452
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,32,512,0.002564444517095884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,65536,10240,0.20089688565995958
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,7168,6144,0.15619199805789524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,65536,768,0.13503021664089626
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,10240,5120,0.01645599967903561
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,3584,768,0.011110222174061669
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,65536,5120,0.1323715580834283
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3584,6144,0.012631110846996307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,8192,128,0.03752177622583177
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2048,768,0.0039022221333450745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,128,512,0.006969778074158563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,7168,10240,0.016706667012638517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,7168,5120,0.013797333670987023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,6144,3584,0.009383111364311641
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,64,768,0.005697777701748743
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,7168,7168,0.02188355558448368
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,3072,16384,0.020750222934616935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,12288,5120,0.021809778279728357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2048,768,0.0251884443892373
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,32,256,0.002187555655837059
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,768,64,0.002032888846264945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,6144,2048,0.011648000114493899
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,16384,768,0.024469332562552557
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,7168,1536,0.006627555522653792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,256,4096,0.0052968888647026485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3072,5120,0.009773333039548662
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,12288,5120,0.03877955675125122
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,7168,768,0.0169831116994222
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,16384,1024,0.014195554786258273
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,65536,12288,3.2441323598225913
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,4096,6144,0.015876443849669564
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,2560,6144,0.00904799997806549
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,10240,2048,0.01660977800687154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,3072,128,0.014480888843536377
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1024,6144,0.010199999643696679
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,10240,7168,0.0473671125041114
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,512,2560,0.017045333981513977
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,5120,10240,0.02310400042268965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,12288,4096,0.01425244410832723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,2048,512,0.01686044368478987
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,7168,3584,0.008935111264387766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,128,3072,0.007301333049933116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,16384,16384,0.07276444302664863
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,4096,1536,0.006235555642180973
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,32,65536,0.04562044474813673
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,256,16384,0.07765155368381076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3584,10240,0.10582666926913792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3584,8192,0.014480888843536377
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,3072,512,0.005596444424655702
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3584,512,0.006262222097979651
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,6144,6144,0.025270221961869135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,64,5120,0.004531555705600315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,4096,2560,0.008618666893906063
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,768,2560,0.006990222467316522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,6144,512,0.01594577729701996
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,32,16384,0.01584533353646596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,3584,3072,0.039150221480263606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,6144,2560,0.021881777379247878
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,4096,1536,0.0086666668454806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,7168,16384,0.05234044459131029
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2048,2560,0.005286222116814719
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,5120,5120,0.03166488806406657
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,10240,6144,0.0978435542848375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,3072,128,0.005633777628342311
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,8192,1536,0.0299288895395067
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,32,65536,0.05564000209172567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,768,3072,0.016198222835858662
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,64,32,0.0018915554715527429
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1024,32,0.005600000007285013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,10240,16384,0.10628622108035618
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,3072,2048,0.020534222324689228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,768,1024,0.007848000360859765
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,2560,4096,0.010395555860466426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,7168,3584,0.010048000348938836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3072,6144,0.016726222303178575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1536,1536,0.008011555506123437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,16384,7168,0.043733331892225474
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,12288,3584,0.03791022300720215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,5120,16384,0.03138666682773166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,4096,12288,0.020273778173658583
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,4096,16384,0.024381332927280005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3584,16384,0.03273066547181871
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,4096,2560,0.007302222152551015
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,6144,64,0.0035502223504914176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,64,32,0.006223111102978389
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,256,65536,0.11746933725145127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,1536,2560,0.01346933344999949
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2560,5120,0.010556444525718689
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,128,8192,0.005693333430422678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,5120,12288,0.0237137774626414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2048,2048,0.006986666884687211
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,12288,3584,0.014192000031471252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,3072,1024,0.030563556485705908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,16384,2560,0.016584000653690763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,768,128,0.014127999544143677
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,16384,16384,0.07364000214470758
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1024,1024,0.009696889254781935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,5120,16384,0.03528177738189697
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,64,5120,0.026850667264726426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,256,128,0.005776888794369168
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1024,512,0.008985777695973715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,8192,1536,0.024480889240900677
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,3584,7168,0.03942044576009115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2048,2048,0.005070222334729301
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,5120,12288,0.024099555280473497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,6144,5120,0.010759111079904767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,2048,128,0.005466666486528184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2048,16384,0.015282667345470853
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,65536,768,0.015617777903874716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,1024,256,0.015497777197096082
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,128,65536,0.5149377716912164
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1024,1536,0.00425155543618732
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,3584,8192,0.014271999398867289
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,128,5120,0.008805333740181392
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,2560,3072,0.007304000357786815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,8192,2560,0.07306844658321805
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,16384,2560,0.01347644461525811
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,768,6144,0.02148088812828064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,256,2048,0.01593777702914344
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,256,16384,0.028399109840393066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,12288,2048,0.047784000635147095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,2560,256,0.0052471111218134565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,10240,1024,0.03016177813212077
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,1024,3072,0.007335999773608313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,12288,5120,0.02139644490347968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,5120,512,0.015970667203267414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1024,10240,0.010791111323568555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,10240,1024,0.0063324446479479475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,3072,128,0.014915555715560913
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,7168,4096,0.012483555409643384
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2560,4096,0.008340444829728868
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,64,1024,0.0032035555276605818
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,12288,65536,0.16105688942803278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,10240,128,0.0061919999619325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,65536,3072,0.03596533338228861
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,16384,128,0.006649777707126405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,16384,256,0.0035457776652442086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,8192,2048,0.0086986662613021
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,3584,512,0.019296889503796894
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,4096,65536,0.1324275599585639
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,12288,1536,0.13681156105465359
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,16384,1024,0.025979555315441553
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,65536,2048,0.43495999442206484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,6144,16384,0.03032533327738444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,65536,3584,0.13403733571370444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,128,128,0.005676444619894028
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,65536,16384,0.3455048931969537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,2560,3072,0.007336888876226213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,256,256,0.01311288856797748
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,512,3072,0.014856000741322836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,12288,12288,0.037435554795795016
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,12288,4096,0.030961778428819444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,3072,256,0.005623111294375525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3072,16384,0.024136000209384497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,64,512,0.002579555536309878
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,768,12288,0.012659555508030785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,12288,32,0.002537777854336633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,8192,3072,0.01702933344576094
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,16384,5120,0.020995555652512446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3072,32,0.002250666626625591
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,64,64,0.0021555556191338436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,128,10240,0.08548888895246719
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,2560,2048,0.019346666004922655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,10240,6144,0.5844257672627767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,5120,12288,0.030299554268519085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,6144,65536,0.19470755259195963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,768,4096,0.02854666776127285
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3584,5120,0.009680888719028896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,12288,8192,0.039466665850745306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,6144,2560,0.007096889118353526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,512,256,0.005608888963858287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,768,1536,0.008350222474998897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1024,768,0.00562755556570159
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,3072,2560,0.016977777083714802
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,64,3072,0.007332444190979004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,4096,7168,0.11014222436481053
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,8192,7168,0.02461333407296075
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,6144,16384,0.023585778143670823
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,2560,6144,0.022146667043368023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,3584,3072,0.023337778117921617
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2560,3584,0.012440889245933957
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,65536,16384,0.28311999638875324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,16384,2048,0.15219644705454508
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,768,512,0.005573333137565189
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,7168,1536,0.007423111134105259
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,4096,64,0.005909333212508096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,2048,768,0.016151999433835346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,1024,10240,0.02309866746266683
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2560,64,0.006931555353932911
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,128,5120,0.008653333617581261
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,5120,6144,0.032772445016437106
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,2560,16384,0.02288444505797492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,768,16384,0.017512889371977914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,3072,3072,0.007585778004593319
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,3584,3072,0.023648000425762598
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,2560,256,0.015116444892353482
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,128,256,0.015838222371207345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2560,512,0.0032088889016045462
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,4096,2048,0.006617777877383762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,10240,4096,0.016538666354285348
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2048,3072,0.005929777605666055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,3584,65536,0.08873244126637776
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,6144,3072,0.02211911148495144
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1024,3072,0.0388648874229855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2560,7168,0.012699555191728802
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,1536,3072,0.022447999980714586
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,16384,3072,0.07069333394368489
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,512,128,0.0021484444538752237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3584,12288,0.02071288890308804
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,1536,7168,0.0123857781291008
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,128,256,0.008303110798199972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,65536,8192,0.33580711152818465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,7168,16384,1.4446755515204535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,1536,3072,0.03828088773621453
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2560,768,0.006907555378145642
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,128,16384,0.08492177724838257
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,16384,12288,1.2832933002048068
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,4096,3584,0.012508444488048553
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,10240,3584,0.016536888149049547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,1536,2048,0.02149244480662876
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,5120,7168,0.04252533449067009
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,512,12288,0.02346222268210517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,512,10240,0.006445333361625671
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,768,12288,0.012596444123321109
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,2560,10240,0.035756445593304105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,512,512,0.012088889049159156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,8192,3072,0.013332444760534497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,8192,3584,0.029506666792763606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,3072,1536,0.015807999504937064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,4096,4096,0.03300977746645609
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,2048,8192,0.01076533314254549
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,2048,65536,0.06349777513080172
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,256,3072,0.007288888924651676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,12288,768,0.05161066518889534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,5120,2560,0.021338666478792827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3584,7168,0.012809777425395118
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,768,1024,0.013413333230548434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,10240,7168,0.02494844463136461
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2560,2048,0.028055111567179363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,256,16384,0.015840000576443143
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,12288,10240,0.10223733054267035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2560,512,0.0026808888134029177
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,512,8192,0.010665777656767102
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,5120,768,0.018730666902330186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1024,2048,0.0046364445653226645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,6144,8192,0.18174310525258383
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,768,128,0.014103111293580798
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,3072,2048,0.006579555571079254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,256,7168,0.007619555625650618
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2048,16384,0.317941321267022
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,512,65536,0.05522933271196154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,4096,4096,0.047835555341508657
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,10240,12288,0.03320444292492337
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,65536,7168,0.12890488571590847
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3072,12288,0.02718399961789449
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,256,512,0.002491555487116178
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,32,128,0.005606222069925732
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,5120,10240,0.031231999397277832
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,2048,65536,0.1816880040698581
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2048,5120,0.009305777649084726
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,4096,4096,0.016583111551072862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,6144,8192,0.022699554761250813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,768,7168,0.010124444133705562
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,512,128,0.0052284445199701525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,65536,768,0.03124533428086175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,2560,128,0.014478221535682678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1024,7168,0.00907911111911138
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,12288,16384,0.0827546649509006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,10240,5120,0.018604444132910836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,5120,10240,0.06825955708821614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,1536,5120,0.028903110159768
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,5120,5120,0.01242844429281023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,65536,64,0.010021333065297868
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,3072,65536,0.09084444575839573
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,8192,1536,0.05318311187956068
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,4096,64,0.011583111352390714
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,6144,256,0.002480888946188821
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,10240,4096,0.01722577710946401
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,7168,8192,0.18129689163631865
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,3584,2560,0.006933333145247565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2048,256,0.005622222191757626
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,2560,3584,0.007369777394665613
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,2048,65536,0.17681777477264404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3584,4096,0.009689778089523315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,10240,512,0.006612444503439798
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,3072,512,0.005604444278611078
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,7168,768,0.01789066692193349
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,128,65536,0.013995556367768182
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2560,3072,0.007263110743628607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,12288,2560,0.05844533443450928
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,10240,768,0.04196622305446201
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,256,768,0.0063893331421746155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,6144,768,0.002875555513633622
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,64,16384,0.015909334023793537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,512,12288,0.01313333296113544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,6144,8192,0.01687111126052009
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,5120,16384,0.03459555572933621
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,5120,12288,0.20053954919179282
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,10240,3584,0.013988445202509562
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,64,65536,0.008384889198674096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,3584,12288,0.05676888757281833
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,16384,256,0.0025439999169773525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,128,768,0.0029004443850782183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,32,1536,0.003559111307064692
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,65536,7168,0.13060088952382407
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,2560,3584,0.022443556123309668
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1536,256,0.006481777876615524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,32,3584,0.004946666873163647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,10240,1536,0.014350222216712104
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2560,65536,0.06227022409439087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,10240,7168,0.052341332038243614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3584,3584,0.00887288898229599
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,256,32,0.0018613334331247543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,6144,32,0.007899555895063613
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,1024,7168,0.02836266822285122
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,1536,1024,0.015164444843928019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,65536,10240,0.1351813342836168
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2048,5120,0.010473777850468954
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,65536,7168,0.14541955788930258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,6144,8192,0.014448889427714877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,12288,4096,0.01882222294807434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,4096,3072,0.02538577715555827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,12288,2560,0.021388444635603163
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,4096,10240,0.016864000095261466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,5120,3072,0.02053955528471205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,1536,3072,0.007689778175618913
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,12288,32,0.0022400000856982335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,1024,7168,0.018625777628686693
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,7168,7168,0.016570667425791424
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,2560,128,0.0053137776752312975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,256,3072,0.005285333428117964
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,65536,16384,0.13272799385918513
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,256,5120,0.01035911093155543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,1536,3072,0.023967999551031325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,10240,16384,0.06726933187908597
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,8192,512,0.02565333412753211
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,512,5120,0.031522668070263334
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,4096,4096,0.019711999429596793
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,5120,4096,0.014655111564530266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,5120,1024,0.03727911247147454
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,65536,2048,0.02734400000837114
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,6144,256,0.03677333394686381
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3072,65536,0.06380177868737115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,768,2560,0.006931555353932911
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,10240,6144,0.017449777987268236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2048,32,0.0022417778770128884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2048,3584,0.007648000286685095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3072,512,0.005179555465777715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,32,3072,0.004559999952713649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3584,1024,0.004181333300140169
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2048,4096,0.009036444127559662
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,2048,256,0.014939554863505893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,7168,1536,0.04859377940495809
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,5120,2560,0.04211289021703932
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,2048,2048,0.00639111093348927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,128,3584,0.007987555530336168
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3584,1536,0.007982222570313348
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,5120,64,0.0028737777223189673
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,1024,12288,0.07076355483796862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,8192,16384,0.04435288906097412
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,512,4096,0.00832266691658232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3584,12288,0.16548532909817165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,10240,512,0.006575999988449945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,10240,6144,0.024977778395016987
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,8192,1536,0.007345777418878343
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,16384,4096,0.036835554573271014
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,10240,1536,0.008634666601816813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,6144,8192,0.08810133404201931
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,128,768,0.006446222050322427
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,32,2048,0.004250666747490565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,128,12288,0.02628266645802392
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,512,256,0.0023128889087173673
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,7168,10240,0.01977333260907067
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,8192,1024,0.00517866677708096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,12288,4096,0.16344977749718562
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,8192,32,0.041972445117102734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,6144,5120,0.013792888985739814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,256,512,0.011399110986126794
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,65536,3072,0.10017066531711155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,4096,768,0.01628444426589542
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,3584,16384,0.020824000239372253
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,2048,5120,0.00832088871134652
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1024,8192,0.008656889200210571
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,128,768,0.005380444642570283
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3584,1024,0.006259555617968242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,4096,2048,0.006595555692911148
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,5120,65536,0.11239022678799099
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,5120,6144,0.011129777464601727
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,12288,256,0.0031831111345026228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,64,8192,0.010690666735172272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,512,128,0.01310222182008955
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,4096,128,0.0021679999513758556
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,7168,3072,0.010450666977299584
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,7168,512,0.0035084444615576
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,6144,8192,0.022416889667510986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,6144,1024,0.00461422238085005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2560,65536,1.5892053180270727
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,1024,2560,0.03831200136078729
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,2048,512,0.014513777361975776
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,1536,3584,0.009132444030708736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,16384,1536,0.01352355546421475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,256,6144,0.010804444551467896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,8192,16384,0.027796443965699937
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,256,10240,0.011720889144473605
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,768,128,0.005939555664857228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,1536,8192,0.010382222632567087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,256,6144,0.009398221969604492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,64,12288,0.005215999980767568
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,16384,3072,0.01928622192806668
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,3072,3584,0.018927999668651156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,4096,7168,0.02369244396686554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,256,5120,0.015503111812803479
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,16384,8192,0.2010586659113566
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,5120,7168,0.012115555504957834
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,16384,3072,0.013296888934241401
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,768,2560,0.03607111175855001
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3584,65536,0.2134924464755588
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,16384,7168,0.05162400007247925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,16384,7168,0.03650488787227207
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,2048,7168,0.009308444129096137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,4096,1536,0.006625777731339137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,12288,16384,0.16454399956597224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,2048,5120,0.008655110994974772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,6144,3072,0.007736889024575551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,256,6144,0.017208000024159748
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,5120,6144,0.03341599967744615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,32,12288,0.013148444394270578
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,65536,6144,0.14450489150153265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,5120,7168,0.04777244395679898
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,4096,2048,0.007081777685218387
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3072,65536,0.06621599859661527
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,10240,3072,0.29009511735704213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3072,768,0.003014222201373842
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,256,6144,0.009104889300134447
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,4096,16384,0.3172364499833849
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,128,2560,0.007714666426181793
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2048,128,0.002924444360865487
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,10240,256,0.00646311127477222
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1024,2560,0.014553777045673795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,128,8192,0.010696888797812991
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,4096,3072,0.007258666886223688
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,1536,8192,0.020632000433074105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,512,256,0.014497778481907315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,512,512,0.0025519999778932994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,512,8192,0.01848088867134518
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,4096,16384,0.025283555189768474
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,65536,128,0.01925155520439148
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,12288,6144,0.02083555526203579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,4096,5120,0.012136889000733694
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,12288,6144,0.18129866653018525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,12288,512,0.00664533343580034
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3072,10240,0.013472889032628802
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,64,1024,0.005991111199061076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,512,2048,0.012478222449620565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,2048,12288,0.01440000037352244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1536,512,0.006610666712125142
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,3584,16384,0.2527093357510037
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,768,1024,0.0032275555034478507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,16384,12288,0.10841066969765557
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,12288,65536,1.5634533564249675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,8192,6144,0.015383111106024848
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1536,65536,0.08582488695780437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,1024,10240,0.02455111179086897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,1024,1024,0.005600888695981767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,65536,768,0.02460533380508423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,2048,4096,0.008038221961922115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,8192,64,0.005773333211739858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,10240,256,0.05247822072770861
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,3584,4096,0.02139555580086178
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,10240,5120,0.3227857748667399
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,128,2560,0.004942222187916438
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2048,512,0.0028408887899584244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,8192,16384,0.044129778941472374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,3584,512,0.015645333462291293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2560,64,0.01185066666867998
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,2048,1536,0.007272889216740926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,768,5120,0.017883555756674874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,5120,3584,0.021408889028761122
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,64,16384,0.02535999980237749
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,4096,7168,0.013693333499961428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,12288,6144,0.018397332893477548
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,4096,3072,0.007864889171388414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,768,16384,0.03668355610635545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,768,64,0.002234666711754269
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,3584,5120,0.008355555435021719
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,6144,3072,0.0074026667409473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,256,4096,0.005325333525737126
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,7168,10240,0.03818044397566054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,512,7168,0.009563555320103964
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,8192,8192,0.06506311231189303
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,16384,7168,0.33838756879170734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,5120,512,0.008307555483447181
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,12288,5120,0.01682844426896837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2048,8192,0.013455111119482251
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,2048,5120,0.020230222079488967
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,1536,512,0.020971554848882887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,12288,2048,0.03293955657217238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,5120,1536,0.008106666306654612
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,4096,10240,0.017303110824690927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,128,7168,0.03471022182040744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,7168,512,0.005610666755172942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,64,8192,0.005630222045713001
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,128,8192,0.005636444522274866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,2048,7168,0.013787555197874704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,7168,10240,0.08739378054936726
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,6144,7168,0.014856000741322836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,16384,5120,0.021393777595625982
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1024,16384,0.04754399922158983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,10240,1536,0.011216888825098673
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,12288,128,0.005917333480384614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,768,7168,0.010738666686746808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,12288,16384,0.04340799980693393
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,65536,6144,3.806145985921224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,5120,65536,0.07337333096398248
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,65536,512,0.10072711441251968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2560,3072,0.007261333366235097
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,4096,65536,0.2291475666893853
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,6144,1024,0.06530488861931695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,3072,128,0.010538666612572141
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,12288,768,0.00490133340160052
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,512,8192,0.01034311122364468
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,512,3072,0.007394666473070781
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,65536,3584,0.6675990952385796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,4096,12288,0.017084444562594097
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,128,256,0.0028773333049482773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,32,1024,0.004917333523432414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2048,7168,0.010861333045694562
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,65536,512,0.10957688755459255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,32,10240,0.012181333369678922
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,6144,7168,0.012817777693271637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,128,512,0.005631999837027655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,3584,8192,0.04312977857059903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,2560,128,0.0052791109515561005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,1536,3072,0.014583999911944071
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,12288,3584,0.015040889382362366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,8192,8192,0.019063111808564927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,128,5120,0.009724444813198512
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,12288,128,0.006258666515350342
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,64,1536,0.0063235556913746735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,2048,7168,0.013042666845851474
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,256,4096,0.0439280006620619
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,12288,1536,0.010400888820489248
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,6144,512,0.006414222220579783
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,256,128,0.012141333685980903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,512,65536,0.06319200330310397
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3584,7168,0.025576889514923096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,12288,64,0.0029644444584846497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,128,64,0.0018542222678661346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,3072,65536,0.079312887456682
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,7168,32,0.020968000094095867
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,1024,5120,0.008447111480765873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,4096,7168,0.037998222642474704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,128,256,0.00536088893810908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2048,16384,0.019882667395803664
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,7168,1536,0.01994400057527754
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,5120,512,0.006983111302057902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,64,256,0.005589333259397083
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,65536,16384,0.16072355376349554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,65536,10240,0.21395466062757704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,512,6144,0.009335111412737105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,256,12288,0.013176889055305056
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,768,64,0.0021786666992637846
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,16384,16384,4.528353797064887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,128,4096,0.00904444439543618
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,65536,512,0.014530667000346713
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,32,256,0.0022880000372727713
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,6144,7168,0.013884444203641681
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2560,8192,0.014916444818178812
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,64,768,0.005924444645643234
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,8192,2048,0.02318488889270359
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,512,8192,0.02625244524743822
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,5120,10240,0.02124622298611535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,2560,8192,0.012053333222866058
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,7168,16384,0.051303111844592623
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,128,10240,0.05004622207747566
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1024,32,0.002152888932161861
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,128,8192,0.07075644201702543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,7168,3072,0.00906666699382994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,32,12288,0.004952888935804367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3072,32,0.00871377769443724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,7168,10240,0.02652799917591943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,3072,4096,0.024826667375034753
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3584,1024,0.003666666646798452
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,768,7168,0.006637333167923822
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,3072,12288,0.03248622351222568
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,64,10240,0.005663110978073544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,512,2560,0.00490399988161193
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,128,7168,0.02481511069668664
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,65536,3072,0.037645333343082
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,65536,128,0.02440711193614536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,256,65536,0.05528088741832309
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,7168,2560,0.008058666355080074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,2560,128,0.005375999957323074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3072,2048,0.006671999891599019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,64,768,0.005604444278611078
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,10240,1536,0.02419111132621765
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2560,4096,0.008411555654472774
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,32,256,0.0022515555222829184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,12288,6144,0.05170933405558268
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,1024,12288,0.033759110503726535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,1024,768,0.00562755556570159
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,5120,5120,0.023734221855799358
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2048,65536,0.3236142264472114
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,4096,1024,0.01650044487582313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,2048,128,0.005294222384691238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,1536,2048,0.00741155528359943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,8192,6144,0.15422488583458796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,4096,2048,0.007028444773621029
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,16384,256,0.002542222125662698
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,1024,65536,0.1308008829752604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,768,2560,0.014162666267818876
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,64,12288,0.02146400014559428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,5120,7168,0.012551999754375882
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,128,7168,0.009768000079525841
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,4096,768,0.018815999229749043
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,1536,128,0.00535111129283905
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,3072,768,0.016376000311639573
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,2560,3584,0.007693332930405934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,1024,2048,0.006672888994216919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,128,2560,0.01574577722284529
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,32,768,0.006650666809744305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,4096,4096,0.01019200020366245
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,65536,3072,0.042071110672420926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2048,2048,0.005633777628342311
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3584,65536,0.3240568902757433
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,65536,1024,0.018969777557584975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,8192,3072,0.011360000405046674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,768,10240,0.022595556245909795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,128,16384,0.005848000033034219
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,1536,2560,0.015570667054918079
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,8192,5120,0.1969511111577352
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3584,64,0.0024177777684397167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,8192,512,0.008023999631404877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1024,4096,0.006325333482689328
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,8192,1536,0.01569333341386583
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,16384,3072,0.017145777742067974
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,256,16384,0.04468889037768046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2048,4096,0.008399999803966945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,4096,65536,0.23099290000067818
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,768,4096,0.008460444708665213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,64,3584,0.004947555561860402
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3072,65536,0.064264886909061
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2560,10240,0.013552000125249227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2048,1024,0.006017777654859755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,10240,1536,0.009342222577995723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,12288,6144,0.027437332603666518
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,65536,256,0.008377778033415476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,12288,10240,0.029913776450686987
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2048,12288,0.011376000112957425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,10240,1024,0.007416889071464539
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,256,12288,0.03888533181614346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,6144,65536,0.3658551110161676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,256,1024,0.011433777709801992
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,2048,2560,0.0070266665683852295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1024,3584,0.005622222191757626
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,4096,12288,0.16871821880340576
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,32,1024,0.006335999816656113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,768,1536,0.027650665905740526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1024,2560,0.0070053330726093715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,768,12288,0.013823111024167804
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,12288,512,0.008023111356629265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,128,256,0.0052426668504873914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,5120,1536,0.008377778033415476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,128,5120,0.00489155575633049
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,16384,32,0.010092444717884064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,16384,4096,0.02338933282428318
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,5120,512,0.01554666625128852
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,8192,12288,0.025400888588693406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,32,32,0.0052355556852287715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,16384,3072,0.037428445286220975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,16384,32,0.02475822303030226
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,1024,10240,0.012716444830099741
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,256,1536,0.013032000097963544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,10240,512,0.007297777467303806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,65536,512,0.020590222544140287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,1024,3072,0.007290666302045186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,2560,6144,0.009004444711738164
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,1024,4096,0.033653332127465144
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,256,16384,0.08383466800053914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,512,32,0.0052720002002186244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3072,65536,0.06232355700598823
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,12288,16384,0.0686275561650594
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3072,6144,0.012263110942310758
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,3072,2048,0.021015110943052504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,12288,3072,0.027201778358883325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3584,768,0.007314666277832455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,6144,512,0.01588355501492818
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,10240,7168,0.026720000637902155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,6144,4096,0.06347022453943889
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2560,768,0.005586666779385672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,7168,8192,0.022788445154825848
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3584,3072,0.006968888971540663
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1536,768,0.0028239999794297745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,256,3072,0.007266666326257918
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,8192,10240,0.02121688922246297
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,2560,1024,0.005927111125654644
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,5120,5120,0.015975111060672335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,32,2048,0.0063982220987478895
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,4096,128,0.007647111184067196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,128,7168,0.011720889144473605
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,65536,7168,0.13196888234880236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,16384,7168,0.02346044447686937
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2048,2048,0.008963555925422246
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,2048,12288,0.021342222889264423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,32,6144,0.014171555638313293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3072,5120,0.016978666186332703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2560,768,0.005934222290913264
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,1536,4096,0.036890665690104164
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,2048,1536,0.022543110781245764
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2560,2048,0.006626666833957036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,1536,128,0.01757066614098019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,4096,2560,0.02723466687732273
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,65536,8192,0.18927110566033256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,6144,65536,0.07394666804207696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,16384,256,0.01850133306450314
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,8192,16384,0.044300443596310086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3072,128,0.0021564445147911706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3584,64,0.00555377784702513
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,16384,3584,0.03379733363787333
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1024,5120,0.005579555614127054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,65536,3072,0.0856328871515062
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,5120,3072,0.007637333538797166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,12288,768,0.007273777491516537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,3584,8192,0.04252800014283922
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3072,16384,0.02162488963868883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,256,6144,0.006251555350091722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,128,65536,0.04863377743297153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,32,256,0.006581333362393909
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,768,64,0.008267555799749162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,128,3072,0.007304000357786815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,10240,32,0.029614223374260798
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,3584,2560,0.020136000381575692
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1024,2560,0.004998222407367495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,5120,12288,0.04105955693456862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,8192,768,0.007594666547245449
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,1536,768,0.016134222348531086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,8192,128,0.0052764444715446895
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,512,1536,0.008672888908121321
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,256,3072,0.006635555376609166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,128,5120,0.018890667292806838
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,65536,2048,0.07477333148320515
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,4096,5120,0.01847288840346866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,5120,6144,0.011815999945004782
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,2560,65536,0.23383021354675293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,12288,65536,3.57444339328342
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,2560,512,0.021448888712459143
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,1536,7168,0.15476000308990479
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,65536,10240,0.16646577252282038
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,64,768,0.005942222144868638
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,65536,5120,0.10905066463682388
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,512,3584,0.01592177814907498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,256,8192,0.0070675553547011475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,8192,6144,0.014986667368147107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,512,3072,0.008640888664457533
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,5120,1536,0.006275555739800136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1536,6144,0.011429333024554782
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,2560,5120,0.022341334157519873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,64,6144,0.010019555687904358
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3584,7168,0.014018666413095264
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,8192,16384,0.029176887538697984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,1024,1024,0.015249777171346875
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,768,65536,0.1282542281680637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,256,10240,0.011789333489206104
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,7168,12288,0.045350223779678345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,256,768,0.0031475555151700974
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,4096,6144,0.050294223758909434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,4096,6144,0.01752711170249515
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,8192,6144,0.014027555783589682
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1024,128,0.00299822228650252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1024,768,0.007108444141017065
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,16384,768,0.008293333152929941
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,768,6144,0.009008000294367472
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,256,64,0.0022462221483389535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,5120,32,0.0021359999146726397
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,3072,16384,0.05616355604595608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3584,8192,0.013791999883121915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,1536,5120,0.02882666720284356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3072,1024,0.006931555353932911
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2048,128,0.00216977774269051
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,2560,1024,0.005991111199061076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,6144,5120,0.013531555732091268
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,768,7168,0.021709332863489788
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,8192,8192,0.03035911255412632
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,12288,2560,0.030986666679382324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,128,1536,0.003579555700222651
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,65536,512,0.020682666036817763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,768,64,0.0018817778262827133
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,64,3584,0.004998222407367495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,8192,2560,0.009674666656388177
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,3072,4096,0.008038221961922115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,4096,4096,0.012991999586423239
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,16384,512,0.006688888702127669
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,512,8192,0.020450666546821594
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1536,5120,0.008781332936551835
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1024,768,0.0029093333416514923
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,5120,3072,0.025477333201302424
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,6144,512,0.012328888807031842
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,6144,1024,0.020485333270496793
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,5120,3584,0.012158222496509552
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1536,2560,0.005622222191757626
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,65536,768,0.059954669740464955
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,512,2048,0.009378666679064432
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,32,256,0.0021768889079491296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,256,8192,0.005445333404673471
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,5120,3072,0.008605333666006723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,16384,1024,0.03958311014705234
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,512,3072,0.007335999773608313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,7168,1536,0.045703109767701894
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3584,256,0.005945777727497949
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,2560,6144,0.029919998513327703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,4096,6144,0.02304977840847439
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,4096,8192,0.02555911077393426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,6144,16384,0.022900443938043382
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,8192,6144,0.024699555502997503
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,4096,65536,0.21410932805803087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,4096,1024,0.016191111670600045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,16384,256,0.062441779507531055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,64,2048,0.0069671107663048645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,768,32,0.00564444437623024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,12288,2048,0.012727999852763282
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,8192,10240,0.01825777855184343
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,65536,256,0.03199022346072727
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,16384,768,0.006355555521117316
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,2560,2560,0.007224000162548489
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,12288,16384,0.043671111265818276
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,4096,128,0.014864888456132678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,12288,16384,0.05505333344141642
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,256,16384,0.029959112405776978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,65536,3072,0.031410667631361223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3584,7168,0.01366133325629764
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,512,5120,0.008991999758614434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,12288,5120,0.016598222984208
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,256,32,0.005967999911970562
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,65536,4096,0.051046222448349
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,6144,512,0.016899555921554565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3072,3584,0.007683555285135905
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,3584,6144,0.026269333230124578
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,1024,1536,0.00630666646692488
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3584,5120,0.010207110808955299
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,768,2048,0.004596444467703502
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,7168,2048,0.014451555079884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,6144,12288,0.03805422120624118
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,2560,3072,0.02628533376587762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,2560,4096,0.01846844454606374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,32,256,0.0025439999169773525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,10240,5120,0.029247111744350855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,5120,5120,0.03515911102294922
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,5120,768,0.01794577803876665
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,256,4096,0.015249777171346875
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,128,3584,0.007657777931955125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,7168,3584,0.02377066678471035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1536,8192,0.010791111323568555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,5120,256,0.0025031111306614345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,1024,16384,0.043228444125917226
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,5120,10240,0.056448002656300865
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2560,10240,0.0158595558669832
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,2560,128,0.014498665928840637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,2048,65536,1.587503963046604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,768,2048,0.017913776967260573
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,5120,2560,0.06634310881296794
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3072,1536,0.00830577810605367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,7168,5120,0.016208888755904306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,7168,4096,0.009770666559537252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,768,768,0.005928888916969299
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,64,32,0.005607999861240387
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,32,32,0.002171555534005165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,768,16384,0.02027466727627648
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,12288,1024,0.008086221913496653
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,256,256,0.0123848890264829
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,768,2048,0.0041893331540955436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,8192,768,0.011026666396194033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,16384,7168,0.02676888969209459
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,6144,2560,0.021957332889238994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,65536,2560,0.068904889954461
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,10240,65536,0.18206932809617785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,12288,1024,0.01000088867213991
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,128,16384,0.044554667340384595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,64,8192,0.010372444159454769
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,65536,2560,0.08859644333521526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,5120,2560,0.02028888960679372
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,768,8192,0.034684442811542086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3072,4096,0.010585777461528778
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,3584,6144,0.1915715535481771
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,12288,256,0.014755555325084262
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,256,768,0.005600000007285013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,128,5120,0.08443999952740139
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,16384,6144,0.02019377715057797
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,16384,768,0.015263999501864115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,7168,65536,0.07907466755972968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,12288,256,0.01717066599263085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,16384,12288,0.04136177897453308
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,128,5120,0.008310221963458592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,512,4096,0.01588800052801768
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3584,1024,0.003600888782077365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,65536,7168,0.16342666414048937
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,1536,7168,0.010623999767833285
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,5120,512,0.015463110473420886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3072,2560,0.01129155523247189
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,768,1536,0.0038373333712418876
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,16384,1024,0.007263110743628607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,256,512,0.005895111295912001
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,4096,2048,0.015821332732836407
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,2048,3072,0.007264888948864407
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,128,10240,0.005560889012283749
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3584,8192,0.014856000741322836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3072,32,0.002830222249031067
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,6144,65536,0.20588710572984484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,32,10240,0.011784888803958893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,12288,2048,0.010851555400424533
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,8192,5120,0.3850399918026394
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,512,5120,0.01204177737236023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,6144,10240,0.02437955637772878
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,768,256,0.003834666477309333
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,4096,6144,0.010068444742096795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,10240,1024,0.009324444664849175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,8192,64,0.002514666567246119
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,256,3072,0.007268444531493717
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,8192,7168,0.030565334690941706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,3072,12288,0.3374977906545003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,128,10240,0.05601155757904053
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,10240,4096,0.04311644368701511
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,2560,65536,0.09312978055742051
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,7168,16384,0.046535111135906644
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,4096,10240,0.04168355464935303
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,3072,1024,0.005931555396980709
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,7168,3584,0.025472000241279602
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,2048,6144,0.018896889355447557
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,3584,4096,0.01866666641500261
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,16384,1024,0.014793778459231058
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,12288,2560,0.0302817788388994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,1024,256,0.01813155578242408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,16384,16384,0.13506933053334555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,10240,7168,0.022942221826977197
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,32,256,0.005573333137565189
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2048,768,0.0028159999185138275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,8192,512,0.013927999469969006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,8192,12288,0.03490755624241299
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,10240,128,0.006247111078765657
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,6144,512,0.00534400012758043
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,12288,1024,0.008269333177142674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2560,5120,0.011355555719799466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,10240,5120,0.01850311126973894
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,768,12288,0.21869777308570015
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,512,7168,0.019235556324323017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,5120,2560,0.019574221637513902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,128,128,0.005904888941182031
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,128,3072,0.007318221860461765
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,512,6144,0.009342222577995723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,8192,64,0.00589155571328269
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,10240,32,0.0076248885856734375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,5120,4096,0.008397333323955536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1024,3584,0.006679111056857639
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,3072,6144,0.04135555691189236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1536,5120,0.008809777597586313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,10240,7168,0.047092444366878934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,16384,4096,0.015802666544914246
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,1536,65536,0.08042044772042169
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,10240,10240,0.08628000153435601
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,128,12288,0.013071999781661563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,3584,512,0.005559111220969095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,64,128,0.0058702222175068325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,8192,8192,0.027111111415757075
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,64,5120,0.008630221916569604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,768,1024,0.013758222262064615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,5120,2048,0.008849778109126622
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,7168,1024,0.019316444794336956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,7168,64,0.0021848889688650766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,5120,6144,0.036906666225857206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,32,1536,0.006253333141406377
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,4096,5120,0.0797377758555942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,10240,512,0.006585777633719974
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,10240,3072,0.015831111205948722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,64,768,0.0029057777590221832
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,6144,768,0.005912000106440649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,32,8192,0.011136000355084738
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,7168,128,0.004209777961174647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,8192,16384,0.039056890540652804
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3584,12288,0.0209840006298489
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,256,128,0.0022408889813555614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2560,16384,0.3952488899230957
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,7168,5120,0.016362667083740234
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1024,3584,0.006353777729802662
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,64,2560,0.005903999838564131
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,1536,1024,0.00739822205570009
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,3072,10240,0.014726221561431885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,12288,2048,0.011388444238238864
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,10240,3072,0.012455999851226807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,2560,65536,0.09420533312691583
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,16384,256,0.011532444092962476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,3584,2560,0.017893332574102614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,65536,32,0.016342222690582275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,6144,1536,0.016759999924235873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1024,3584,0.007679111427730984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,16384,128,0.0062773335311147906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1024,16384,0.02123822271823883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1536,3072,0.010073777702119615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,64,1536,0.008015111088752747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,65536,3072,0.03152799937460158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,768,128,0.0055440002017551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,16384,10240,0.05246222350332472
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,64,2560,0.004910222358173794
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,8192,1024,0.019157333506478202
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,16384,16384,0.07060977485444811
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,64,256,0.002149333349532551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3072,768,0.0038577777643998465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,1024,65536,0.06225866741604275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,2048,512,0.015188443991872998
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,4096,1024,0.0063991112013657885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,512,7168,0.019554666346973844
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,12288,6144,0.07405155234866671
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,8192,512,0.005567111074924469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,5120,12288,0.02134311033619775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,16384,7168,0.055431110991372
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,256,16384,0.07834221919377644
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,65536,6144,0.05750755469004313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,10240,3072,0.013193777865833707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,6144,1024,0.003949333396222857
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,2048,12288,0.05200799968507555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,256,5120,0.004886222382386525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1024,16384,0.013256000147925483
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,1536,3584,0.007274666594134436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,256,10240,0.02534488836924235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,65536,5120,0.12100888623131646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,768,65536,0.09919110933939616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,1536,8192,0.031480001078711614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,1024,256,0.015224888920783997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,768,1536,0.01484444406297472
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1024,16384,0.019832889238993328
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,8192,768,0.003354666754603386
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,512,12288,0.02475022276242574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,3584,5120,0.008352888955010308
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,12288,12288,0.04435377650790744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,12288,512,0.06975822316275702
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,768,3072,0.012169777519173093
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,10240,256,0.012677333421177335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,256,32,0.005225777626037598
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,8192,512,0.007225777539942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,2048,2048,0.006670222100284364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,10240,7168,0.027678221464157104
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,128,12288,0.004922666483455234
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,512,16384,0.011561777856614856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,768,6144,0.009639999932712978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,16384,16384,0.12311732769012451
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,10240,12288,0.04033244318432278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,5120,64,0.006623999940024481
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,2560,2560,0.04760355419582791
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,65536,2048,0.026947556270493403
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3584,64,0.002163555473089218
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,12288,512,0.009806222385830348
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2048,3072,0.008221333225568136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1536,16384,0.03170133299297757
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,128,16384,0.016903999778959487
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,4096,3072,0.014120888378885059
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,7168,64,0.020743111769358318
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,64,512,0.002554666664865282
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,12288,65536,0.15643021795484754
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3584,32,0.006695999867386288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,10240,5120,0.02015022271209293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,3072,10240,0.018009778526094224
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,768,2048,0.004418666578001446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3584,3584,0.01945600079165565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,10240,12288,0.0696293314297994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,768,32,0.00591466658645206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,65536,768,0.029556443293889362
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,3584,3072,0.028239998552534316
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1024,6144,0.00979200005531311
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,64,64,0.002218666589922375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,7168,128,0.01517511076397366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3584,8192,0.015407111909654407
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,1536,3584,0.007656888829337225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,3584,10240,0.050457778904173106
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,512,128,0.005271111097600725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,64,5120,0.00902844468752543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,128,256,0.005582222094138463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,1024,1024,0.010319111247857412
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,32,12288,0.025051555699772302
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,7168,12288,0.561837355295817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,256,512,0.005611555443869696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,16384,3584,0.019321777754359774
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,768,512,0.015358222855461968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2048,128,0.006659555352396435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,16384,5120,0.048241777552498706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,8192,2560,0.02271911170747545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,65536,12288,0.2145457797580295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,128,10240,0.018763555420769584
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,8192,1024,0.020071999894248117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,7168,65536,0.13125867313808864
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,7168,65536,0.1863937775293986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,12288,12288,0.07179199986987643
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,5120,10240,0.02602133320437537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,10240,10240,0.03137866655985514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,5120,7168,0.015248888068728976
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,3584,128,0.005614222337802251
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,512,768,0.02125599980354309
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,512,2560,0.007284444239404466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,32,768,0.002502222235004107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3584,1536,0.0063075555695427795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1024,6144,0.01904177831278907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,12288,5120,0.11123643981085883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,16384,128,0.015474667151769003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,2560,256,0.005327111317051782
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,5120,3584,0.010392000277837118
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,3072,3072,0.009025777379671732
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,12288,8192,0.03819466630617777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,6144,6144,0.026345777842733595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,256,3072,0.01862133377128177
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,4096,10240,0.01595288846227858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,12288,10240,0.7115626864963107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,4096,32,0.0028862222615215513
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,32,8192,0.020241777102152508
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,10240,256,0.006635555376609166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,10240,65536,0.21337244245741102
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,32,6144,0.009552000297440423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,16384,12288,0.03916977842648824
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,2560,2560,0.01722577710946401
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,768,1536,0.0063182223174307085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,10240,8192,0.11162488990359837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,10240,1536,0.02257333364751604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,1536,1536,0.031552000178231135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,5120,768,0.005889777921968036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,2048,12288,0.01459377838505639
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,10240,8192,0.04196977615356445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,65536,32,0.02567111121283637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,6144,128,0.005567111074924469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,16384,65536,0.38718933529324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1536,4096,0.013742222554153867
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,512,6144,0.016887111796273124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,10240,1536,0.008734222087595198
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,768,1024,0.0062773335311147906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,768,10240,0.09781777858734131
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,32,64,0.0059075554211934405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,128,3584,0.014146667387750415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,512,12288,0.02882666720284356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,512,128,0.005882666756709416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1536,512,0.0028488888508743713
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,6144,2560,0.008928000099129146
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,3072,1536,0.016791999340057373
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,1536,3584,0.010255999863147736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,7168,768,0.005935110979610019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,65536,32,0.010223999619483948
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,16384,65536,0.22649688190884062
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,32,3072,0.004933333231343163
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,2048,10240,0.18460088306003145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1536,1024,0.007279111279381647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,512,256,0.005345777918895085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,8192,2560,0.008623999853928884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,8192,65536,0.13998133606380886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,12288,3072,0.01332355539004008
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,768,65536,0.05596711238225301
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,5120,128,0.005601777798599667
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,32,16384,0.007305777735180325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2560,10240,0.11058488819334243
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,10240,2560,0.012406222522258759
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,768,12288,0.026913776993751526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,2560,3584,0.007714666426181793
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,64,10240,0.0058808889653947614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,65536,16384,1.2893617418077257
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,16384,1024,0.015615999698638916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,7168,5120,0.011844444606039258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,7168,10240,0.025775111383861963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2048,10240,0.1088542209731208
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,512,1536,0.003681777666012446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,12288,65536,1.5995101928710938
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,12288,5120,0.016494222813182406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,768,7168,0.00961688905954361
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,10240,512,0.00666222224632899
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,3584,3072,0.007302222152551015
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,256,7168,0.006724444528420766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,768,128,0.0064399999876817065
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,16384,3584,0.022643556197484333
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,5120,7168,0.015627556376987033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,256,5120,0.005280888742870755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,64,256,0.006007110906971826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,5120,32,0.0021911110315057966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,16384,128,0.002226666650838322
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2048,10240,0.013608000344700284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,128,3072,0.009353777600659264
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,2560,65536,0.06561777989069621
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,10240,65536,0.14672088623046875
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,12288,2048,0.017103110750516255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2560,2560,0.03532355692651536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,6144,2560,0.015273777974976433
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,512,7168,0.009738666315873465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,2560,32,0.002228444442152977
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,512,8192,0.019083556201722886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,4096,10240,0.016727111405796476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,64,8192,0.011559999651379056
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,10240,10240,0.025694222913848028
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,3584,512,0.015174221661355762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,6144,3584,0.2026453283098009
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,4096,256,0.0035235554807715942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,4096,5120,0.011746666497654386
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1536,3072,0.009675555759006077
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1536,3072,0.0508284436331855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,3584,2048,0.015541333291265698
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,512,3584,0.015963556038008798
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1024,65536,0.024759110477235582
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,4096,4096,0.029559112257427637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,8192,1536,0.0069724445541699724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,1024,256,0.0059368887709246735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,3072,3072,0.05594755543602837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,64,8192,0.06573510832256742
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,2560,16384,0.024458666642506916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1024,512,0.0029013332807355454
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,2048,8192,0.02292799949645996
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,2048,65536,0.14525777763790554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,10240,2048,0.025546666648652818
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,768,2048,0.006631999793979857
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,3584,8192,0.04505155483881632
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,10240,2560,0.01000711073478063
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,6144,2048,0.011780444118711682
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,256,3584,0.02899200055334303
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,512,3072,0.007290666302045186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,512,5120,0.00906311141120063
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,6144,2560,0.02607911162906223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,64,8192,0.015523556205961438
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2560,3584,0.00830666638082928
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2560,512,0.002537777854336633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,768,1024,0.006248888870080312
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,512,768,0.0028533333291610083
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,8192,8192,0.03383200036154853
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,256,768,0.006284444282452266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,1024,10240,0.03748266564475165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,6144,7168,0.012425777812798819
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1536,128,0.005527110977305307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,3072,8192,0.025597333908081055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,8192,768,0.00598933340774642
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,64,512,0.005389333185222413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,10240,128,0.0364586677816179
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,4096,3072,0.010538666612572141
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,256,128,0.005249777601824866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,65536,256,0.024878222081396315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,6144,4096,0.012433778080675336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,768,5120,0.009015999734401703
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,65536,6144,0.13354755772484675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,6144,16384,0.04890755481190152
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,128,6144,0.022463111413849726
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2560,10240,0.01594577729701996
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1536,65536,1.0327609380086262
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2560,2560,0.0110871113008923
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,5120,512,0.0041982221106688184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1536,16384,0.1867013374964396
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,5120,6144,0.013221333424250284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,64,4096,0.008669333325492011
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,6144,8192,0.014988443917698331
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,12288,128,0.004272888931963178
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,65536,8192,0.08948889043596055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,65536,3584,0.03972355524698893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,32,7168,0.01792977750301361
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2560,3584,0.007355555891990662
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,4096,16384,0.0440906650490231
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,6144,5120,0.027183110515276592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,2048,256,0.014856000741322836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,6144,8192,0.016889777448442247
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,12288,2048,0.012777778009573618
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,7168,1536,0.020724443925751578
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,7168,2048,0.006990222467316522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,7168,8192,0.0236826671494378
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1024,12288,0.013904889424641928
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,3584,3072,0.009370666411187913
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,768,1536,0.0063075555695427795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,1536,128,0.01462044484085507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,128,1536,0.0038933331767717996
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2048,1024,0.005969777703285217
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,65536,5120,0.04783466789457533
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,10240,4096,0.018569777409235638
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,1024,512,0.00565422202150027
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,64,32,0.002209777840309673
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,10240,256,0.019626667102177937
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,512,16384,0.015509333875444202
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1536,3584,0.01351288871632682
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,3072,7168,0.03242222136921353
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,512,4096,0.006284444282452266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,512,12288,0.013452444639470843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,2048,1536,0.0063742221229606206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1536,12288,0.014332445131407844
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,7168,768,0.016872000363137987
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,5120,16384,0.8824071354336209
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,256,3584,0.02169422143035465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,64,64,0.01036622209681405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,7168,256,0.005559111220969095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2048,2560,0.018928888771269057
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,768,256,0.0148062225845125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,512,256,0.0058542220956749385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,6144,128,0.034296890099843345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,4096,3072,0.009326222042242685
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2560,2560,0.006240889016124938
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,8192,128,0.04012889001104567
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,128,65536,0.05963377820120918
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,12288,768,0.08566311332914565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,32,1024,0.002934222213096089
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,16384,65536,0.4498746660020616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,5120,768,0.0031519999934567344
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1536,128,0.002218666589922375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,512,16384,0.015526221858130561
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,1536,10240,0.02439911166826884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,16384,7168,0.037216888533698186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,16384,3584,0.04081866807407803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,8192,512,0.03720622261365255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2048,128,0.002505777817633417
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,6144,512,0.007319110963079665
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,3072,16384,0.021009777983029682
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,5120,10240,0.016522667474216886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,6144,3584,0.07878933350245158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,16384,128,0.022817777262793645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,3584,1024,0.016187555260128446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,3584,2048,0.009903111391597325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,1024,512,0.006263999889294307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,128,5120,0.00905777762333552
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,5120,256,0.017481777403089736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,512,5120,0.012504888905419244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,4096,64,0.005897777775923411
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,3072,3584,0.007649777664078607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1536,12288,0.012674666941165924
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,12288,1536,0.17457866668701172
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,1024,128,0.005136888888147143
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,10240,768,0.047994666629367404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,768,12288,0.023374223046832617
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,6144,256,0.01679644485314687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,10240,1024,0.0069448889957533936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,3072,3584,0.028367112080256145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,32,768,0.014835556348164877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,1024,1536,0.013873777455753751
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,8192,12288,0.3256506654951307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,12288,128,0.006235555642180973
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,12288,12288,0.051028443707360156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,8192,6144,0.015750222735934787
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,768,2048,0.015527110960748462
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,768,4096,0.014134221606784396
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,65536,16384,0.35915467474195695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,10240,128,0.0029048888633648553
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,768,16384,0.018235555953449674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,256,5120,0.022647110952271357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3584,5120,0.025731555289692346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,8192,6144,0.03674488928582933
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,7168,1536,0.006298666612969504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,4096,64,0.003544888976547453
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,4096,512,0.015433778365453085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,65536,32,0.011977777712874942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,12288,32,0.007576000359323289
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3584,2048,0.00814666681819492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,3072,5120,0.00869955536392
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,2048,2560,0.02035466664367252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,512,2560,0.004956444518433677
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2048,1024,0.0032684444967243406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,7168,3584,0.029807110627492268
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,128,6144,0.00627377794848548
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,65536,2560,0.05776266919242012
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,4096,10240,0.01711466742886437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,7168,2560,0.007349333001507654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,8192,2560,0.0094373333785269
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,65536,1024,0.07989955610699125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,65536,10240,0.11242400275336371
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1536,4096,0.008035555481910706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,768,2048,0.007299555672539606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1024,7168,0.01390044391155243
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,768,10240,0.01200977795653873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,65536,4096,0.18733155727386475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,12288,3584,0.042280889219707914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,65536,64,0.013743110828929476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,2560,16384,0.021730666359265644
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1024,64,0.005679111099905438
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,256,7168,0.005623999983072281
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,64,3072,0.007646222081449296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,6144,1536,0.006760889043410619
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,1536,128,0.022322666313913133
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,8192,7168,0.013623110949993134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,32,2560,0.013233777549531726
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,10240,1536,0.008403555386596257
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3072,1536,0.00905600024594201
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1536,32,0.0021822222818930945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,3072,3584,0.0173662222094006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,128,768,0.014145778285132514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,6144,1536,0.04292266567548116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,4096,8192,0.023243554764323767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,3584,2560,0.007330666813585494
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,10240,7168,0.06380355358123779
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,1024,2560,0.018862222631772358
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2560,8192,0.01898755629857381
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,2560,256,0.005616000129116907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2048,65536,0.044529779089821704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,3584,8192,0.04247377647293938
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,768,64,0.0018942221585247253
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,4096,256,0.005609777652555042
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,8192,65536,0.13004355960422093
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3072,128,0.005888888819350137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,10240,1024,0.00759733302725686
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,65536,1024,0.01814311080508762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,65536,10240,0.0857422219382392
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,5120,5120,0.026531555586391028
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,3072,16384,0.02130488885773553
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,16384,12288,0.06414488951365153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,5120,8192,0.032920890384250216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,7168,768,0.0035706667436493766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,128,1024,0.011721777419249216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2560,65536,0.06921422481536865
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,768,64,0.005919999960396025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,768,32,0.00627377794848548
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,256,1024,0.01240355521440506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2560,768,0.007134222322040134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,6144,10240,0.019537778364287484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,128,16384,0.03323644399642944
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1024,256,0.005234666582610872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,12288,3072,0.03172889020707872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,10240,512,0.009107555780145857
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,16384,2048,0.012815111213260226
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,10240,4096,0.018052443861961365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,5120,2048,0.008285333712895712
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,8192,32,0.002530666689078013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,64,512,0.0069297779765393995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,512,8192,0.008030222521887885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,512,512,0.01221333361334271
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,7168,64,0.005910222315125995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3584,65536,0.07436088720957439
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,10240,768,0.018574222922325134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,4096,12288,0.11224533451928033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,8192,12288,0.019673777951134574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3584,6144,0.011996444728639392
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2048,5120,0.007349333001507654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,8192,8192,0.03277066681120131
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,512,4096,0.01108088923825158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2560,2560,0.006577777779764599
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,256,256,0.006948444578382704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,512,10240,0.011790222591824003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,512,4096,0.01714933415253957
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,4096,128,0.0074417781498697065
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,4096,64,0.0022906667242447534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,768,16384,0.017975111802419026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,512,1024,0.0061928890645504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,5120,128,0.0053004444473319584
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,2048,65536,0.06383555465274386
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,4096,16384,0.0225591113169988
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,256,12288,0.011911110745535957
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,3584,256,0.023240889112154644
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,12288,1536,0.00867377801073922
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,512,8192,0.045259555180867515
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,5120,65536,0.12211910883585612
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,5120,512,0.005718222094906702
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,768,32,0.0020666666742828158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,1024,3072,0.007706666986147563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,65536,4096,0.19943643940819633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,16384,10240,0.4676862292819553
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,128,16384,0.012864888542228274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,32,3584,0.008976000050703684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,512,512,0.006602666858169768
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,64,256,0.005939555664857228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,256,128,0.002534222271707323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,2048,65536,0.06679733594258626
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,5120,65536,0.07292355431450738
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,8192,7168,0.02071555621094174
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,256,768,0.0053324442770746015
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2048,4096,0.040496001640955605
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,8192,128,0.005568888866239124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,768,6144,0.014476444986131458
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,10240,6144,0.045138667027155556
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,8192,3584,0.014553777045673795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3072,7168,0.010936000280910067
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,64,7168,0.010348444183667501
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1024,64,0.00833333366447025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,256,12288,0.013411555853154926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,4096,6144,0.012773333324326409
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,3072,1536,0.015178667174445258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3584,16384,0.02444800072246128
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,3072,4096,0.0181342214345932
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,10240,1536,0.06859466764662
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1024,1024,0.005604444278611078
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,3584,1024,0.014838222000333997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,3072,16384,0.020624000165197585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,64,5120,0.008681777450773451
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3584,256,0.0024888888001441956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,2560,7168,0.009317333499590555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,10240,8192,0.0440613329410553
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,7168,2048,0.02896711230278015
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,5120,128,0.005299555758635203
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,128,768,0.014422222971916199
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,1024,65536,0.09157688750161065
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,65536,5120,0.05807377894719442
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3584,768,0.005592888842026393
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,768,5120,0.010132444401582083
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,1024,8192,0.011364444262451597
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,12288,256,0.009383111364311641
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,16384,10240,0.21540710661146376
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,5120,7168,0.02167644434505039
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3072,2560,0.007618666523032718
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,7168,7168,0.029116445117526587
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,3072,2048,0.017154667112562392
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1024,256,0.002463111033042272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,16384,128,0.0024622221373849446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,512,10240,0.010648888846238455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,768,65536,0.056827558411492236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,10240,256,0.0025253333151340485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,16384,12288,0.08025778002209134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,64,3072,0.010320888625250923
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,10240,16384,0.05673333009084066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,5120,1024,0.006649777707126405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,768,65536,0.06389511293835111
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,128,16384,0.00591555568906996
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,512,3584,0.01514933341079288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2560,768,0.00904444439543618
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1536,4096,0.007954667011896769
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,6144,512,0.01923644377125634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,65536,5120,0.05718133184644911
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,3072,65536,0.15580088562435573
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,4096,16384,0.024195555183622573
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,16384,10240,0.035070220629374184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,12288,5120,0.016418667303191293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,10240,12288,0.06185599830415514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,3584,4096,0.008446222378147973
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,3072,2560,0.009468444519572789
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,1536,2560,0.01517511076397366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,16384,512,0.003988444391224119
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,16384,256,0.0028444443725877335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,128,64,0.0053688887920644544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,7168,65536,3.864841037326389
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2560,2048,0.008290666672918532
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,2560,5120,0.010431999961535135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,6144,6144,0.011448000040319232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3584,65536,0.06557155317730375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,512,3584,0.00795111142926746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,8192,64,0.0024008889579110676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,512,2560,0.007311111523045435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3072,1536,0.005807111246718301
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,512,65536,0.018423999349276226
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,32,6144,0.005333333379692501
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,768,4096,0.018441778090265062
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,1024,10240,0.012086222569147745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,768,64,0.006680888848172293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,128,3072,0.010059555371602377
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,6144,4096,0.00960799968904919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,2048,5120,0.027756444282001917
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,256,4096,0.01347733371787601
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,256,128,0.005678222411208683
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1536,512,0.005335111171007156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,10240,10240,0.47522486580742734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,7168,7168,0.02030044462945726
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,16384,12288,0.0542791121535831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,512,65536,0.11539378431108262
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,32,2048,0.006911110960774952
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,8192,512,0.0037866665257347953
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,128,10240,0.029046220911873713
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,10240,5120,0.018159111340840656
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,3584,65536,0.06657333506478204
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,32,128,0.007993777592976889
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,256,6144,0.016775111357371014
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,4096,12288,0.021915555000305176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,768,6144,0.01864888932969835
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,8192,128,0.015028445257080926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,32,128,0.0053093334039052325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,2048,768,0.017490666773584154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,10240,6144,0.02068444424205356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,16384,5120,0.13040622075398764
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,5120,2048,0.05519466598828634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,10240,1536,0.026497777965333726
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,128,2048,0.008934222161769867
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,512,2560,0.005232000102599462
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,8192,8192,0.04875555634498596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,12288,65536,0.1461875571144952
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,1024,10240,0.011715555356608497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2560,128,0.00217333332531982
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2560,768,0.003528888854715559
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3072,3584,0.0073520003093613525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,2560,2048,0.018581334087583754
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,1536,8192,0.06609866354200575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,6144,7168,0.10994578070110744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,65536,7168,0.09754133224487305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,1536,2560,0.00692622239391009
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,256,65536,0.01354311075475481
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,128,3072,0.007278222176763747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2048,64,0.0028000000036425064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,256,7168,0.03017510970433553
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,16384,512,0.009299555586444007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,4096,1024,0.003549333247873518
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,16384,4096,0.03342577815055847
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1024,12288,0.021095999413066443
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2560,5120,0.008613333106040955
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,768,4096,0.017543110582563613
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,256,128,0.005586666779385672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1536,768,0.002879111096262932
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,10240,16384,0.055162668228149414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,1024,2560,0.0069226668112807805
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,768,128,0.002143111079931259
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,3584,1536,0.00630044440428416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,10240,2560,0.01866311166021559
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,1536,512,0.027081777652104694
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3584,768,0.005921777751710679
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,16384,3584,0.5404746797349718
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,16384,32,0.011038222246699862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2560,7168,0.03561066587766012
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,128,10240,0.024102222588327196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,2048,8192,0.07284800211588542
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,512,1536,0.0038924444880750445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,16384,65536,0.14380889468722874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,65536,12288,0.19032977686987984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,8192,3584,0.03915111223856608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,4096,2048,0.016335111525323655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,12288,3584,0.014177777700954013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2048,4096,0.009114666945404476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,5120,1024,0.005912888795137405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1024,3072,0.011659555964999728
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,12288,7168,0.05457955598831177
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,128,7168,0.004946666873163647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2560,12288,0.017239999439981248
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,32,6144,0.009338666995366415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2048,128,0.005947555518812603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1536,128,0.002481777841846148
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1024,1536,0.003912888881233004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,2560,3584,0.019596444235907663
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,768,1024,0.013398222625255585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,512,8192,0.01034311122364468
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,10240,768,0.005888888819350137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,512,1024,0.0031600000543726813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2560,128,0.002480888946188821
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1536,4096,0.008375999828179678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,3584,5120,0.1614417764875624
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,768,65536,0.06161777840720283
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,256,16384,0.005565333283609814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,12288,512,0.004265777766704559
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1536,1024,0.005930666708283954
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,768,3072,0.010418666733635796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,10240,6144,0.0432604451974233
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1536,768,0.004223111189073986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,5120,3072,0.010664000279373594
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,7168,3072,0.030287110143237646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,256,2048,0.007237333390447829
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,1536,65536,0.5022995736863878
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3584,4096,0.008736888567606607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,1536,4096,0.022993778189023335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,6144,6144,0.017543110582563613
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,4096,6144,0.12333688471052383
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,8192,1024,0.05013155606057909
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,16384,1024,0.02607022225856781
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,6144,128,0.005695999910434087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,6144,10240,0.305730660756429
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,1536,2560,0.00831822223133511
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,4096,5120,0.03415288858943515
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3584,64,0.013391111459996967
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,512,128,0.0058764442801475525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,2560,8192,0.010429333481523724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,65536,12288,0.23302221298217773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,16384,7168,0.6368133227030436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1024,3584,0.005960000058015187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,8192,7168,0.038605332374572754
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,12288,8192,0.03170133299297757
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,4096,64,0.024400888217820063
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,64,32,0.005294222384691238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,6144,7168,0.028246220615175035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3072,2560,0.005610666755172942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,768,2048,0.004226666771703296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,3584,4096,0.008392000363932716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,4096,128,0.005639111002286275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,3072,7168,0.033923556407292686
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2560,65536,0.06404799885219999
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,5120,10240,0.035008000002966985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,3072,8192,0.0292231109407213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3072,7168,0.013822221921549903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,1024,16384,0.03605155481232537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,7168,768,0.007624000310897827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,1536,10240,0.042493333419164024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,2560,7168,0.009762666291660732
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,12288,3072,0.016247999336984422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,12288,512,0.027936001618703205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,4096,8192,0.015213333898120456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,256,7168,0.02478222217824724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,65536,7168,0.16577066315544978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,4096,8192,0.01482222146458096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,3584,16384,0.02165511084927453
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,8192,10240,0.0551546679602729
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,65536,128,0.011687999798191918
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,128,1536,0.006263999889294307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,1536,1536,0.006339555399285422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1536,10240,0.015811555915408664
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,16384,65536,0.254784001244439
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,65536,128,0.004925333377387789
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,3072,65536,0.06462488571802776
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,6144,1024,0.018953777021831937
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,16384,4096,0.05827733543184069
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,32,1536,0.007689778175618913
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1024,2048,0.006664888726340399
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,12288,256,0.022661333282788593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3584,256,0.0023457778410779103
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,8192,5120,0.017109332813156974
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,64,32,0.005364444520738389
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,5120,16384,0.022973333795865376
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,6144,65536,0.6673137876722547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,7168,6144,0.07512978050443861
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2560,12288,0.018477333916558158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,768,768,0.015992000699043274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,4096,8192,0.016744888491100736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,4096,6144,0.012451555165979596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,512,5120,0.008662222160233391
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,2048,65536,0.3061102231343587
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,128,64,0.005966222120655908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1536,2560,0.02752799954679277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,12288,2560,0.03331111205948724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,256,1536,0.007491555478837755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,10240,65536,0.13162222173478869
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,8192,16384,0.02644266684850057
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,128,1536,0.006061333335108227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,16384,4096,0.0499199993080563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,2048,12288,0.02716000046994951
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,3584,10240,0.015112888481881885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,128,8192,0.005653333332803514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2048,16384,0.017093333933088515
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,4096,3072,0.007304888632562425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,65536,8192,0.169469330045912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,3072,7168,0.02276888820860121
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,2048,128,0.014473777678277759
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1536,128,0.005292444593376583
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,7168,65536,0.14275466071234807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,5120,2560,0.008101333346631791
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3072,5120,0.012540444731712341
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1536,16384,0.021663111117151048
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,4096,10240,0.028215110301971436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,6144,6144,0.016528000434239704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,2048,65536,0.06348088714811537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,3584,6144,0.03482044405407376
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,768,128,0.0021466666625605687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,12288,768,0.013350221845838757
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,5120,1536,0.03304622239536709
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,3584,3072,0.02544622288809882
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,2560,512,0.013506666653686099
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,3072,2560,0.007255111303594377
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,8192,2048,0.008752889103359645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,768,7168,0.009399111072222391
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,8192,3072,0.01016533292002148
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,4096,12288,0.055155555407206215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,3584,512,0.01627111103799608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,5120,4096,0.019161777363883126
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,768,5120,0.008980444735950893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,4096,8192,0.01519733336236742
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,7168,7168,0.012563555604881711
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,6144,1536,0.00618488879667388
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,128,10240,0.01962577799956004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,7168,256,0.005984888722499211
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,16384,16384,0.07779822084638807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,1024,768,0.01883288886811998
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,6144,12288,0.041153778632481895
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3584,3584,0.01069777790043089
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,7168,1024,0.007087111473083496
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3072,1536,0.0069724445541699724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,10240,1536,0.00867911097076204
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1536,2560,0.006981333096822103
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,4096,3584,0.038763556215498186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,10240,512,0.018901333212852478
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,5120,6144,0.016907556189431083
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,512,512,0.006280888699822956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,5120,10240,0.0374257763226827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,8192,1024,0.06851910882525973
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,1024,1024,0.013443555268976422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,512,12288,0.013454222016864352
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,128,2048,0.006607999818192587
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,10240,1024,0.011731555892361535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,12288,1024,0.007357333269384172
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,5120,768,0.005936000082227919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,7168,3584,0.04913066824277242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,4096,768,0.005611555443869696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,10240,65536,0.18945156203375924
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,512,3072,0.005972444597217772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,6144,256,0.004211555752489301
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,256,10240,0.024064888556798298
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,16384,1536,0.034091555409961276
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,10240,5120,0.020191111498408847
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,2048,2560,0.007641777396202087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,7168,2560,0.007292444507280986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,12288,512,0.045718222856521606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,768,1024,0.008085333638721043
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,64,8192,0.0395066671901279
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1024,10240,0.012464889221721225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,8192,8192,0.02420622275935279
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,768,2560,0.01994755533006456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1536,5120,0.0087333329849773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,2560,12288,0.03028977910677592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,768,8192,0.011761777930789523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,128,768,0.005643555687533484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,256,2048,0.009157333109113906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,5120,3584,0.009418666362762451
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,16384,1536,0.010028444230556488
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3072,1024,0.005949333310127258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,3072,7168,0.010781333678298526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,3072,3072,0.024907555845048692
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2560,128,0.007678222325113084
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,64,6144,0.017676444517241586
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,6144,12288,0.01943911115328471
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3072,8192,0.014533332652515836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2048,2048,0.005909333212508096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,6144,4096,0.032066666417651706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,128,12288,0.023656888140572443
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,16384,768,0.006129777679840724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,7168,512,0.04330933425161573
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,512,6144,0.010508444574144153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,1024,5120,0.00832177781396442
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,7168,4096,0.019820445113711886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,128,3584,0.014168888330459595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,4096,65536,0.09592977497312759
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,64,32,0.0058746664888328975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,4096,12288,0.01904444396495819
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,768,3584,0.008676444490750631
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,10240,128,0.006335999816656113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,6144,10240,0.06247911188337538
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,2560,6144,0.008736888567606607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,64,16384,0.011513777905040316
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,2560,768,0.017108443710539076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,6144,2560,0.009278222090668148
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,4096,2048,0.006607111129495833
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,768,65536,0.16232889228396946
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,8192,3584,0.011391110718250275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,8192,7168,0.014927110738224454
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,12288,5120,0.032631112469567194
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,1024,512,0.012824888858530255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,12288,2560,0.013219555219014486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,12288,1024,0.007663999994595845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,6144,10240,0.03629422187805176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,65536,256,0.018988443745507132
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,256,7168,0.015486222174432544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,3584,4096,0.02552177839808994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,8192,65536,0.8340257538689507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,6144,128,0.005587555468082428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2048,1536,0.006292444550328785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2560,1536,0.006344888773229387
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,10240,512,0.03939200109905667
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,768,5120,0.036687112516827054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,64,6144,0.009420444567998251
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,128,2048,0.0046364445653226645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,64,1024,0.003165333428316646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,256,6144,0.009321778184837764
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,6144,10240,0.01947733263174693
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,10240,1536,0.008948444492287105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,512,5120,0.009475555684831407
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,12288,64,0.0034168888297345904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,2560,512,0.014422222971916199
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,12288,1536,0.026808889375792608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,12288,2560,0.030497776137457952
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1024,128,0.011743110915025076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,512,1024,0.005930666708283954
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,12288,2560,0.012357333468066322
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,2048,16384,0.06657866636912028
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,10240,12288,0.033246222469541765
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,16384,4096,0.022249778111775715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2560,8192,0.013625777430004544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,4096,6144,0.00963199966483646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,256,3584,0.00740533322095871
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,1024,12288,0.03158844510714213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,768,6144,0.06578844123416476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,10240,1024,0.021756443712446425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,4096,6144,0.009493333597977957
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,512,8192,0.010774222513039907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,65536,6144,0.0630542238553365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,64,3584,0.00795111142926746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,12288,768,0.006844444407357111
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,768,8192,0.009485333330101436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,768,10240,0.016175111134847004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,512,2048,0.015792000624868605
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,5120,8192,0.01625955601533254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1024,128,0.0022320000247822865
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,512,1536,0.01056977775361803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,8192,7168,0.18115199936760795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,512,64,0.0021484444538752237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,8192,2048,0.021681777305073206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,2560,2048,0.016562667157914903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3584,1536,0.003920000046491623
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,256,1024,0.013624000052611033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,2560,128,0.0052355556852287715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,6144,3072,0.021977777282396953
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,768,2048,0.004568888909286923
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,32,2048,0.006610666712125142
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,65536,768,0.01734044485621982
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,12288,2048,0.012129777835475074
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,5120,3584,0.008803555534945594
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,6144,1024,0.005221333354711533
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,1024,256,0.005607111172543631
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,12288,16384,0.04251733422279358
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,7168,3584,0.011759111450778114
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,8192,1024,0.010701333483060202
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,12288,10240,2.0748248630099826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,768,768,0.028396444188223943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,64,12288,0.028532445430755615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,10240,65536,0.14428977171579996
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,16384,8192,0.19430666499667695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1536,256,0.005628444254398346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,1536,3072,0.007639110916190677
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,768,6144,0.010453333457310995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,16384,1024,0.014107555150985718
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,16384,2560,0.1437564425998264
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,768,4096,0.00810222244924969
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,16384,5120,0.038000000847710505
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2048,3584,0.013059555656380124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,3584,6144,0.011383111278216044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3072,512,0.0035128887328836652
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1024,4096,0.006211555666393704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,256,2560,0.005324444423119227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,64,1024,0.0032195556494924757
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,128,16384,0.04600444436073303
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,65536,4096,0.08715911044014825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,10240,3072,0.013297778036859302
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2048,8192,0.01422044469250573
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,16384,16384,0.19743733935885957
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,512,7168,0.027931554449929133
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,8192,768,0.016132444143295288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,7168,8192,0.03188622328970168
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,3072,7168,0.023834667272037927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,12288,8192,0.03552000059021844
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,65536,512,0.11877155303955078
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,65536,1536,0.0689893364906311
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,64,5120,0.004625777817434735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,12288,512,0.010063110954231685
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,5120,1536,0.009484444227483537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,512,7168,0.010838222172525195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1024,1536,0.00629333323902554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,7168,16384,0.038973334762785174
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,4096,1024,0.00627377794848548
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,512,5120,0.035785777701271906
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,128,3072,0.004595555365085602
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,16384,128,0.004231111043029361
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2048,4096,0.04390933447413974
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,2048,2560,0.018569777409235638
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,16384,8192,0.024961777859263953
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,128,16384,0.0052639999323421055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,64,65536,0.10500266816880967
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,32,32,0.008007110820876228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3584,2048,0.005040888984998067
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,512,1536,0.006088888893524806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,16384,4096,0.02948533495267232
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,12288,3072,0.014820444915029736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,768,7168,0.022184888521830242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,16384,1536,0.03031555480427212
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,6144,64,0.0019368889431158702
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,768,2560,0.004978666702906291
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,256,10240,0.011920889218648275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,1024,128,0.005260444349712796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3584,64,0.011161777708265515
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,256,128,0.006622222148709827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,2560,12288,0.04042222102483114
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,768,2048,0.006631111105283101
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,12288,3072,0.01275199982855055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,12288,512,0.003560888684458203
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,2560,6144,0.027830223242441814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3584,2048,0.0070266665683852295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,12288,128,0.003958222352796131
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,512,2560,0.007378666765160031
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2560,6144,0.010761777559916178
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,7168,512,0.005704000178310607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,4096,512,0.0032426667296224167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,12288,768,0.005994666781690385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,8192,64,0.005382222019963794
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,4096,768,0.006038222048017714
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,128,4096,0.011752888560295105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,768,128,0.0021893332401911416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,16384,64,0.005516444643338521
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,7168,1536,0.0069546666410234236
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,12288,3584,0.016553777787420485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,12288,6144,0.04502844479348925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,4096,2048,0.006639110959238476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1024,6144,0.016135111451148987
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,16384,65536,0.6533697976006402
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,128,512,0.01664888858795166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,128,5120,0.008001777860853408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3584,128,0.006000888844331105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,4096,6144,0.01774666706720988
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,8192,65536,0.09088000324037339
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2560,2560,0.0070062221752272705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,10240,3072,0.012400888734393649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,5120,4096,0.011043555206722684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,5120,10240,0.015204444527626038
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,768,4096,0.007672000262472365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,512,7168,0.018037334084510803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,512,256,0.005640000104904175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,7168,10240,0.03279555506176419
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,10240,512,0.018590221802393597
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,65536,7168,0.18248889181349012
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,128,4096,0.005407111098368962
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,1536,1536,0.017268444101015728
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,4096,6144,0.013537777794731988
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3072,32,0.003896000070704354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,4096,2560,0.009114666945404476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,1024,5120,0.009125333693292405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,768,7168,0.010496888723638324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,768,32,0.005950222412745158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,2560,7168,0.009822222093741098
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,1536,10240,0.016002666619088914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,3584,768,0.005692444327804778
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,8192,32,0.00556355549229516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,2048,3072,0.015899555550681222
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,65536,6144,0.054027554061677724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3584,12288,0.018726223044925265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,7168,2560,0.07744710975223117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,256,32,0.005568888866239124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,16384,4096,0.016140444411171805
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,5120,3072,0.007300444775157505
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,256,2560,0.013795555465751223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1536,8192,0.01236088905069563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,2048,16384,0.04663822385999891
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,256,16384,0.016211556063758004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,3072,3072,0.01623288955953386
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2048,6144,0.00867733359336853
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,4096,12288,0.018425777554512024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,10240,2560,0.028071999549865723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,12288,256,0.006291555447710886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,7168,2560,0.007302222152551015
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,12288,16384,0.15160532792409262
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2560,2560,0.00832444429397583
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,3584,10240,0.01609688831700219
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,2048,10240,0.039911111195882164
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,5120,512,0.04259555538495382
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,128,10240,0.014479110638300577
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,4096,3072,0.009006222089131674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,6144,32,0.0021866667601797315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1024,2560,0.00489155575633049
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,7168,16384,0.26954221725463867
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2048,32,0.007664889097213745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,64,3072,0.005600000007285013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,512,2048,0.014145778285132514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,65536,2560,0.13358399603101942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,65536,512,0.012460444536474017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,128,256,0.005611555443869696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,256,6144,0.00959822204377916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,1536,256,0.0052435555391841466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3584,10240,0.015999111864301894
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,256,1536,0.006081777728266186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1024,768,0.006002666635645761
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,8192,2560,0.007609777980380588
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,512,512,0.00564177789621883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,64,12288,0.0052826665341854095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,3584,256,0.0074551113777690465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,1536,10240,0.011761777930789523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,768,3072,0.007317333585686154
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,1536,7168,0.03169599837727017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,512,2048,0.006616000086069107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,512,32,0.005648889061477449
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,768,5120,0.008348444269763099
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,65536,1024,0.01773066653145684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,4096,2560,0.007373332977294922
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,1024,4096,0.009008000294367472
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,512,2048,0.021220443977249995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,512,65536,0.028229332632488672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,256,768,0.005629333357016246
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,16384,256,0.002973333415057924
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,256,7168,0.017667555146747164
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,8192,3072,0.008835555778609382
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,4096,10240,0.016543999314308167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,256,12288,0.013457777599493662
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,6144,4096,0.03071911136309306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,6144,16384,0.061662223603990346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,64,7168,0.005548444473081165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3072,1536,0.00453333349691497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,512,65536,0.062262223826514355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,768,1024,0.007349333001507654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,3072,128,0.005692444327804778
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,6144,256,0.031297776434156634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1024,10240,0.012002666791280111
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1024,2560,0.008464888566070134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,65536,1536,0.01940000057220459
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3072,2048,0.006232888748248418
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,8192,128,0.014169777433077494
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,3072,6144,0.00923999978436364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,128,16384,0.016209777858522203
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,4096,512,0.006252444452709622
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,64,2048,0.006670222100284364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,12288,512,0.00885333369175593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,256,512,0.006312888943486744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,5120,512,0.005625777774386936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,12288,2048,0.017474666237831116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,7168,10240,0.020536889632542927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,12288,256,0.006647999915811751
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2048,2560,0.007340444458855524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,16384,65536,0.29288355509440106
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,6144,2048,0.0306328899330563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3072,3072,0.00943022221326828
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2560,10240,0.014182221558358936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,10240,8192,0.0222053329149882
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1024,5120,0.013224889006879596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,7168,768,0.0035502223504914176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,2560,8192,0.02517866591612498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,7168,256,0.0052577778697013855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,4096,5120,0.01033688916100396
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,16384,16384,0.20578222804599336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,5120,8192,0.02270222206910451
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2048,768,0.0032879999942249725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,1024,16384,0.021293333835071985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,3584,5120,0.031035555733574763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,65536,16384,0.3019591172536214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2048,6144,0.009686222506894005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,256,65536,0.04827910992834303
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,768,32,0.010769777827792697
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,16384,3072,0.01776088939772712
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,512,3072,0.022118222382333543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,8192,256,0.006263111200597551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,8192,3584,0.027311111489931744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,4096,2560,0.0083155557513237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1536,3584,0.007728888756699032
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,16384,2560,0.03224622209866842
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,64,7168,0.024783111280865137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3584,7168,0.013478221992651621
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,3584,512,0.014427555931939019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,8192,2048,0.007650666766696506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2048,768,0.004029333177540037
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,128,128,0.005615111026499007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,4096,512,0.015841777125994366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,2048,128,0.014519999424616495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,7168,2048,0.006973333656787872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,32,7168,0.013083555632167392
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,2048,768,0.02128622266981337
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,4096,1024,0.014144889182514615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3072,6144,0.012718222207493253
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,128,1024,0.006014222072230444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,3584,12288,0.016865778300497267
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,128,16384,0.016131555040677387
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1024,2048,0.004644444419278039
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,65536,1536,0.020264888803164165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,7168,7168,0.021379555265108745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,8192,3072,0.011743110915025076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,1536,12288,0.013451555536852943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,128,7168,0.01032977799574534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3072,10240,0.16153155432807073
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,128,7168,0.022294221652878657
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,6144,1536,0.017243555850452848
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,16384,5120,0.01752088963985443
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,32,3072,0.005583999885453119
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3072,10240,0.01555999947918786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,6144,2560,0.00739822205570009
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,4096,3072,0.007709333466158972
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,8192,12288,0.03291199935807122
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,128,2560,0.022674666510687932
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1536,2048,0.006893333461549547
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,3584,128,0.005600888695981767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,7168,8192,0.02581866747803158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2048,128,0.005655999812814925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,6144,6144,0.016498666670587327
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,128,65536,0.05337688989109463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,4096,5120,0.01629955569903056
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,16384,3072,0.03638489047686259
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,1024,128,0.01575377749072181
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,512,2560,0.015111999379263984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,5120,16384,0.04043466515011258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,8192,16384,0.20472799407111275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,32,1024,0.0029004443850782183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,64,65536,0.0455075568623013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,2048,512,0.0138568886452251
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,768,3072,0.005592888842026393
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,4096,65536,0.06730666425493029
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,7168,7168,0.015844444433848064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,16384,3584,0.040047112438413836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,16384,65536,4.50749757554796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,12288,768,0.02749066717094845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,768,7168,0.011755555868148804
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,65536,768,0.18588088618384468
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,32,128,0.002586666701568498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,8192,768,0.0059057776298787855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,128,32,0.005902222047249476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,16384,1024,0.027218666341569688
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,128,1536,0.014694222145610385
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,5120,6144,0.017447999782032438
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,5120,16384,0.025292444560262892
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,16384,12288,0.07564089033338758
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,5120,64,0.005560889012283749
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,512,128,0.006958222223652734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,10240,5120,0.16649066077338323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,10240,768,0.00489777781897121
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,6144,4096,0.019476445184813607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,8192,2560,0.008968888885445064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,65536,12288,0.23467111587524414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,2560,7168,0.02648799949222141
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,3584,3072,0.047467556264665395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3072,256,0.005879999862776862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,7168,6144,0.061421334743499756
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,6144,5120,0.013968888256284924
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,256,512,0.013096888860066732
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,8192,6144,0.013411555853154926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,4096,12288,0.02001688877741496
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,768,4096,0.006641777853171031
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,512,1536,0.0038373333712418876
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,12288,3072,0.016233777006467182
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,1536,7168,0.03481511274973551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,7168,3584,0.03618222143914964
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3584,2048,0.006654222392373615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1024,3584,0.007677333222495184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,8192,12288,0.020631111330456205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2560,6144,0.011034666664070554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,2048,12288,0.028167112006081477
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,1024,65536,0.07437955670886569
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,256,5120,0.00902844468752543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,5120,7168,0.014056000444624158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,128,128,0.0021911110315057966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1024,10240,0.01367288910680347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,32,5120,0.011924444801277585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,65536,768,0.21826577186584473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,128,2048,0.040792889065212674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,16384,1024,0.01200177768866221
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,16384,2560,0.01752444439464145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3072,8192,0.05041422115431892
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2560,12288,0.018403554956118267
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,64,12288,0.02696888811058468
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,3072,768,0.016851555969980028
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,10240,3072,0.03402933478355408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,2048,5120,0.0299928883711497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3072,32,0.002534222271707323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,7168,7168,0.015882665912310284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2560,2048,0.004615111069546806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,1536,2048,0.025947555899620056
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,10240,12288,0.034741332133611046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,16384,3072,0.0233297778500451
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2048,2048,0.006268444574541516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,65536,768,0.01723911033736335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,32,3584,0.00738666703303655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,1024,7168,0.009736888938479954
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2048,128,0.0021964444054497615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,4096,16384,0.033907555871539645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,128,256,0.007654222349325816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,4096,3072,0.007300444775157505
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,64,768,0.007003555695215861
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,10240,2048,0.023192889160580103
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3072,7168,0.01386666629049513
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,128,5120,0.00906311141120063
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,3072,7168,0.011993777420785693
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,65536,12288,0.2418959935506185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,8192,768,0.011247999966144562
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,8192,2560,0.04264355699221293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,12288,12288,0.04857688811090258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,3584,128,0.015208888385030957
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,7168,512,0.00425155543618732
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,16384,256,0.0111191107167138
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,6144,5120,0.016601777738995023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,65536,2048,0.047282665967941284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,256,768,0.01757422255145179
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,8192,8192,0.038945777548684016
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,3584,1536,0.01903555625014835
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,3072,1024,0.0059164443777667145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,5120,512,0.00286311118139161
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,12288,32,0.0028435554769304064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,128,1536,0.006670222100284364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,10240,3072,0.013217777841620974
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,65536,5120,0.1645173364215427
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,512,128,0.005693333430422678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,8192,10240,0.02181511123975118
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,256,256,0.0025271111064487034
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,16384,768,0.007640889121426477
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,3584,2048,0.01908177799648709
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1024,32,0.002161777681774563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,8192,65536,0.0814817746480306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,1024,5120,0.008423999779754216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,1024,10240,0.011769777370823754
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,6144,768,0.017804443836212158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,6144,5120,0.01792711185084449
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,512,7168,0.009714666340086196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,768,256,0.014813333749771118
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,12288,768,0.0069777775141927935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,4096,2560,0.009020444419648912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,128,1536,0.006594666590293248
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1536,6144,0.054409778780407377
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,2560,1536,0.04579910967085096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,1536,256,0.015232000086042615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,128,16384,0.006224888894293044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,512,128,0.003975999852021535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,3072,128,0.014472888575659858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,128,3584,0.007301333049933116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,7168,10240,0.05419022176000807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,128,6144,0.010374222364690568
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,5120,128,0.002571555475393931
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,3584,3072,0.025065778030289546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,768,8192,0.007747555772463481
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,768,512,0.005955555372767978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3584,256,0.0025173332542181015
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3072,10240,0.04192355606291029
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,65536,6144,0.07600266403622098
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,256,256,0.005896889087226655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,7168,1024,0.006268444574541516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,12288,3584,0.02823022339079115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3584,12288,0.06320177846484713
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,12288,2560,0.011668444507651858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,7168,128,0.015113777584499784
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,128,6144,0.01275111072593265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,128,2560,0.004961777892377642
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2560,1024,0.0032462223122517266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,64,2048,0.004562666846646203
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,512,16384,0.02819733487235175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2560,6144,0.04052088989151849
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2048,3584,0.008803555534945594
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,7168,1536,0.006599111275540457
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,65536,10240,0.2155991130405002
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,65536,3584,0.16735733879937065
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,2048,3584,0.09650222460428874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,12288,1536,0.011782222323947482
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,6144,3072,0.007631111476156447
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,512,5120,0.010035555395815108
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,6144,2048,0.023001778456899855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,2048,2560,0.006945778098371293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,4096,3584,0.025087111526065405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,4096,12288,0.017151999804708693
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,16384,768,0.010407999985747868
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,7168,4096,0.008983111215962304
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,3072,2560,0.0069422221018208405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,12288,128,0.002206222257680363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,10240,8192,0.08645155694749619
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,6144,5120,0.023005333211686876
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,5120,3584,0.016373333003785875
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,8192,2048,0.013543999857372709
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,12288,1024,0.019066666563351948
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2048,7168,0.011391110718250275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,3072,128,0.018187556001875136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,2048,5120,0.0730480021900601
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,12288,1024,0.03470044334729513
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,768,1536,0.004955555415815777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1536,16384,0.0192284451590644
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,6144,65536,0.13887733883327907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,65536,2560,0.027878221538331773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,256,7168,0.019519999623298645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1024,3072,0.007308444215191736
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,2048,128,0.016344888342751395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1024,2560,0.0107342220014996
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,65536,512,0.057743999693128795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,12288,12288,0.04983999994066027
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,7168,256,0.0033200000309281876
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,2048,4096,0.025183111429214478
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,32,8192,0.004587555511130227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2048,2048,0.014157333307796054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,16384,10240,0.06854399707582262
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2560,64,0.0022257777551809945
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,128,2048,0.004911111046870549
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,3072,256,0.015174221661355762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,8192,5120,0.030626667870415583
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3072,4096,0.05297422077920702
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,2048,8192,0.03055466545952691
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,10240,10240,0.03993955585691664
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,32,8192,0.01069155583779017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,32,7168,0.004927111168702443
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2560,5120,0.0353724459807078
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,768,3072,0.007304000357786815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,3072,7168,0.11090844207339817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,256,7168,0.009378666679064432
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,5120,768,0.016894222961531747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,4096,6144,0.026698667142126296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,768,5120,0.006979555719428592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,768,16384,0.016149333781666227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,512,12288,0.025799999634424847
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,1536,1024,0.015153777268197803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,1024,3584,0.007471111085679796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,6144,1024,0.009240888886981541
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,7168,16384,0.07943644126256307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,768,5120,0.023765333824687537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,7168,32,0.005519111123349931
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,3072,1024,0.01516266663869222
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,1536,8192,0.05659911367628309
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,6144,65536,0.12852356168958876
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3584,8192,0.014932443698247274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,6144,32,0.01072355525361167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,2560,10240,0.03308799862861633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,1024,3584,0.015992000699043274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,5120,256,0.0031999999450312722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,6144,32,0.0026933333526055017
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,2560,3584,0.007995555798212687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3072,512,0.01904088921017117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,10240,768,0.0069511110583941145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,256,65536,0.03169155451986525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,512,128,0.007931555310885111
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,12288,4096,0.03194933467441135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,8192,3584,0.011370666325092316
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,12288,5120,0.06330133146709867
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,4096,4096,0.011023999916182624
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,3584,768,0.02420266634888119
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,2560,256,0.006322666588756774
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,2560,5120,0.03234311276011997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,16384,7168,0.044257776604758374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,6144,256,0.016178665889634024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,32,1024,0.0028568889117903183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2560,32,0.0018808889306253856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,4096,12288,0.030072887738545735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,2560,2048,0.016189333465364244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,4096,768,0.003879111260175705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,7168,12288,0.01871555546919505
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,128,6144,0.0090506664580769
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,128,32,0.00217333332531982
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,12288,12288,0.05329422156016032
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,7168,6144,0.015230221880806817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,7168,12288,0.0188417782386144
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,6144,2048,0.030541333887312148
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3072,1536,0.006635555376609166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,12288,5120,0.03982133335537381
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3072,64,0.002499555548032125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,6144,5120,0.012723555167516073
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,64,1024,0.006317333214812809
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,1536,2560,0.0069351109365622205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2048,65536,0.06051555607053968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,12288,5120,0.025617778301239014
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,2048,8192,0.010764444039927589
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,4096,12288,0.017130666308932833
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3072,768,0.006072888771692912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,7168,8192,0.03443466623624166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3072,5120,0.06549421946207683
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3584,5120,0.009930666950013902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,768,2048,0.009386666946940953
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3072,7168,0.04670222269164192
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,12288,65536,0.2506248950958252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,16384,8192,0.025729777084456548
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,768,4096,0.007786666353543599
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,8192,6144,0.028535998529858057
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,256,65536,0.024488889508777197
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3072,1536,0.004261333495378494
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,5120,128,0.007340444458855524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,3584,128,0.015096889601813423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,4096,4096,0.009335999687512716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,6144,1536,0.009670221971140968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,256,3584,0.007716444631417592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,512,512,0.005603555589914322
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,128,65536,0.06959022416008843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,3072,12288,0.030069334639443293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1024,8192,0.01037599974208408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,128,2560,0.014597333139843412
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,12288,512,0.0069644442862934535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2048,12288,0.013999111122555204
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,8192,3584,0.025046222739749487
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,512,7168,0.03150666753451029
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,16384,5120,0.020032889313167997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3072,1536,0.04897511005401611
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,6144,3072,0.05120889014667935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,12288,7168,0.027064889669418335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,6144,12288,0.03784888982772827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,4096,512,0.0040693332751592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,10240,16384,0.05238577723503113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,12288,128,0.016166221764352586
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,512,128,0.002141333288616604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,3072,5120,0.0243857784403695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,768,12288,0.01346666696998808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,512,4096,0.013752000199423896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,2560,3072,0.023391111029518977
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,2048,2048,0.018591110905011494
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2048,4096,0.009726222190592024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2560,32,0.005993777679072485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,4096,1536,0.005960000058015187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,512,65536,0.5210293134053549
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,128,10240,0.011774222056070963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,16384,256,0.006647999915811751
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,5120,256,0.005478222337034013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,8192,6144,0.013791110780504016
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1536,6144,0.09661155276828343
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,4096,64,0.002014222244421641
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,512,256,0.002510222295920054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,4096,5120,0.012905778156386482
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,2048,5120,0.008411555654472774
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,8192,6144,0.02898933490117391
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,128,128,0.0052639999323421055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,512,1024,0.0036071112586392295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,3072,16384,0.02202044427394867
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,12288,256,0.012171555724408893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,3584,3584,0.007376000285148621
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,32,65536,0.013817778064144982
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,5120,2560,0.030682666434182063
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,256,768,0.00562755556570159
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,4096,128,0.014552889598740472
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,5120,1536,0.004304000073009067
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,128,3072,0.015235554840829639
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,65536,16384,0.16524444686041936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3072,6144,0.011316444310877057
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,7168,2560,0.023221333821614582
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,128,6144,0.009390222529570261
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,5120,8192,0.028902222712834675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,4096,10240,0.01680266691578759
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,3072,8192,0.02598222262329525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,4096,256,0.016146666473812528
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,6144,4096,0.014635556273990206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,16384,10240,0.0337413317627377
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1024,64,0.005956444475385878
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,3072,768,0.018974221414989896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,10240,512,0.006735999965005451
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,4096,6144,0.035982221364974976
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,10240,3072,0.056271996763017446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,8192,8192,0.01568711135122511
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,8192,256,0.005251555393139522
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,128,8192,0.02013777858681149
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,2560,256,0.0052915554907586836
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,8192,3584,0.027664888236257765
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,1536,4096,0.007675555845101674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,512,4096,0.005886222339338726
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,256,768,0.014090667168299357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,256,5120,0.030185778935750324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,10240,10240,0.02548444436656104
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,1024,8192,0.028877334462271795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,16384,256,0.0038862224254343244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,12288,2560,0.012475555141766867
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,4096,6144,0.041181332535213895
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1024,2560,0.007585778004593319
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,1024,65536,0.0789404445224338
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,6144,256,0.005909333212508096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,8192,65536,0.0830391115612454
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,7168,2560,0.011431111229790581
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,12288,1536,0.008372444245550368
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3072,768,0.003104000041882197
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1024,5120,0.009010666774378883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,2048,512,0.010620444185203977
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,512,6144,0.008987555901209513
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2560,2048,0.006261333409282897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,65536,5120,0.12898043791453043
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,10240,3584,0.01477866702609592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1536,64,0.006767999794748094
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,512,256,0.005628444254398346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,6144,512,0.0032213332338465583
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,32,5120,0.004541333350870344
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2048,5120,0.049357334772745766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1536,128,0.006210666563775804
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,256,256,0.0021840000732077495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,1536,256,0.012327111429638334
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,4096,8192,0.02539199921819899
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,16384,12288,0.050267553991741605
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,3584,7168,0.014727999766667684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,6144,2048,0.007257777783605788
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,128,2048,0.006264888991912206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3072,3072,0.007607999775144789
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,10240,512,0.0069102222720781965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,2048,1024,0.005962666538026597
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,10240,3072,0.012072888513406118
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,512,2560,0.005487999982304043
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,128,128,0.0021555556191338436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,128,4096,0.0058986664646201665
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2560,512,0.0035377778112888336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,6144,3584,0.008328888979223039
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,65536,10240,0.23707111676534018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,8192,3072,0.02827289038234287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,10240,7168,0.01905155513021681
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,8192,5120,0.01420177850458357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,64,16384,0.006641777853171031
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,3072,768,0.005919111271699269
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,5120,3584,0.027423111928833857
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,1024,6144,0.1229226721657647
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2560,64,0.002790222151411904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2560,65536,0.35577599207560223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,512,65536,0.12237511740790473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,10240,256,0.006537777682145436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1024,1536,0.005951111101441913
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,32,8192,0.021726222501860723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,256,6144,0.01722400055991279
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1536,10240,0.014119999276267158
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,6144,7168,0.02223644488387638
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,256,8192,0.013088000317414602
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,4096,2048,0.006608888920810487
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,128,6144,0.017873777283562552
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,7168,65536,0.14958755175272623
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,768,65536,0.39248087671067977
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,4096,2560,0.006970666348934174
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,12288,3584,0.01519733336236742
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,1024,65536,0.06256533331341214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,10240,10240,0.049250665638181895
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,2560,10240,0.027468444572554693
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1536,64,0.0021386666016446217
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,12288,5120,0.041768001185523145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3584,6144,0.018459555175569322
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,256,8192,0.0299893319606781
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,12288,256,0.0026604444202449587
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,16384,768,0.027003554834259882
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,8192,16384,0.02661688956949446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,256,32,0.008408000071843466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,12288,2048,0.02770400047302246
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,2048,3072,0.018940443793932598
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,768,10240,0.06683288680182563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,5120,3072,0.010566222170988718
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,6144,2560,0.0229751103454166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,256,32,0.0018168888572189545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,1536,6144,0.008983111215962304
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,7168,256,0.020249777370029025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,768,7168,0.025316443708207872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,2048,4096,0.007622222105662028
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,1024,1024,0.015804444750150044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,128,512,0.011025778121418424
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,512,16384,0.03051111102104187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,8192,256,0.005970666805903117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,3072,5120,0.02335022224320306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,65536,8192,0.10890311002731323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,65536,16384,0.2871093220180935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,65536,10240,0.10835822423299153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2560,768,0.005347555710209741
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,10240,512,0.022729777627521094
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1536,7168,0.01035288886891471
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,3584,10240,0.048194666703542076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,10240,6144,0.01775911119249132
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,5120,7168,0.012136889000733694
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,10240,1536,0.00777333312564426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,65536,3584,0.04464177621735466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,1024,1536,0.008654221892356873
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3584,32,0.0021848889688650766
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,16384,10240,0.033067554235458374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,128,5120,0.008579555484983657
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,512,16384,0.02683466672897339
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,12288,3072,0.07539466354582045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,5120,12288,0.024061333802011278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,512,8192,0.012079111403889127
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,65536,7168,0.1511546638276842
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,128,512,0.002577777744995223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,16384,7168,0.03551822238498264
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,3584,10240,0.015486222174432544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,5120,1024,0.028967999749713477
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,6144,6144,0.01991911066903008
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,6144,1024,0.06349688768386841
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,6144,10240,0.03219644559754266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,10240,65536,0.1416204505496555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2048,64,0.002521777732504739
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,128,3072,0.005296000176005893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,65536,4096,0.09476800097359551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,256,256,0.00831911133395301
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2048,2048,0.004693333473470476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,256,768,0.002855111120475663
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,32,128,0.002149333349532551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,4096,5120,0.14853510591718885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,5120,65536,0.07557066943910387
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,5120,64,0.005929777605666055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,12288,10240,0.04134666588571336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,64,3584,0.004931555440028508
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,12288,128,0.015580443872345818
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,5120,16384,0.06004799736870659
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3072,2560,0.007638221813572778
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,5120,768,0.023733332753181458
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,5120,768,0.016926222377353244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1024,3584,0.00599911105301645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,768,12288,0.052312887377209134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,1536,768,0.017278222574128043
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,768,256,0.0052764444715446895
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,512,768,0.0032400000426504347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3072,1536,0.006896889044178857
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,10240,128,0.008004444340864817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,128,2560,0.009640889035330879
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,8192,2560,0.011357333097192975
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2560,256,0.005888000130653381
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,1536,768,0.005698666804366642
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,768,256,0.00249777775671747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,64,1024,0.003216888962520493
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,768,3072,0.005587555468082428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,128,256,0.012763555679056378
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1024,5120,0.0069751110341813825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,64,7168,0.018075555562973022
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,768,2560,0.0070186663005087115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,2560,12288,0.017080000705189176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,128,256,0.005510222166776657
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,6144,768,0.05167644553714328
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,32,10240,0.012447111308574677
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,8192,3584,0.012264000044928657
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,8192,768,0.006232888748248418
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,3072,768,0.005613333235184352
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,7168,3584,0.06109422445297241
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,256,64,0.005673777725961473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,5120,7168,0.019342222147517733
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,768,10240,0.022287999590237934
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,7168,8192,0.017419555120997958
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,16384,768,0.006238222122192383
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,2048,65536,0.10489688979254828
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,256,1024,0.003212444484233856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,32,10240,0.02663199934694502
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3584,64,0.004889777965015835
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,12288,768,0.01945955554644267
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,256,5120,0.014936889211336771
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,128,16384,0.03937955697377523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,256,64,0.006030222194062338
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,7168,2560,0.02422044508987003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1536,4096,0.007799111306667328
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,512,8192,0.011170667078759937
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,3584,4096,0.008064000142945183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,10240,64,0.006024888820118374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,7168,10240,0.02648177742958069
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,8192,512,0.005984888722499211
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,65536,1536,0.3409893247816298
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,12288,16384,0.05884889099333021
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,4096,5120,0.024940444363488093
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,8192,256,0.008774222599135505
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,5120,512,0.03984355595376756
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,128,8192,0.12815111213260227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,16384,7168,0.05359911256366306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,2560,16384,0.02035288843843672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,12288,8192,0.5878382258945042
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,2048,16384,0.02040088839001126
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,2048,10240,0.014498665928840637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,3584,65536,0.06653777758280437
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3072,256,0.002558222247494592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,512,256,0.007658667034573025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,65536,3584,0.06534400251176622
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,7168,10240,0.034075554874208235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1536,2560,0.005569777968857024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,32,768,0.005666666560702854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,768,512,0.0031635556370019913
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,256,2560,0.01592177814907498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,768,1024,0.014704000618722705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,512,1536,0.01348533315791024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,32,1536,0.010086221827401055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,10240,768,0.0069617778062820435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,256,5120,0.012786666552225748
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,6144,7168,0.02196088929971059
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,2048,768,0.005655999812814925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,10240,768,0.006632888896597757
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,65536,2048,0.21005511283874512
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,16384,6144,0.03197244471973843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3072,2560,0.010074666804737514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,8192,10240,0.019433778193261888
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,2560,2560,0.03432888785998026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,2048,65536,0.145142224099901
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1536,2048,0.004961777892377642
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,128,8192,0.005902222047249476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1536,3584,0.03110755483309428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,512,2048,0.004839111119508743
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1536,8192,0.012818666795889536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,7168,1536,0.006219555520349079
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,16384,256,0.03416533271471659
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,3584,32,0.005952000204059813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3072,64,0.019949333535300363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,16384,10240,0.05519022213088142
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,10240,4096,0.037208000818888344
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,12288,4096,0.014542222023010254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,8192,512,0.00700444479783376
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,10240,8192,0.02307022280163235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3072,1536,0.006313777632183499
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2048,32,0.005557333429654439
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,5120,1024,0.017875555488798354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,256,2048,0.006991110742092133
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,64,3584,0.007939555578761632
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,3072,32,0.006627555522653792
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,65536,4096,0.22750666406419542
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3072,6144,0.011521778172916837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3072,256,0.0028213332924577924
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,12288,5120,0.01621866722901662
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2560,768,0.007298666569921706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,7168,1024,0.013250666360060373
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2048,2048,0.004292444636424382
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,1024,4096,0.05182400014665392
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,8192,16384,0.026234666506449383
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2560,12288,0.08326933119032118
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,4096,1536,0.005986666513813867
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,2560,5120,0.009588444398509132
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,6144,3072,0.009498666558000777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,2560,4096,0.007962666451931
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,12288,16384,0.48569689856635195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,12288,1024,0.007420444654093848
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,5120,10240,0.01665422154797448
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,2048,3072,0.04950044552485148
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,128,16384,0.01590399940808614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,512,256,0.0052871112194326185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,12288,1536,0.010036444498433007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,2048,16384,0.021714667479197185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,16384,4096,0.3051831192440457
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,256,2048,0.006344888773229387
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,12288,3584,0.04047377904256185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,128,6144,0.022384888596004907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,5120,1024,0.004951111144489712
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,4096,8192,0.01332355539004008
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,1024,2048,0.0063324446479479475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,6144,4096,0.02661955522166358
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,768,512,0.005647999958859549
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2560,3072,0.007990222010347579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,512,16384,0.01625955601533254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,65536,4096,0.039213332864973284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,512,3584,0.00601244428091579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,128,768,0.0028702221396896574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,768,3584,0.006523555351628198
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,768,8192,0.019965334071053397
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,3584,10240,0.03221155537499322
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,65536,3072,0.07405600282880995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,6144,65536,0.07581066422992282
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1536,6144,0.00943911075592041
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,1536,1536,0.006301333506902059
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,256,256,0.005290666802061929
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3072,1024,0.003922666526503033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,64,32,0.006597333484225803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,3072,6144,0.011056000159846412
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,2048,10240,0.02757155564096239
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,12288,12288,0.06934755378299289
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,16384,6144,0.024114666713608637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,2560,7168,0.03170311119821336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,768,2560,0.015549333559142219
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,128,65536,0.023938667443063524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,128,768,0.00591466658645206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,768,256,0.006268444574541516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,10240,128,0.007337777978844113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,16384,512,0.012486221889654795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,6144,1536,0.020015110572179157
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,4096,8192,0.015428443749745687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,16384,16384,0.04574222366015116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,3072,12288,0.017138666576809354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,16384,32,0.008339555727110969
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,256,6144,0.012081777883900536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,1024,768,0.005559111220969095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,256,16384,0.016192888220151264
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3584,7168,0.013782222237851886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,4096,6144,0.011997333003415002
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,512,512,0.013108444710572561
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,256,128,0.0151591118839052
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1024,2048,0.007311111523045435
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,5120,256,0.01960355540116628
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3584,768,0.005598222215970357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2560,4096,0.007977777885066139
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,768,4096,0.007711999946170383
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,768,128,0.005548444473081165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,768,256,0.00527555536892679
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,65536,2560,0.04767733481195238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1024,12288,0.025054223007626
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,6144,3072,0.035755554835001625
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,10240,8192,0.06890844636493258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,1024,5120,0.018539556198649935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,8192,1536,0.00794222205877304
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1536,1024,0.011062222222487131
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,10240,512,0.006288888967699475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1536,128,0.00555288874440723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,1536,10240,0.011830222275522021
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,256,32,0.006237333433495627
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,64,65536,0.04556711183653938
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,16384,5120,0.020639111598332722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,768,1024,0.003175111073586676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,768,3072,0.06436977783838908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,512,8192,0.010346666806273991
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,12288,12288,0.04982577761014303
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,12288,2560,0.011425777441925473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,768,512,0.013105777402718862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,5120,32,0.005848888721730974
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,16384,3584,0.09584977891710068
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,12288,256,0.0028942223224374983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,7168,128,0.0052426668504873914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,16384,12288,0.03552088803715176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,2560,1536,0.03578311204910278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1024,6144,0.007039111521508958
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,512,5120,0.015293333265516492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,2048,512,0.006587555425034628
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,128,1536,0.006252444452709622
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,64,3584,0.004905777672926585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,1536,256,0.002191999927163124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,4096,8192,0.01498133275243971
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,32,64,0.005945777727497949
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,5120,768,0.009187555147541894
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,3584,2048,0.006311111152172089
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,1536,256,0.006973333656787872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,512,768,0.01626044511795044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,4096,2560,0.007615110940403408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,5120,3072,0.14859910806020102
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,4096,7168,0.028901331954532202
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,4096,6144,0.013580444786283704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1536,3584,0.008395555946562026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,1024,1536,0.004108444270160464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,8192,1536,0.021195555726687115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,128,6144,0.03089511063363817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,6144,512,0.01625333395269182
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,768,1536,0.013927999469969006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,1536,768,0.005966222120655908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,12288,256,0.01754399968518151
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,5120,32,0.0018728888697094387
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,32,12288,0.013168888787428537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,5120,128,0.005568888866239124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1024,3072,0.007695111135641734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,2048,5120,0.020000888241661918
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1536,12288,0.016920000314712524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,1024,2560,0.007336888876226213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,8192,6144,0.017228444417317707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,10240,2048,0.010283555421564314
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,1024,6144,0.00962400022480223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,4096,32,0.002185777864522404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,256,8192,0.010693333215183683
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,1024,768,0.023709333605236475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,8192,7168,0.04567111200756497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,3072,5120,0.008356444537639618
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,10240,768,0.007694222033023834
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2048,256,0.0046364445653226645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,16384,8192,0.08075200186835395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,2560,7168,0.022466666168636743
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,6144,12288,0.01867644488811493
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,4096,768,0.0029724445194005966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,512,4096,0.00564177789621883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,5120,12288,0.04252444373236763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,10240,6144,0.024284443921513025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,768,3072,0.00703111125363244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,256,32,0.0022408889813555614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2560,256,0.0025235555238193935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,6144,2560,0.02411288850837284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,128,1024,0.00601333338353369
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1536,4096,0.034550223085615374
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,512,2560,0.007631999750932057
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,256,1024,0.005979555762476391
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,6144,128,0.0025395556456512874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,2560,7168,0.009753777749008602
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,768,16384,0.011208888557222156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,4096,32,0.002209777840309673
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,4096,256,0.015978667471143935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,512,3584,0.007668444679843054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,10240,7168,0.051889777183532715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,512,4096,0.011417778001891242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,128,5120,0.00941955546538035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,256,128,0.0021973333011070886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,4096,2560,0.007053333024183909
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,16384,2048,0.02233155568440755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,768,16384,0.0210506667693456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,5120,65536,0.6809626685248481
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,256,12288,0.013270222478442721
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,7168,10240,0.02281066609753503
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,128,3584,0.02514755560292138
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,10240,6144,0.03228800164328681
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,65536,2560,0.02736088964674208
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,65536,768,0.01386044422785441
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,128,12288,0.013560000393125745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,8192,128,0.005577777822812398
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,256,7168,0.005295111073387994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,5120,3584,0.012895999683274163
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,128,1024,0.006312888943486744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,768,128,0.002500444443689452
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,512,65536,0.05595466825697157
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2048,768,0.0032799999333090256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,3584,12288,0.038260443343056574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,7168,16384,0.13253422578175864
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,3584,512,0.005954666684071223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1024,10240,0.010846222440401712
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,1024,1024,0.013215111361609565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,32,32,0.0019128889673286015
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,64,1024,0.0032088889016045462
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,128,2048,0.007164444360468123
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,16384,3072,0.02676711148685879
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,512,16384,0.15353955162896052
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,16384,3584,0.09308089150322808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3072,1024,0.00979733301533593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,6144,1024,0.004432888908518685
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,8192,1024,0.01868888901339637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,2048,2048,0.006643555644485686
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,65536,1024,0.03114933437771267
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,3584,256,0.0052577778697013855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,4096,6144,0.012508444488048553
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,1024,2560,0.015788444214397006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2048,8192,0.010025777750545079
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,8192,2560,0.011838222543398539
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,7168,5120,0.06455644634034899
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2048,6144,0.009361777868535783
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,10240,65536,0.15137599574195013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,7168,4096,0.02553422252337138
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1024,256,0.0025759999536805684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,5120,256,0.016920889417330425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,12288,4096,0.0473582214779324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,256,10240,0.03776444329155816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,2560,1024,0.006416000011894438
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,5120,65536,0.07050222158432007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,1024,8192,0.020234665936893888
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,2560,768,0.016149333781666227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,2048,5120,0.008271999657154083
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,16384,512,0.013357333011097379
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,512,256,0.01907911068863339
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,8192,7168,0.035616000493367515
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,12288,1024,0.022015111313925848
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,8192,5120,0.01440355512830946
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,3072,128,0.014464888307783337
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,5120,65536,0.11206044091118707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,256,768,0.013783111340469785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,256,3072,0.007651555869314406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,5120,1024,0.006264888991912206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3584,2560,0.008624888956546783
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2560,64,0.0021466666625605687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,256,65536,0.2787368827395969
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3072,5120,0.010646222366227044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,64,256,0.0021768889079491296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,8192,12288,0.04127644499142965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2560,3584,0.04187377625041538
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,256,2048,0.00867733359336853
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,65536,1024,0.02141511109140184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,2560,512,0.014144000079896716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,10240,5120,0.018792889184421964
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,10240,5120,0.172017772992452
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,768,3584,0.07307200299368964
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3584,10240,0.01796711153454251
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,5120,16384,0.023102222217453852
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,7168,10240,0.03754933343993293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,5120,10240,0.02366311185889774
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,5120,256,0.016501333978441026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,8192,8192,0.04055022199948629
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,10240,16384,0.4943155712551541
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,3072,4096,0.030592888593673706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,6144,8192,0.017421333326233756
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,8192,16384,0.05753778086768257
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,128,1536,0.0036391110883818734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,16384,10240,0.045736889044443764
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2048,12288,0.014840000205569796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,128,768,0.019305777218606737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,12288,7168,0.020065777831607394
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,16384,256,0.015876443849669564
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,65536,10240,0.1481040053897434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,256,10240,0.02069244450993008
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,4096,65536,1.7731929355197484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,256,768,0.013880000346236758
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,2048,4096,0.012112889024946423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,768,256,0.01518311103185018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,3584,3072,0.008585777547624376
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,10240,4096,0.015839111473825242
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,65536,7168,0.15238577789730495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2560,7168,0.011520889070298938
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,1024,65536,0.10245955652660793
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,16384,6144,0.024995555480321247
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,64,16384,0.015562666787041558
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,10240,128,0.0032622222271230486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,256,4096,0.01570577753914727
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3072,64,0.0022124445272816554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,16384,3584,0.019542222221692402
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3072,16384,0.02382133404413859
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,2560,7168,0.009735110733244155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,256,7168,0.01830933325820499
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,512,2048,0.013767110804716745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,128,2560,0.006949333681000604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,2048,2048,0.020025778147909377
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3072,128,0.009181333084901174
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,3584,1536,0.006275555739800136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3072,32,0.005565333283609814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2560,4096,0.008055999875068665
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,12288,1536,0.02700888944996728
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,4096,3072,0.007309333317809635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2560,12288,0.04083022144105699
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,65536,256,0.039162665605545044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,12288,4096,0.014673777752452426
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,6144,768,0.00598933340774642
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,32,32,0.002181333386235767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,512,512,0.006966222491529252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,12288,10240,0.19751200411054823
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,1536,256,0.005323555734422471
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,8192,4096,0.013737777868906656
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,768,3584,0.007385777930418651
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,8192,256,0.007338666253619724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,256,32,0.0019244444039132858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,10240,256,0.01760444376203749
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,10240,10240,0.04893777767817179
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,256,32,0.005408000200986862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,2048,3072,0.007300444775157505
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2560,12288,0.014446222119861178
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,64,128,0.006759111252095964
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,4096,128,0.005923555543025334
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,256,5120,0.009025777379671732
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,64,512,0.005886222339338726
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,6144,768,0.027958220905727808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2560,6144,0.06605955627229479
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,4096,128,0.003911111089918348
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,64,5120,0.004608889006906086
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,512,3584,0.006256000035338932
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,3584,12288,0.01776088939772712
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,64,8192,0.01718133356836107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,7168,3072,0.011336888704035016
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,5120,12288,0.027849776877297297
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,512,2048,0.005511111269394557
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,12288,10240,0.05826933516396416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2048,7168,0.0791084435251024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1024,1536,0.004237333519591225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,8192,16384,0.05120266808403862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,16384,7168,0.04475377665625679
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,128,3072,0.015078221758206686
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,32,5120,0.006073777874310811
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,10240,4096,0.0833946665128072
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,7168,64,0.0022337778160969415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,128,128,0.005944889038801193
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,512,12288,0.04098755452367995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,4096,7168,0.01828444500764211
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,512,768,0.00566133318675889
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,12288,3072,0.01460977726512485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,2048,2560,0.018263111511866253
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,4096,3072,0.019125334090656705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1536,8192,0.01235466698805491
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,4096,10240,0.01482311056719886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,4096,7168,0.012830221818553077
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,3072,4096,0.015840889679061044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1536,16384,0.030328889687856037
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,768,512,0.013222222526868185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,65536,3072,0.0875048902299669
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,7168,64,0.0028515555378463534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,5120,8192,0.07450755437215169
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,128,2048,0.0086666668454806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2560,12288,0.01927288870016734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,12288,7168,0.03640266590648227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,3072,1536,0.01754044493039449
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,7168,1024,0.017493334081437852
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,4096,2560,0.045904888047112354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,768,256,0.005936000082227919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,768,256,0.005621333503060871
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,2560,5120,0.009384888741705153
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,256,768,0.0053839998112784485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,2560,128,0.014145778285132514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,256,5120,0.01568266749382019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,4096,5120,0.011157333023018308
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,5120,1024,0.04698488778538174
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,128,1024,0.006301333506902059
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,256,5120,0.04093955622778998
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,10240,32,0.0069431112044387394
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1536,65536,0.05853244331147936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1024,128,0.0022231110682090125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,512,10240,0.0209840006298489
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,7168,12288,0.0387733346886105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3072,7168,0.011033777561452655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1536,1536,0.006555555595291986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,12288,1536,0.008636444807052612
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,4096,128,0.014923555983437432
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,8192,64,0.00258933338854048
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,5120,6144,0.013169777890046438
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,4096,7168,0.014236445228258768
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,128,10240,0.011787555283970304
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,768,768,0.006007110906971826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2048,64,0.009398221969604492
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2048,65536,0.17421778043111166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,3072,512,0.015491555134455362
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,6144,65536,0.07628266678916083
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,10240,768,0.007054222126801808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,16384,5120,0.017816889617178176
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,10240,8192,0.9713581932915581
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,512,4096,0.008345777789751688
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,128,2048,0.014864888456132678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,8192,1024,0.023212444451120164
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,1024,4096,0.0090515555606948
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,1536,16384,0.030292444758945044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,1024,5120,0.016598222984208
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,6144,2560,0.0177208897140291
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,3584,10240,0.04087911049524943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,256,256,0.0022248888595236675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,32,6144,0.005244444641802046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,1536,4096,0.007812444534566667
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,65536,512,0.013429332938459186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,7168,3072,0.010575999816258749
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,10240,10240,0.30334043502807617
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3072,5120,0.017331555485725403
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,12288,3584,0.07280621925989787
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,6144,5120,0.02236355510022905
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,1536,6144,0.018311111463440787
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,768,256,0.005939555664857228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2048,5120,0.007012444237867991
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,12288,32,0.007080888582600488
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,256,64,0.0018622221218215094
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,256,512,0.0052595556610160405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,512,16384,0.013736888766288757
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,7168,3584,0.05956355730692545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,512,12288,0.014156444205178155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,16384,4096,0.023806222611003455
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,6144,1536,0.025565332836574976
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,3584,128,0.014613333675596448
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3584,7168,0.012464000119103326
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,16384,128,0.006773333168692059
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,4096,6144,0.018741332822375827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,2560,128,0.005985777825117111
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,16384,12288,0.07499555746714275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,3584,1536,0.006529777828190062
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,5120,6144,0.014576888746685453
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,128,32,0.0018915554715527429
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1536,1024,0.006351999938488007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2560,10240,0.01564800077014499
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,16384,3072,0.013481777575280933
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1536,3584,0.03190844588809543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,512,12288,0.02369333306948344
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,65536,4096,0.7963475651211209
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1536,12288,0.019155555301242404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3584,16384,0.027814222706688776
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,8192,3072,0.05170399943987528
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,10240,768,0.020633776982625324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,16384,2560,0.07338666915893555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,16384,256,0.0024862223201327855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,1024,512,0.008237332933478886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,8192,7168,0.022132444712850783
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1024,128,0.00217333332531982
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,7168,1536,0.02162488963868883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,128,8192,0.01035644445154402
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,4096,3072,0.0069226668112807805
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,768,2048,0.00424888895617591
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,5120,1536,0.008250666989220513
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2560,2048,0.005560889012283749
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,8192,3072,0.03394133183691237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,32,6144,0.0059573331640826324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,7168,256,0.003206222214632564
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2048,65536,0.06443555487526788
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,1024,12288,0.015757333901193406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,4096,10240,0.04054488738377889
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,256,6144,0.009340444372759925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,7168,1536,0.02036711076895396
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,256,2048,0.014415999253590902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,10240,3072,0.03228622343805101
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,128,12288,0.02236622240808275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,64,768,0.0058871110280354815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,10240,2560,0.028320001231299505
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,128,10240,0.052728000614378184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,4096,2560,0.055531554751926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,8192,8192,0.03981511129273309
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,3584,7168,0.01108266661564509
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2048,65536,0.04809777935345968
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,128,1024,0.00941066692272822
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,512,6144,0.0090515555606948
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,1024,5120,0.02851822310023838
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,12288,2560,0.20324088467491996
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,10240,512,0.00798044436507755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,64,3584,0.010831111007266574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,256,3072,0.014130666851997375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,128,1536,0.03537333342764113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2048,6144,0.009385777844323052
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,5120,256,0.005977777971161737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2560,32,0.0029484445436133277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,1536,128,0.006687110910813014
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,4096,2560,0.009371555513805812
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,3584,2560,0.008409777449236976
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,2560,3072,0.020618667205174763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,32,3072,0.007250666618347168
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,8192,16384,0.026424888107511733
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,3584,3584,0.0075937774446275495
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,1024,2560,0.0069306666652361555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,3072,512,0.03323910964859857
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,3584,3584,0.0076657773719893555
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,4096,16384,0.021799999806616042
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,512,3072,0.0052355556852287715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2560,5120,0.009008000294367472
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,1024,1024,0.013776889277829064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,768,7168,0.010091555615266165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,1024,5120,0.008336000144481659
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,64,4096,0.007735999921957652
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3584,512,0.0028453332682450614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,6144,8192,0.017254221770498488
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,768,512,0.002871111035346985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,3072,1536,0.006244444598754247
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2048,1536,0.008640000389681922
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,768,8192,0.020630222227838304
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,10240,7168,0.0199937770764033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2048,4096,0.0083155557513237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,2048,2048,0.061936000982920326
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,3072,2048,0.016107555892732408
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,6144,2048,0.022965333527988855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,5120,256,0.005604444278611078
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,6144,6144,0.015977778368526034
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,10240,3584,0.030209776428010728
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,8192,5120,0.07370044125450982
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,5120,16384,0.021819555097156104
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,256,2560,0.014147554834683737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,3584,65536,0.06690933307011922
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,10240,4096,0.03942755526966519
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,5120,10240,0.026955554882685345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,5120,256,0.005948444621430502
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,512,768,0.002867555452717675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,12288,3584,0.07105066378911336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,8192,16384,0.04050222370359633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,4096,65536,0.11468799908955891
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,7168,32,0.002712888850106133
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,4096,65536,0.06643111175960965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2048,12288,0.011726222104496427
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,16384,12288,0.8826319906446668
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,768,5120,0.019256000717480976
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,768,10240,0.011920889218648275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,768,2048,0.00444088876247406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1536,7168,0.01291911138428582
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,5120,5120,0.012710221939616732
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,1024,10240,0.032221333848105535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,7168,3072,0.0103502223889033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2048,2560,0.004881777697139316
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,256,256,0.0052177777720822235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,8192,256,0.006778666542636023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,4096,6144,0.01606399979856279
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1024,3584,0.007643555601437886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,768,16384,0.014246222045686511
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3072,12288,0.019298666053348117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,3072,16384,0.022297778063350256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,2560,256,0.005283555636803309
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,5120,12288,0.13232444392310247
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,8192,2048,0.08156266477372912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,65536,7168,0.14516354931725398
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,32,3072,0.007283555964628856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,2048,2560,0.020587555236286588
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2048,1536,0.006639110959238476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,12288,1536,0.02624711063173082
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,1536,6144,0.02195288903183407
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,32,10240,0.016550223032633465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,7168,65536,0.08334400256474812
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,512,128,0.005473777651786804
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,128,8192,0.03965422179963853
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3072,12288,0.09280533260769314
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,256,4096,0.011599111060301462
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,768,65536,0.2528826660580105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,1024,6144,0.019263111882739596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,128,12288,0.03495555453830295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,16384,5120,0.027109333210521277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,12288,2560,0.040645334455702044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,5120,3072,0.0340924428568946
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1536,5120,0.008948444492287105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,3072,2048,0.006291555447710886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,512,12288,0.02308622168170081
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2048,3584,0.007613333563009898
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,1024,256,0.014488000008794995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,4096,16384,0.04212266537878248
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,128,32,0.006913777854707506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,128,256,0.005199999858935674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,3072,4096,0.008344888687133789
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,1536,5120,0.00830666638082928
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,7168,512,0.01682488951418135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,128,64,0.001810666587617662
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,128,256,0.006248888870080312
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,3584,3072,0.007682667010360294
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,1536,2048,0.007976888782448238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,10240,512,0.0191920002301534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,7168,65536,0.14763200283050537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,512,8192,0.021298666795094807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,4096,12288,0.048511111074023776
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,1024,2560,0.01681066718366411
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,768,6144,0.017858665850427415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,7168,256,0.005931555396980709
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,768,65536,0.021200888686709937
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,32,16384,0.0052871112194326185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,6144,1536,0.005389333185222413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1536,2048,0.006310222049554189
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,3584,5120,0.031962666246626116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,6144,128,0.0090844440791342
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,5120,3584,0.008284444610277811
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,8192,128,0.014801777071423002
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,65536,1536,0.22650043169657388
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,512,7168,0.02029511166943444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,4096,65536,0.07019555568695068
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,5120,12288,0.029489778810077246
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,7168,256,0.005913777897755305
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,2560,128,0.01553422212600708
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,256,5120,0.00656444455186526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,8192,1024,0.006267555471923616
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,512,16384,0.011095999843544431
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,1536,8192,0.013776000175211163
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,2560,1024,0.005952000204059813
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,8192,1536,0.020259555843141343
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,32,2560,0.007332444190979004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,64,1024,0.00319377767542998
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,768,768,0.0028817777832349143
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,65536,1024,0.017310221989949543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,2048,1536,0.006894222150246303
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,6144,64,0.010254222485754225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,65536,512,0.08463555574417114
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,2560,1024,0.010168000227875179
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,3584,8192,0.025799999634424847
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,1536,12288,0.05336088935534159
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2560,32,0.002879111096262932
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,4096,5120,0.00869422240389718
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,3072,10240,0.15148711204528809
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2048,1536,0.0053093334039052325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,16384,1024,0.011445333560307821
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,65536,256,0.009743111001120673
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,8192,7168,0.1482435597313775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,16384,16384,0.04888711041874356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,2560,512,0.005598222215970357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,1024,1536,0.014135999812020196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,8192,2560,0.025054223007626
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,5120,12288,0.059115555551317006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,8192,64,0.025055999557177227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,512,8192,0.011401777466138205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,1024,128,0.002407999916209115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,10240,2560,0.02481866710715824
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,5120,1536,0.0068488890926043195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,65536,512,0.020056888461112976
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1536,2048,0.007002666592597961
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3584,128,0.005615111026499007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,768,10240,0.008338666624493068
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,2048,7168,0.008700444466537898
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,4096,16384,0.050739556550979614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,65536,256,0.024085332949956257
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,256,128,0.005597333527273602
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2560,32,0.005751111027267244
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,10240,2560,0.014126222994592456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,3072,768,0.007000000112586551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,5120,7168,0.012131555212868584
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,6144,4096,0.012055111428101858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,2560,128,0.005324444423119227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,10240,8192,0.03115644388728672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,3072,7168,0.012544000314341651
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,7168,10240,0.01762488815519545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,128,1024,0.006049777898523543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,10240,1536,0.009275555610656738
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,768,12288,0.012480888929631976
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,32,512,0.002542222125662698
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,7168,3584,0.010198222266303169
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,6144,2048,0.03920266694492764
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,768,3072,0.0076453329788313965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2048,3584,0.009706666900051964
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,4096,6144,0.012303999728626676
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,10240,256,0.017290666699409485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,16384,3072,0.23118310504489473
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2560,64,0.010495999621020423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,4096,1024,0.006604444649484422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2560,4096,0.01620533400111728
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2048,1536,0.004810666872395409
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,6144,3072,0.17565777566697863
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,256,3072,0.01773066653145684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,12288,32,0.0062871111763848206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,10240,2048,0.011411555939250521
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,6144,768,0.007281777759393056
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,6144,12288,0.5323395729064941
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1024,65536,0.23440800772772896
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,3072,2048,0.10317155387666489
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,1024,128,0.0052435555391841466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,3072,512,0.002862222285734283
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,7168,16384,0.18212444252438018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2560,3072,0.013360000318951078
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,5120,12288,0.02407199972205692
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,12288,1024,0.011514667007658215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,7168,3072,0.009892444643709395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,1536,7168,0.024273778001467388
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,64,2048,0.011522666447692446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,10240,12288,0.0679253339767456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,2560,1536,0.017515555024147034
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1536,2048,0.005596444424655702
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,16384,8192,0.02520977788501316
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,8192,10240,0.03039466672473484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,4096,128,0.014494222071435718
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,10240,16384,0.04058400127622817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,16384,8192,0.8381351364983453
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,256,8192,0.01794488893614875
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3584,16384,0.02534666657447815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,768,3584,0.01964000033007728
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3584,4096,0.013288000391589271
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1536,16384,0.028798222541809082
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,2048,512,0.029282665914959375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,65536,8192,0.09173511134253608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,512,3072,0.007332444190979004
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,5120,16384,0.043402665191226535
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,1536,3584,0.022470222579108343
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,32,512,0.005747555444637935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,4096,10240,0.17730400297376844
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,256,512,0.019017777509159513
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,2560,6144,0.0210826661851671
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,12288,5120,0.016262221667501662
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,2560,2048,0.019367999500698514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,32,2560,0.009769777456919352
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,512,1536,0.014013333453072442
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,12288,6144,0.02776177724202474
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,65536,3584,0.09851555691825019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,65536,768,0.24144977993435332
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,256,512,0.011454222102959951
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,4096,768,0.02574488851759169
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,16384,32,0.002982222164670626
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,8192,2048,0.007327111230956183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,6144,128,0.007659555309348636
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,2560,3072,0.0063182223174307085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,64,64,0.001863111017478837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,6144,6144,0.028449777099821303
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,256,2048,0.006288888967699475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,5120,4096,0.027656000521447923
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,4096,768,0.003951111187537511
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,3072,16384,0.09099911318884955
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,512,2048,0.00673333348499404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,256,768,0.006693333387374878
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,1536,16384,0.045976890458001025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,10240,12288,0.04154400030771891
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,12288,256,0.006311111152172089
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,6144,64,0.006631999793979857
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,8192,256,0.032263110081354775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,8192,10240,0.026955554882685345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,65536,2560,0.05051555567317539
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3584,5120,0.011223999990357293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3072,32,0.0052800000541739995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,3072,3072,0.01891999940077464
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,65536,128,0.020744888318909537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,4096,3072,0.027136888768937852
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,4096,2048,0.01721244388156467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,3072,4096,0.007699555820888943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,512,768,0.015846222639083862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,3584,5120,0.008363555702898238
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,16384,512,0.0069715554515520734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,16384,1536,0.029913776450686987
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,3584,65536,0.13496622774336073
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2560,512,0.003173333282272021
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,12288,10240,0.026361778378486633
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,8192,12288,0.08835555447472467
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1024,2048,0.00866844422287411
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,12288,768,0.006940444310506185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,3584,768,0.005611555443869696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,5120,8192,0.024895111719767254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2048,8192,0.011778666741318174
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,65536,3584,0.042543109920289784
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,16384,3072,0.014817777607176038
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,65536,2048,0.22915556695726183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,4096,6144,0.023181334137916565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3584,32,0.0025537777692079544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,12288,1024,0.006298666612969504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,8192,1536,0.006600889066855113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,256,1536,0.003579555700222651
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,8192,16384,0.028791109720865887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,65536,2048,0.047664887375301786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2560,1024,0.0038853333228164246
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,16384,6144,0.03699377841419644
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,5120,3072,0.007705777883529663
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,7168,6144,0.014478221535682678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,10240,6144,0.034373333056767784
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,4096,128,0.002152888932161861
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1024,1536,0.0063288890653186375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2048,65536,0.060043553511301674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,8192,4096,0.010491555763615502
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,64,3584,0.007784888976150089
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,768,2048,0.008368888662921058
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,65536,8192,0.09153422382142808
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,10240,3584,0.015973332855436537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,512,8192,0.012496000362767113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,32,1536,0.0041893331540955436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,5120,2048,0.028038223584493
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,16384,512,0.0069982219073507525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,8192,1024,0.008199999729792276
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,768,7168,0.0486968888176812
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,7168,256,0.017498667041460674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,128,2048,0.009119110802809397
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,5120,2048,0.006667555620272954
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,5120,512,0.011208888557222156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,5120,7168,0.013458666702111563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,7168,32,0.0025315555847353404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,10240,512,0.009330666727489894
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,5120,8192,0.029218667083316382
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2048,3072,0.007268444531493717
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3072,8192,0.017370666066805523
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,256,3072,0.013841778039932251
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,65536,6144,0.11966221862369114
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,7168,128,0.02033066584004296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,7168,6144,0.014992000328169929
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3072,1024,0.0052968888647026485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,65536,1024,0.0811671084827847
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,1024,128,0.005335111171007156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,3584,128,0.00536088893810908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,3072,65536,0.09599910842047797
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,6144,5120,0.013818666338920593
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,12288,8192,0.02718844513098399
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1536,512,0.002550222186578645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,1536,3584,0.08347200022803412
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1536,10240,0.011932444241311816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,768,512,0.002838222309947014
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,512,3584,0.005591111050711737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,256,1024,0.0031617778456873367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,2048,8192,0.026541334059503343
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,7168,10240,0.016146666473812528
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,16384,2560,0.01279022213485506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,512,128,0.013062222136391534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,512,12288,0.008959111240175035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,64,64,0.0018115554832749898
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,5120,4096,0.019275556008021038
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,768,4096,0.007732444339328342
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,6144,512,0.005571555760171678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,768,2048,0.006522666662931442
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2560,65536,0.07320000065697564
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,6144,6144,0.017846221725145977
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3072,1536,0.014909333652920194
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,32,2048,0.004558222161398994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,4096,5120,0.017298666967286002
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,7168,768,0.060402664873335093
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,16384,1536,0.012147555748621622
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,10240,128,0.002526222210791376
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,256,65536,0.1918017731772529
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,256,6144,0.038366221719317965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3584,256,0.002833777831660377
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,4096,6144,0.013071999781661563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,3584,2048,0.020094222492641874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,6144,3584,0.009335999687512716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,512,8192,0.01071111112833023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,12288,512,0.008325333396593729
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,16384,4096,0.018906666172875296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,512,7168,0.01036177741156684
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,512,1536,0.006256888724035687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,256,16384,0.006264888991912206
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2048,5120,0.007290666302045186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,768,128,0.00572622236278322
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,4096,64,0.0022737777067555324
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,1024,4096,0.018233777748213876
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,3072,256,0.014819555812411837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,1024,3584,0.0199626667631997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2560,256,0.005984888722499211
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,5120,8192,0.1068720022837321
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3584,3584,0.011086222198274402
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,128,768,0.00591822216908137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,512,2560,0.005021333280536863
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,3072,6144,0.17664978239271376
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1024,1536,0.005562666803598404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,3072,5120,0.008211555580298105
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,7168,512,0.005608888963858287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,512,7168,0.009272000028027428
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,256,4096,0.006269333263238271
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,256,512,0.005374222166008419
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,1024,4096,0.008022222254011366
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,7168,6144,0.03715733355946011
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,6144,768,0.005631999837027655
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,5120,4096,0.017654221918847825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,65536,8192,0.1735964483684964
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2560,1536,0.008424888882372115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1024,8192,0.008932444784376357
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,7168,2560,0.010022222167915767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,256,3584,0.007974222302436829
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,256,64,0.005311111195219888
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,3584,128,0.014896889527638754
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,64,16384,0.015170666906568738
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,65536,32,0.012438221938080259
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1024,10240,0.033424887392256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,12288,8192,0.0391813317934672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,2560,32,0.007999999655617608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,3072,8192,0.0140364451540841
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,6144,8192,0.019650666250122916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1536,256,0.00259911103381051
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,65536,128,0.02527911133236355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2048,2048,0.006575111299753189
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,512,64,0.007100444700982835
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1536,256,0.009764444496896531
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,32,64,0.005255110975768831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,3584,3584,0.00816177742348777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,256,4096,0.007718222008811102
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,1536,1024,0.020799999435742695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,12288,7168,0.03415022293726603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,10240,16384,0.055531554751926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,65536,2048,0.084297776222229
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,512,10240,0.054508444335725575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,6144,2560,0.010775110787815519
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,256,16384,0.027843554814656574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3072,16384,0.021002666817771062
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,32,2048,0.015565334094895257
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,256,5120,0.009309333231714036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,512,16384,0.03464711043569777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,10240,4096,0.016122667325867545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,128,128,0.005602666487296422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,8192,768,0.005944889038801193
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,3072,16384,0.034471111165152654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,1024,4096,0.015909334023793537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3584,3072,0.007678222325113084
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,768,6144,0.009020444419648912
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,256,1536,0.012344000240166983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1024,5120,0.008850666383902231
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,256,128,0.006682666639486949
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,12288,1024,0.027894222074084814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,65536,512,0.05713599920272827
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,8192,5120,0.03184355629814996
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,4096,3072,0.009698666632175446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,4096,1024,0.01571733256181081
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,1024,6144,0.02844711144765218
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,5120,7168,0.015629332926538255
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,256,768,0.00563822231358952
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,256,7168,0.016361777981122334
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,3584,256,0.0052728888889153796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,3072,5120,0.008346666892369589
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,16384,10240,0.03695288962788052
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,10240,3072,0.02624711063173082
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,7168,7168,0.019800888167487252
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,4096,4096,0.010518222219414182
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,1536,65536,0.06240533457862007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,2048,3584,0.02356622285313076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,128,512,0.005327111317051782
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,7168,16384,0.041342222028308444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,128,512,0.005576000031497743
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,65536,64,0.005296000176005893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1024,5120,0.00868711123863856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,4096,4096,0.029754665162828233
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,10240,5120,0.49186309178670246
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2048,3584,0.011774222056070963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,12288,10240,0.0481608874268002
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,256,1024,0.016140444411171805
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,3584,1536,0.01722044414944119
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,32,32,0.0018240000224775737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,65536,2048,0.04487199915779961
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,65536,32,0.014354666074117025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,6144,3584,0.01111644423670239
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,6144,1536,0.02109333376089732
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,3584,65536,2.576706568400065
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,768,3584,0.00757155567407608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,5120,16384,0.03657955593532986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,1024,6144,0.01850755512714386
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,512,7168,0.020943111843532983
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,256,8192,0.02987288766437107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,10240,1024,0.027237332529491846
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,256,64,0.005883555445406172
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3584,2560,0.0069653333889113525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,256,2048,0.004242666479614046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,16384,8192,0.037426667080985174
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,16384,2048,0.01907911068863339
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,6144,7168,0.014913777510325113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,5120,10240,0.029771556456883747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,10240,768,0.008557333714432186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,12288,512,0.020261334048377145
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,1536,512,0.005595555735958948
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,12288,3584,0.01757066614098019
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,16384,768,0.018942221999168396
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,12288,2560,0.019036443697081674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,65536,1024,0.024618667032983568
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,1024,16384,0.01609422266483307
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,8192,2560,0.008613333106040955
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,5120,8192,0.01815111107296414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,65536,32,0.006948444578382704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,3072,6144,0.09749866856469049
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,8192,768,0.024088889360427856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,10240,128,0.006599999964237213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,256,2560,0.006923555499977536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,256,256,0.002475555572244856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,3584,10240,0.01649599936273363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,3072,2048,0.005679111099905438
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3584,2048,0.006594666590293248
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,65536,12288,1.5025919808281794
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,4096,8192,0.01348977784315745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,1536,10240,0.011874666644467248
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,8192,2048,0.01387288918097814
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,16384,2560,0.012777778009573618
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,8192,16384,0.06044266621271769
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,16384,6144,0.022950222094853718
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,4096,128,0.005923555543025334
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,4096,7168,0.012682666381200155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,768,10240,0.03615199857287937
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,128,1536,0.0041537777417235905
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,128,128,0.011346666349305047
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,3584,768,0.018717333674430847
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,128,1536,0.00721866637468338
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,32,1536,0.0035582222044467926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,1536,128,0.005227555417352253
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,1536,256,0.005599110904667113
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,512,1024,0.005931555396980709
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,512,1024,0.005892444401979446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,2048,5120,0.0299271113342709
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,10240,256,0.04295377930005392
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3584,1536,0.011490667031870948
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,3584,1536,0.006268444574541516
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,8192,1536,0.006530666516886816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,64,2048,0.006673777682913675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,2560,256,0.015174221661355762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,4096,512,0.005983110931184556
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,256,2048,0.00490399988161193
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,10240,256,0.006291555447710886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,32,6144,0.010063110954231685
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,512,3584,0.015849777393870883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,1536,2560,0.00737955586777793
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,6144,256,0.006625777731339137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,65536,16384,0.9170106252034506
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,65536,8192,0.26481954256693524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,6144,10240,0.031839112440745033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2048,128,0.002862222285734283
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,128,256,0.01538133290078905
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,2560,128,0.02436977790461646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,3072,16384,0.02065333392884996
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,10240,65536,3.981261359320747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,8192,6144,0.028640889459186133
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,4096,10240,0.030995554394192163
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,16384,768,0.011220444407727985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,16384,8192,0.30320533116658527
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,256,3584,0.01459822224246131
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,4096,5120,0.010229333407349056
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,32,8192,0.005290666802061929
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,2560,3584,0.0073902226156658595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,65536,5120,1.2725519604153104
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,65536,768,0.026084444589085046
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,256,512,0.019855999284320407
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,256,16384,0.036887112591001726
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,7168,128,0.005642666584915585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1536,8192,0.020744888318909537
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,4096,2560,0.007062222394678328
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,256,768,0.015623110863897534
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,256,7168,0.0053297777970631914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,65536,5120,0.055768887201944985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,128,2560,0.005271111097600725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,512,8192,0.03142222099834018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,4096,16384,0.24368890126546225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,2048,64,0.002199999988079071
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,1536,512,0.01462044484085507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,16384,1536,0.027637332677841187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,256,5120,0.00868088917599784
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1024,1536,0.00637511122557852
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,2560,12288,0.028054220808876887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,2560,768,0.019297776950730216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1024,4096,0.007984889050324758
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,512,768,0.0059075554211934405
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,10240,3072,0.011749332977665795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,12288,1024,0.026610665851169165
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3072,256,0.002850666642189026
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,768,1024,0.01352533366945055
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2048,65536,0.06028088596132067
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,768,512,0.005245333330498801
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,512,7168,0.005642666584915585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3584,10240,0.014513777361975776
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,1024,2048,0.008359111017651027
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,4096,128,0.0052968888647026485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,6144,8192,0.024449777272012498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,12288,256,0.01791200041770935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,10240,768,0.007113777928882175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,768,6144,0.009725333087974125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,128,1024,0.0032275555034478507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,2048,128,0.013458666702111563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1024,5120,0.0052897776994440295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,65536,10240,0.18335466914706758
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,2048,5120,0.019253333409627277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,3072,1536,0.017599110802014668
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,10240,7168,0.06405422422620985
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,256,6144,0.009193778038024902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,4096,8192,0.014112888111008538
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,512,768,0.005626666463083691
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,5120,32,0.002671111168132888
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,64,10240,0.03232889042960273
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,768,64,0.005610666755172942
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,6144,6144,0.240485323799981
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,5120,8192,0.031911998987197876
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2048,64,0.005586666779385672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,768,3584,0.007983999947706858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,1024,768,0.03146933184729682
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,65536,256,0.06004533502790663
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,6144,10240,0.018901333212852478
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,128,12288,0.0052177777720822235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,3584,768,0.016968000266287062
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,512,7168,0.010000000397364298
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,16384,10240,0.2478026549021403
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,512,65536,0.09951200087865193
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,768,3584,0.017895110779338412
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,7168,768,0.005928888916969299
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,2560,768,0.022895110978020564
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,768,8192,0.07206044594446818
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,768,8192,0.010543111297819348
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,16384,768,0.007258666886223688
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,12288,128,0.006637333167923822
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,12288,5120,0.06762666834725274
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,5120,12288,0.02599999970859951
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,512,8192,0.010771555205186209
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,8192,2048,0.008357333640257517
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,2560,256,0.005345777918895085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,8192,128,0.015243555108706156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,128,10240,0.012459555433856117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,2048,5120,0.017356443736288283
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,256,512,0.0053342220683892565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,65536,1024,0.06306666798061795
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,768,512,0.017855111095640395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,512,2560,0.014628445108731588
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2560,5120,0.011556444068749746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1536,768,0.006021333237489064
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,768,1536,0.006640000061856375
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,1024,2560,0.007350222104125553
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,10240,1536,0.023296000228987798
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1536,16384,0.012067555553383298
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,3584,16384,0.033997333712048
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2048,5120,0.009083555804358589
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,12288,768,0.007328000333574083
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,512,512,0.0026195556339290407
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,7168,256,0.017251556118329365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,16384,768,0.013535999589496188
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,16384,10240,0.05498844385147095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,8192,2048,0.008793777889675563
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,4096,2048,0.020765332712067496
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,2560,768,0.016546666622161865
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,128,128,0.002179555594921112
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,128,3584,0.007722666694058313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,16384,768,0.007839110990365347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,32,256,0.005611555443869696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,6144,5120,0.025426665941874187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,256,8192,0.020592888196309406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,2560,256,0.006631999793979857
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,65536,5120,0.04450400008095635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1024,7168,0.01035377797153261
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,128,3072,0.008303999900817871
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,512,10240,0.01241422196229299
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,32,5120,0.009681777821646797
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,768,6144,0.006241777704821692
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,1024,8192,0.010591111249393888
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,6144,256,0.005597333527273602
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,65536,4096,0.07795466979344685
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,64,7168,0.010128888818952773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,4096,6144,0.012339555554919772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,6144,10240,0.036539554595947266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,512,7168,0.018428444862365723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,128,16384,0.005592000153329637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2048,32,0.006639110959238476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,128,12288,0.021919111410776775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,512,4096,0.022551111049122278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,8192,65536,0.08647822009192573
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,1024,7168,0.010048888623714447
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1024,2048,0.005476444545719359
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,2560,7168,0.19308088885413277
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,512,16384,0.02036711076895396
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,12288,10240,0.7874773343404134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,64,32,0.006351999938488007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,256,12288,0.013153778182135688
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,128,64,0.0018808889306253856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,512,65536,0.08235822121302287
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,5120,5120,0.022648000054889258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,128,512,0.0025253333151340485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,3584,768,0.006284444282452266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,6144,512,0.016213332613309223
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,12288,6144,0.11259288258022732
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,12288,10240,0.04225066635343763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,4096,4096,0.010043555663691627
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,512,4096,0.005991999887757831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,1536,1024,0.0035342222286595237
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,128,2560,0.006945778098371293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,8192,3584,0.06320799721611871
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,128,3072,0.014175999495718213
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,2560,10240,0.025021332833502028
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,768,65536,0.06344711118274264
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,8192,256,0.0058764442801475525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,6144,512,0.0032515554792351196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,65536,32,0.3032746579911974
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,256,1024,0.012070222033394707
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,3072,12288,0.1790462202495999
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,4096,12288,0.030108445220523413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,2048,2048,0.00657155571712388
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,512,10240,0.02167822255028619
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,256,7168,0.0052986666560173035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3072,8192,0.014375999569892883
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,2048,10240,0.027319999204741582
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,10240,1536,0.03655644589000278
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,2560,768,0.006292444550328785
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,10240,16384,0.05421688821580675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1024,2560,0.004939555707905027
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,1536,16384,0.019499555230140686
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,7168,16384,0.027875555886162653
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,7168,128,0.0052391112678580815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,12288,256,0.0028924445311228433
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,256,12288,0.01310488912794325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,1024,768,0.0069671107663048645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,5120,128,0.0024871110088295406
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,12288,6144,0.027458666099442378
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,7168,3584,0.010834666589895884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,7168,1024,0.010093332992659675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,10240,2560,0.012848888834317526
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,65536,512,0.01482488877243466
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,256,65536,0.3117075496249729
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,5120,2560,0.010486221975750394
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1024,256,0.0031804444475306403
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,1536,65536,0.6815102365281848
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,7168,16384,0.02439022229777442
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,7168,10240,0.03769244419203864
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,1536,64,0.0059262220230367445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,4096,1024,0.006592000110281839
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,7168,128,0.009383111364311641
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,8192,16384,0.026901332868470088
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,8192,2560,0.009357333183288574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,128,1536,0.012472000386979846
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,10240,16384,0.057586669921875
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3584,768,0.005574222240183089
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3072,65536,0.06396622127956815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,3584,10240,0.027552000350422327
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,16384,128,0.0058017778727743365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,32,2560,0.004643555730581284
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,512,3584,0.007994666695594788
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,8192,65536,0.12853422429826525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,768,64,0.006729777902364731
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,4096,128,0.01960355540116628
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,64,2560,0.0070159998204973005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,2048,65536,0.06404622395833333
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,512,2560,0.00536088893810908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,256,1536,0.013146667016877068
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,8192,8192,0.025624889466497634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,1024,16384,0.030092444684770372
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,4096,6144,0.01146577795346578
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,16384,8192,0.18464799722035727
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,512,768,0.0028568889117903183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,2560,7168,0.013975999421543546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,128,5120,0.03308355477121141
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2560,32,0.005223999834722943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,2560,1536,0.014525334040323893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,512,16384,0.011490667031870948
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,3584,1536,0.006688000013430913
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,12288,12288,0.06595377789603339
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,128,2048,0.0064897777305708984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,7168,65536,0.12028977606031631
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,6144,768,0.02075199948416816
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2560,2560,0.006625777731339137
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,1024,1024,0.013995556367768182
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,4096,7168,0.018970666660202872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,12288,8192,0.2907786634233263
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,12288,4096,0.015059555570284525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,16384,3072,0.028209778997633193
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,512,3072,0.007324444750944774
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,6144,2560,0.03335288829273648
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,512,512,0.005564444594913059
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,2048,128,0.0023742222951518162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,3072,7168,0.01255644443962309
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,16384,6144,0.03950577643182542
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,3584,5120,0.024604444702466328
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,3072,8192,0.013011555704805585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,2560,6144,0.008656000097592672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,10240,2048,0.009337777892748514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,12288,1024,0.007125332951545715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,3584,12288,0.05482844511667887
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1024,65536,0.024493333366182115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,8192,128,0.006596444381607904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,65536,32,0.004239999999602635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,256,16384,0.03327644533581204
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,10240,5120,0.01903288894229465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3584,512,0.008292444050312042
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,4096,64,0.0021866667601797315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,12288,7168,0.05167288912667168
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,8192,6144,0.01865600049495697
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,10240,64,0.017277333471510146
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,12288,12288,0.08044977982838948
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,2048,1024,0.027694221999910142
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,4096,6144,0.03397599856058756
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,1536,2560,0.017743999759356182
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,1536,3072,0.020346666375796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,2560,4096,0.02539377742343479
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,6144,256,0.005988444305128521
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,12288,256,0.006304889089531369
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2560,16384,0.02312444481584761
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,64,6144,0.0052737775776121355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,65536,256,0.012865777644846173
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,2048,8192,0.009341333475377824
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,8192,2560,0.010124444133705562
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,2048,32,0.0022248888595236675
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,3072,16384,0.1827439996931288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,12288,12288,0.12066133817036946
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,12288,10240,0.031847112708621554
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,8192,16384,0.052207999759250216
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,64,1024,0.006291555447710886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,8192,1024,0.02366311185889774
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,256,6144,0.030611554781595867
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,8192,768,0.004559999952713649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,10240,3584,0.017502221796247695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,2560,2048,0.03037422233157688
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,5120,1536,0.006276444428496891
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,10240,1024,0.009916444619496664
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,768,1536,0.006481777876615524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,6144,4096,0.025452444950739544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,12288,1024,0.009033777647548253
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,64,12288,0.014327110515700446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,5120,65536,0.07473778062396579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,12288,4096,0.05271377828386095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2560,256,0.005940444353553984
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,6144,2560,0.020791111720932853
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,512,1024,0.01315822203954061
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,512,64,0.0021946666141351065
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,2560,1024,0.003913777983850903
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,256,6144,0.018255111243989732
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,128,3072,0.007323555648326874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,256,512,0.002530666689078013
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,128,256,0.005946666830115848
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,4096,128,0.015174221661355762
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,7168,32,0.006344888773229387
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2560,8192,0.0148026661740409
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,16384,32,0.0029022221763928733
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,2560,768,0.018002667360835604
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,4096,7168,0.024079110887315538
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,4096,1024,0.015810666812790763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,4096,2048,0.007297777467303806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,256,32,0.006676444576846228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,16384,1536,0.029327998558680218
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,12288,5120,0.03128266665670607
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,3072,256,0.0053688887920644544
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,64,1024,0.0063271112740039825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,65536,1024,0.04351733459366692
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2560,6144,0.011401777466138205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,16384,2560,0.01592177814907498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,5120,1024,0.0058906666106647914
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,512,7168,0.019537778364287484
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2560,65536,0.0613386697239346
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,1024,10240,0.12678578164842394
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,3072,16384,0.019627556204795837
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,12288,12288,0.03616888986693488
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,5120,10240,0.01516533394654592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,2048,32,0.005559111220969095
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,256,768,0.0028320000403457214
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,3072,10240,0.04003644320699904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,512,2560,0.004866666677925321
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,512,128,0.0021706666383478376
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,6144,5120,0.014861333701345654
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,4096,5120,0.18947111235724556
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,3584,12288,0.01756622228357527
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,128,256,0.011597333682907952
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,2560,1536,0.006298666612969504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,3072,1024,0.0032595555401510666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1536,512,0.002943111169669363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,64,128,0.0021404443929592767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,16384,16384,0.04507466819551256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3584,65536,0.10218489170074463
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,8192,32,0.002185777864522404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,6144,6144,0.026728888352711994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,768,2048,0.006636444479227066
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,3072,3072,0.0075697774688402815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,10240,2560,0.10036177767647637
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,32,2560,0.0075191110372543335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,3584,6144,0.00902844468752543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,4096,3072,0.007309333317809635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,16384,10240,0.055312001042895846
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,8192,16384,0.02882666720284356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,32,768,0.005943999936183293
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,512,256,0.002513777878549364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,65536,256,0.0074480002125104266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,512,12288,0.012991111311647626
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,2560,65536,0.17316355970170763
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,16384,6144,0.04161599940723843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,512,512,0.012451555165979596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,7168,5120,0.022616000639067754
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,512,1536,0.003535110917356279
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,512,768,0.014818666709793938
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,768,10240,0.011695110963450538
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,8192,6144,0.013800000150998434
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,16384,2048,0.014506666196717156
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,8192,10240,0.05689777930577596
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,3072,4096,0.12483289506700303
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,32,12288,0.005292444593376583
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,512,3072,0.011154666543006897
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,10240,2048,0.010821333361996545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,3072,32,0.0021564445147911706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,8192,3584,0.011405333048767514
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,1536,1536,0.013519110778967539
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,64,4096,0.01238222254647149
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,5120,12288,0.03668177790111966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,10240,3072,0.02660177813635932
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,128,32,0.005889777921968036
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,12288,6144,0.07745689153671265
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,12288,3072,0.016564443707466125
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,8192,32,0.005621333503060871
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,5120,3072,0.0251893334918552
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,768,2048,0.014448000325096978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,16384,16384,0.9095253414577908
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,768,512,0.0052675555149714155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,1024,256,0.030223111311594646
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,65536,7168,0.0910266637802124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1536,256,0.006647999915811751
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,256,3072,0.015871110889646742
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,12288,8192,0.0244862222009235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,3584,64,0.005963555640644497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,65536,1024,0.018241778016090393
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,16384,65536,0.21039377318488228
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,128,2048,0.013210666676362356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,5120,32,0.005260444349712796
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,1536,768,0.017867555220921833
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,768,3584,0.006298666612969504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,8192,256,0.008340444829728868
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,512,1024,0.012808000048001608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,10240,5120,0.026887999640570745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,12288,2048,0.012379555238617791
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,7168,3584,0.11796977784898545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,1536,65536,0.4693617820739746
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,6144,3072,0.031679110394583807
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,10240,1024,0.00868444475862715
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,6144,1024,0.005958222266700532
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,12288,1024,0.034280889564090304
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,768,128,0.010638222098350525
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,7168,6144,0.01475911173555586
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,5120,12288,0.03861155443721347
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,6144,1024,0.016571554872724745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,1024,8192,0.02549422283967336
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,5120,32,0.005942222144868638
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,3072,2048,0.006934222247865465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,512,3072,0.007302222152551015
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,5120,512,0.0158595558669832
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,2048,3072,0.007132444116804335
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,7168,5120,0.011041777829329172
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,1536,4096,0.00869955536392
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,4096,8192,0.015152888165579902
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1536,10240,0.010777778095669217
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,3072,2048,0.018962666392326355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,4096,128,0.0029004443850782183
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,128,7168,0.005305777821275923
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,2560,8192,0.019097778532240126
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2560,10240,0.011801777614487542
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,4096,512,0.0058062221441004015
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,12288,7168,0.2683466540442573
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,12288,512,0.006672888994216919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,12288,12288,0.21423644489712187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,16384,8192,0.10282399919297959
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,12288,128,0.005923555543025334
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,8192,2048,0.026856889327367146
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2560,7168,0.009962666365835402
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,768,7168,0.02405422263675266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,768,4096,0.019959110352728102
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,128,2048,0.016159110599093966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,10240,512,0.03800800111558702
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,256,256,0.00535111129283905
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,256,16384,0.015481778317027621
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,32,2560,0.00508622204264005
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,32,64,0.0019804444164037704
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,65536,1024,0.02776977750990126
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,10240,2048,0.008993777963850234
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,32,65536,0.0076453329788313965
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,8192,1536,0.007003555695215861
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,10240,6144,0.042916443612840444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,4096,8192,0.02609688871436649
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,8192,256,0.046366221374935575
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,2560,1536,0.01684977776474423
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,512,16384,0.2696133454640706
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,64,8192,0.005327111317051782
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,10240,3584,0.03653600149684482
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,8192,8192,0.019652444455358718
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,2560,16384,0.12140444914499919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,65536,1536,0.07060177458657159
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,16384,1024,0.008655110994974772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,2560,12288,0.028992888000276353
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1536,64,0.0021866667601797315
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,3584,3584,0.008050666915045844
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,512,7168,0.04372622238265144
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,32,768,0.002579555536309878
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,2048,1536,0.011573333707120685
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,768,6144,0.008524444368150499
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1536,512,0.005585777676767773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,7168,8192,0.03138400117556254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,3072,768,0.005624889085690181
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,1536,4096,0.021007999777793884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,16384,64,0.002543111021320025
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,6144,16384,0.11303022172715928
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,768,10240,0.012087999946541257
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,3584,3584,0.008332444561852349
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,2048,12288,0.016843555702103507
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,1024,12288,0.011066666907734342
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,10240,3072,0.04353866643375821
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,7168,65536,0.13186311721801758
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,512,65536,0.06371644470426771
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,256,65536,0.049119111564424306
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,5120,8192,0.018064889642927382
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,12288,2048,0.01053600013256073
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,2048,5120,0.009580444130632613
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,4096,1024,0.005953777581453323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,4096,128,0.014512888259357877
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3072,4096,0.01347555551264021
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,12288,2560,0.011762667033407422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,3584,2048,0.00629688882165485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1536,2560,0.005312888986534542
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,12288,2048,0.02910755409134759
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,32,128,0.0022115556316243275
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,65536,1536,0.02197155521975623
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,3584,7168,0.018882667024930317
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,1536,32,0.0021457777669032416
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,64,128,0.006647111227114995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,7168,6144,0.0120000003112687
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,8192,10240,0.018072000808186002
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,65536,2048,0.07775111330880059
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,8192,12288,0.03286044465170966
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,64,2048,0.007000000112586551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,4096,6144,0.012702222499582501
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,5120,5120,0.028880887561374243
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,16384,10240,0.12916799386342367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,8192,2560,0.07712355587217543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,32,768,0.005978666659858491
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,5120,4096,0.008143111235565608
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,7168,65536,0.07572799921035767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,1024,2048,0.015846222639083862
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,64,1536,0.003608889049953885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,32,16384,0.005413333161009683
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,1536,12288,0.016328889462682936
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,32,5120,0.00925866680012809
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3584,7168,0.012870222330093384
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,65536,2048,0.027870221270455256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,65536,5120,0.05232888791296217
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,65536,16384,4.3000772264268665
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,65536,768,0.06680622365739611
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,10240,10240,0.03717155588997735
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,8192,16384,0.03966044386227926
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,1024,6144,0.009017777939637503
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,2048,3584,0.016922665966881644
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,512,2560,0.01891911029815674
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,1024,2048,0.006702222343948152
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,10240,2048,0.028303109937243994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3584,3072,0.018784888916545443
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,64,12288,0.0057795556883017225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,5120,1536,0.006286222073766921
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,768,12288,0.012450666891203986
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,12288,7168,0.031239999665154353
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,16384,64,0.0021884443445338146
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,65536,64,0.009748444788985783
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,8192,65536,0.09125066465801662
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2048,8192,0.009347555538018545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2048,8192,0.014857777290874057
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,2048,3072,0.007698666718271043
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,768,16384,0.016311999824311998
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,4096,32,0.002208000048995018
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,65536,128,0.018929777873886954
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,2048,1024,0.016168889072206285
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,65536,5120,3.136895921495226
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,12288,2048,0.01909244391653273
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,80,6144,2560,0.007293333609898885
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,8192,3584,0.011693333586057028
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,256,8192,0.02130311065249973
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,8192,6144,0.03600710961553786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,12288,7168,0.0366924438211653
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,3072,2048,0.006299555715587404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,1024,2048,0.015489778584904142
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,1536,4096,0.007262222468852997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,5120,7168,0.027849776877297297
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,3584,1536,0.006334222025341458
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,256,8192,0.019609777463806998
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,6144,16384,0.036333332459131874
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,256,5120,0.00924355536699295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1536,5120,0.008560889297061497
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,128,4096,0.0052177777720822235
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,3072,3072,0.007296000089910295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,16384,3584,0.0384204453892178
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,256,7168,0.005232000102599462
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,512,4096,0.0454693337281545
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,4096,1024,0.0035413333939181436
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,768,768,0.0031786666562159858
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,768,2048,0.014900444282425774
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,1024,768,0.007620444728268518
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,256,3072,0.006953777538405524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,256,65536,0.016525333126386006
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,10240,8192,0.024618667032983568
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,512,65536,0.05899288919236925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,16384,4096,0.38209599918789333
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,2560,5120,0.03393955694304572
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,256,3072,0.0052782222628593445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,6144,6144,0.018589332699775696
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,1024,65536,0.060673779911465116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,1024,3072,0.009451555709044138
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,12288,1536,0.08018666505813599
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,12288,12288,0.0418853329287635
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2560,6144,0.011758222348160215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,5120,3584,0.00871377769443724
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,5120,1024,0.01688799924320645
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,512,128,0.002584888910253843
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,4096,1024,0.015614221493403116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,4096,7168,0.01421955558988783
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,4096,64,0.0021884443445338146
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,128,7168,0.01975288821591271
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,64,256,0.002954666606254048
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,12288,1536,0.009848000274764167
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,6144,3584,0.009344889058007134
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,3072,16384,0.03494400117132399
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,6144,3584,0.009093333449628618
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,256,12288,0.013171555267439948
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,6144,3584,0.008378666308191087
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,12288,7168,0.03423377871513367
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,256,128,0.01311644415060679
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,2560,768,0.013516444298956128
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,1536,4096,0.008249777886602614
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2,7168,2048,0.007113777928882175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3072,3584,0.010464888479974536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,128,2048,0.004905777672926585
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,10240,3072,0.012802666260136498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,256,6144,0.009672889278994666
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,2048,10240,0.015528000063366361
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,6144,32,0.0032666667054096856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,65536,5120,0.8389671113755969
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,256,512,0.018963555494944256
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,10240,6144,0.017613333132531907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1536,10240,0.016120889120631747
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,1536,32,0.0022239999638663395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,7168,16384,0.02586666742960612
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2048,64,0.0019866666860050624
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,3584,768,0.00628266649113761
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,4096,256,0.01649599936273363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,8192,3584,0.011745777395036487
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,2048,12288,0.017252445220947266
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,1536,10240,0.010397333237859938
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,2048,1024,0.015840889679061044
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,128,768,0.0028106667515304354
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2048,2560,0.007264888948864407
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,1024,6144,0.033788445923063494
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,8192,3584,0.03115644388728672
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,2048,12288,0.017128888103697035
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,16384,5120,0.055359999338785805
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,12288,3072,0.03128088845147027
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,3584,2560,0.043452445003721446
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,4096,2048,0.023528888821601868
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1024,16384,1536,0.018400889303949144
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,128,4096,0.004257777912749184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,5120,64,0.028552889823913574
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,3072,4096,0.028027554353078205
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,12288,65536,0.1569733354780409
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,2048,4096,0.009159111314349705
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,8192,7168,0.05667199691136678
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,5120,8192,0.03010666701528761
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,1024,2048,0.006616000086069107
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,12288,64,0.03542755709754096
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4096,512,10240,0.03254133462905884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,6144,1024,0.016897777716318767
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,1536,7168,0.009669333696365356
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,256,768,0.014144000079896716
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,6144,1536,0.020124443703227572
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,16384,1536,0.06337422132492065
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,7168,768,0.003564444267087512
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,65536,7168,0.1512426667743259
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1024,16384,0.017826666434605915
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,512,6144,0.018202667435010273
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,6144,512,0.007295110987292395
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,2048,2560,0.007352888584136963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,3072,256,0.00258577780591117
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,10240,10240,0.04090311129887899
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,5120,5120,0.015301333533393012
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,5120,1024,0.009759999811649323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,1024,2048,0.017251556118329365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,512,1024,0.01313688854376475
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,12288,10240,0.06944711340798272
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,65536,128,0.010062221851613786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,7168,6144,0.028079112370808918
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,6144,3584,0.1648542218738132
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,5120,256,0.002895111011134254
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,64,65536,0.007646222081449296
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,384,3584,7168,0.018623110320832994
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,1536,3072,0.015480889214409722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,12288,768,0.005712888720962737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,65536,10240,0.23218133714463976
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,768,1536,0.01682044400109185
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,7168,64,0.002157333410448498
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,3072,16384,0.04987644486957126
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,256,7168,0.009705777797434065
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,1024,3584,0.0052897776994440295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,1024,1536,0.003953777667548922
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,65536,16384,2.235853407118056
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,16384,64,0.013563555147912769
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,768,3584,0.01145155562294854
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,5120,1536,0.09267910983827378
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,64,64,0.0029928889125585556
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,64,12288,0.008052444292439355
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,1024,2048,0.01657511128319634
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,7168,512,0.016522667474216886
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,5120,7168,0.027132444911532935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,8192,7168,0.01738222274515364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,256,2048,0.006642666541867786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,768,2560,0.00776533368561003
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,256,4096,0.008390222158696916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,5120,5120,0.01660355594423082
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,1536,1536,0.004029333177540037
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,64,4096,0.008206222620275285
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,10240,3584,0.014160888062583076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,5120,8192,0.0576488905482822
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,8192,128,0.017694221602545846
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,512,4096,0.03152444296412998
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,128,7168,0.02545333405335744
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,3072,512,0.002536000063021978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,2048,1024,0.0063688887490166565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,768,1024,0.005497777627574072
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,512,10240,0.032349334822760686
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,2048,7168,0.013751111096805997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,7168,1536,0.017687999539905123
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,512,16384,0.008403555386596257
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,16384,1024,0.008374222450786168
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3584,65536,0.16448622279696995
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,2560,32,0.001812444378932317
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,3072,2048,0.006682666639486949
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,16384,12288,0.2748026582929823
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,6144,256,0.018259555101394653
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,512,64,0.0024906665914588505
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,3584,32,0.00598933340774642
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,5120,64,0.008653333617581261
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,128,5120,0.009047110875447592
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,10240,12288,0.04491022229194641
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,768,2048,0.004562666846646203
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,6144,2048,0.01943288909064399
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,128,1024,0.00592533333433999
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,6144,2560,0.008381333616044786
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,1536,1024,0.006010666489601135
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,2048,768,0.005616000129116907
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,256,6144,0.017895999881956313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,32,1024,0.0029128889242808023
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,16384,512,0.09349599811765884
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,3072,3072,0.02234755622016059
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,1024,128,0.002549333290921317
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,4096,16384,0.11581956015692817
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,4096,2560,0.006978666616810693
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,768,10240,0.024404444628291663
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,1024,5120,0.00868711123863856
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,256,256,0.005467555589146084
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,2048,32,0.0021600000974204806
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,512,7168,0.011286222272449069
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,1536,1536,0.00666133314371109
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,512,3584,0.020933333370420668
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,2048,12288,0.11051999860339695
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,16384,3584,0.055030223396089345
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,65536,7168,0.2780284351772732
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,10240,16384,0.05474222368664212
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,16384,512,0.03193333413865831
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,4096,2560,0.021919111410776775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,4,64,16384,0.005234666582610872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,4096,8192,0.015809777710172866
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,3072,5120,0.01038933296998342
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,7168,1536,0.02124533388349745
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,8192,1536,0.00666222224632899
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1024,256,0.00656088896923595
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,5120,2560,0.023634667197863262
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,768,256,0.002507555608948072
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,256,65536,0.09740977817111546
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3584,2560,0.011758222348160215
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,128,65536,0.07796977625952826
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,3584,256,0.005579555614127054
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,3584,1024,0.024680000212457445
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,65536,1536,0.10550133387247722
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,8192,4096,0.05987822347217136
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,12288,2048,0.011134222149848938
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,512,512,0.002554666664865282
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,10240,1024,0.007266666326257918
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,2048,2048,0.03870044483078851
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,1536,10240,0.012404444317022959
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,64,128,0.002209777840309673
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,5120,65536,0.06928088929918078
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,3072,7168,0.00980000032318963
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,256,512,65536,0.06408444378111097
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,5120,3072,0.026742221580611333
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,64,8192,0.011372444530328115
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,5120,7168,0.04099199838108487
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,256,6144,0.00980088859796524
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,16,1536,3072,0.005574222240183089
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,768,2560,0.01516711049609714
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,4096,10240,0.017691555950376723
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,4096,4096,0.011528000235557556
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,256,1024,0.012420444852775998
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,3072,1024,0.006346666564544042
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,7168,768,0.0329368876086341
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,128,16384,0.03536444571283128
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,768,7168,0.020834666159417894
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,4096,7168,0.05680177609125773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,128,8192,0.04528266522619459
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,768,65536,0.025198222862349615
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,5120,2560,0.007270221908887227
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,4096,2560,0.01758666667673323
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8192,256,7168,0.05798400110668606
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,65536,3584,0.10800088776482476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,256,65536,0.0793164438671536
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,8192,12288,0.03685866792996725
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,10240,12288,0.035079111655553184
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,64,5120,0.006597333484225803
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,3584,128,0.015124445160230001
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,128,128,0.0052471111218134565
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,7168,64,0.0021262222694026097
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,6144,128,0.01516355574131012
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,80,768,4096,0.016580444243219163
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,8192,512,0.007033777733643849
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,192,1536,2560,0.015615999698638916
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,6144,1536,0.006495111104514863
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,256,10240,0.012404444317022959
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,768,6144,0.009762666291660732
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,2048,3072,0.03423822257253859
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,7168,8192,0.03245866629812453
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2,3584,4096,0.030760000149408977
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8192,7168,128,0.034908443689346313
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2,128,10240,0.005411555369695027
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,10240,16384,0.11529155572255452
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,3072,8192,0.03431111243036058
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,512,1024,0.007281777759393056
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,160,256,5120,0.006638222270541721
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,512,8192,0.021351999706692163
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,2560,64,0.005933333188295364
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,256,4096,0.015480000111791821
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,2560,1536,0.007623111208279927
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,8192,2560,0.02981333269013299
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,65536,6144,0.1206595500310262
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,512,65536,1.029810693528917
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,7168,2560,0.007223111059930589
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,2048,5120,0.010083555347389644
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,6144,5120,0.024846222665574815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,128,512,0.0052897776994440295
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,12288,32,0.007986666427718269
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,768,128,0.0053119998839166425
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,8192,3072,0.01107022249036365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,5120,8192,0.04424533247947693
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,768,128,0.005254222287072076
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,256,6144,4096,0.02606311109330919
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,3072,12288,0.01815733313560486
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,768,512,0.002578666640652551
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,65536,3072,3.6556837293836804
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,12288,1024,0.10419199864069621
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,3584,2048,0.008408889174461365
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,512,8192,0.019294222195943195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,65536,3584,0.04137244489457872
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,2560,1024,0.005906666732496685
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,64,128,1024,0.0031591111587153543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,128,2048,0.006686222222116258
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,1536,256,0.006311111152172089
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,6144,256,0.016528889536857605
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,2560,4096,0.014624888698259989
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2,10240,3584,0.013783999615245394
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,32,2048,1536,0.0038746665749284956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,768,7168,0.018575110369258456
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,6144,65536,0.13108978006574842
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,8,7168,256,0.005589333259397083
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,2560,12288,0.03953866826163398
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,2048,16384,0.04780266682306925
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,3584,256,0.004006222304370669
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,128,2048,0.011940444509188334
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,1024,3072,0.008558221989207799
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,1,8192,6144,0.013880000346236758
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1,256,3584,0.014909333652920194
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,512,1536,5120,0.010115555591053432
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,32,5120,0.009734222458468543
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,512,2560,0.0069777775141927935
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,16,5120,8192,0.014183999763594734
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,2560,65536,0.0517484446366628
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,768,7168,0.009768000079525841
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,128,12288,16384,0.045392000012927584
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,512,1024,4096,0.01179377817445331
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,256,65536,0.061624891228146024
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,512,256,16384,0.02259644369284312
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,768,128,0.005280888742870755
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,16,8192,10240,0.03527199890878465
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1,3584,65536,0.07180266910129122
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,1024,8192,0.012879999975363413
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,80,256,512,0.006190222170617845
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2048,256,0.006133333262470033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,1024,7168,0.022521777285469904
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,64,768,0.0029120000286234748
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,16,4096,1536,0.006679111056857639
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,128,3584,0.005320000151793162
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,6144,6144,0.0670915577146742
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,5120,1536,0.007999111380841997
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,2048,7168,256,0.009017777939637503
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,64,7168,1024,0.018557333283954196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,32,256,0.002226666650838322
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,2048,5120,3072,0.0458826654487186
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,3072,256,0.005296000176005893
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,128,1024,0.005955555372767978
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,1536,128,0.014550222290886773
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,16384,10240,0.031125333574083116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8192,3584,3584,0.12002844280666775
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,2048,128,3072,0.010773333410422007
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,128,128,768,0.013451555536852943
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,2560,1536,0.006796444455782573
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,48,2560,16384,0.020826665891541373
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,384,7168,7168,0.03846044341723124
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,32,10240,6144,0.03100444542037116
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,128,128,0.002206222257680363
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4,2560,3584,0.007637333538797166
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,5120,768,0.0041759999261962045
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,1536,7168,0.012510221865442065
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,2560,4096,0.007992888490358988
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,3072,768,0.017290666699409485
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,7168,4096,0.04248266749911838
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,96,256,12288,0.006620444357395172
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,6144,1024,0.006298666612969504
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,256,2560,0.013573333621025085
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,2048,256,65536,0.1756044493781196
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,256,256,2560,0.004944889081848992
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,48,512,6144,0.01754133403301239
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,768,12288,768,0.011532444092962476
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,80,65536,8192,0.16642577118343777
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,8192,2048,0.03284800052642822
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,12288,6144,0.053674668073654175
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4,512,4096,0.016474665866957772
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,768,2048,0.014153778553009033
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,512,1024,0.0059057776298787855
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,32,3072,12288,0.016869333055284288
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,48,16384,512,0.004237333519591225
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,48,1024,12288,0.014815110299322339
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,64,2560,0.004920888692140579
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,32,2560,7168,0.012434666355450949
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,768,10240,768,0.013160000244776407
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1024,512,65536,0.1263377798928155
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,192,512,4096,0.00831288927131229
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,384,5120,4096,0.015783111254374187
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,64,768,768,0.005591111050711737
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,96,5120,1536,0.006592000110281839
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,64,128,768,0.005924444645643234
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,8,6144,7168,0.04054666558901469
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8192,7168,5120,0.320304897096422
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,384,3072,512,0.005947555518812603
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,96,5120,128,0.005682666682534748
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4,4096,65536,0.06726044416427612
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,160,65536,2048,0.0351893338892195
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,8,10240,1024,0.006280888699822956
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,128,65536,2560,0.02735288937886556
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,1,1024,65536,0.05705600314670139
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,160,3072,2048,0.006863111009200414
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,192,64,6144,0.0052391112678580815
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,1024,512,512,0.013776000175211163
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,4096,8192,3584,0.10990222295125325
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,192,12288,128,0.006731555693679386
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,256,65536,8192,0.07253244188096789
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,1024,2560,7168,0.025841777523358662
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8,4096,6144,16384,0.3199671109517415
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,768,8192,65536,0.49736886554294163
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,16384,768,0.041659557157092623
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,float16,128,65536,1536,0.039868444204330444
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,160,12288,128,0.01591022147072686
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,96,7168,5120,0.026524444421132404
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,8,7168,1536,0.0063271112740039825
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,nvfp4,4096,65536,3072,0.2995244397057427
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,512,1024,5120,0.021331555313534204
SGLang,0.5.9,NVIDIA GB200,gemm,sglang,fp8_block,768,2560,7168,0.03554755449295044
